E., 073 881

TITLE

INSTITUTIONREPORT NO

PUB DATENOTE.

AVAILABLE FROM

EDRS PRICE

DESCRIPTORS

IDEty

ABSTRACT

DOCUMENT RESUME

RC 006 844

Teachers and Students: Differences in TeacherInteraction with Mexican American and Anglo Students.Mexican American Education Study.Commission on Civil Rights,. Washington, D.C.MAES-5Mar 7368p.

Superintendent of Documents, U.S. Government PrintingOffice, Washington, D.C. 20402 (no price quoted)

MF-$0.65 HC-$3.29*Anglo Americans; -Ciarsroom Integration; EducationalEquality; Grade 4; Grade 8; Grade 10; Grade 12;

*Interaction Process Analysis; *Mexican Americans;*Teacher Influence

*Flanders System of'Interaction Analysis

Teacher pupil verbal behavior in the classroom wasassessed to dis-cover if any significant differences existed in theinteractions of teacher toward Mexican American and Anglo students.Classrooms were observed in 52 rural, urbane= and suburban schools inCalifornia, New Mexico, and Texas with a substantial enrollment ofMexican American students. Data from 429 English and social studiesclassrooms in grades 4, 8, 10, and 12 were analyzed using theFlanders Interaction Analysis system to code teacher-pupilinteractions and 22 school, classroom, and teacher characteristics.The differences in the way the teacher interacted with students weremeasured by the disparities in the Anglo and Mexican American perpupil measures which were calculated for each type of behavior. Thecoding of the classroom interaction was done every 3 seconds for 10minutes in each classroom by 5 observers. The observers also codedobserved teacher and classroom characteristics, briefly interviewedthe teacher, and made a seating. chart of the .location of each chairand the ethnicity of each occupant while in the classrooM. Data werealso collected as to whether each observed.classroom was tracked andto what degree. Significant differences were found between MexicanAmerican and Anglo .students in terms of teacher praise orencouragement, acceptance or use of student ideas, questioning,positive response, all noncriticizing talk, and all student speaking.Related documents are ED 052 849, ED 056 821, and ED 062 069. (NQ)

al

.6a

1S I

IP'

,ii Am

....,

- --:,-,--

MI

TEACHERS AND STUDENTSDifferences in Teacher Interaction With Mexican American and Anglo Students1_0

REPORT V: MEXICAN AMERICAN EDUCATION STUDYA Report of the U.S. Commission on Civil Rights

Photos by:Robert a Moeser,

Department of laborJoe Mancias, Jr.,

For sale by the Superintendent of Documents, U.S. Government Printing OfficeWashington, D.C. 20402

U.S. COMMISSION ON CIVIL RIGHTSThe U.S. Commission on Civil Rights is tem-porary, independent, bipartisan agency estab-lished by Congress in 1957 and directed to:Investigate complaints alleging that citizens ai-being deprived of their right to vote by reasonof their race, color, religion, or national origin,or by reason of fraudulent practices;Study and collect information concerning ledevelopments constituting a denial of equal pro-tection of the laws under the Constitution;isppraise Federal laws and policies with respectto equal -protection of the laws;Serve as a national clearinghouse for informationin respect to denials of equal protection of thelaws; andSubmit reports, findings, and recommendationsto the President and the Congress.M. embers of the CommissionStephen Horn, Vice ChairmanFrankie M. FreemanMaurice B. MitchellRobert-S. RankinManuel Ruiz, Jr.John A. Suggs, Staff DirectorCR1.2 :/v157/3p t.4

LETTER OF TANSi

THE U,S. COMMISSION ON CIVIL RIGHTSWASHINGTON, D.0March 1973

THE PRESIDENT

THE PRESIDENT OF THE SENATETHE SPEAKER OF THE HOUSE OFREPRESENTATIVES

Sirs:

The Commission on Civil Rights presents to youthis report pursuant to Public Law 85-315 asamended.

This is the fifth in the Commission's series of re-ports investigating barriers to equal educationalopportunities for Mexican Americans in the publicschools of the Southwest. It focuses on the denialof these opportunities as reflected in the differ-ences in the classroom verbal interactions ofteachers with Mexican American and Anglo chil-dren.

The Commission's findings are based upon in-formation from actual observations and interviewsobtained by its personnel in 429 classrooms ofschools in three geographical areas of CaliforniaNew Mexico, and Texas.

The picture of verbal interaction that emerges inthis report is one in which Mexican Americanstudents are neglected in comparison to Anglostudents. The Commission found that teacherspraise or encourage Anglo children considerablymore often than Mexican Americans. They useand build upon the ideas of Anglo students muchmore frequently than those of Mexican Ameri-cans. Moreover, teachers direct questions to Mexi-can American students much less often than theydo to Anglo students. In light of these findings, it

ITTAL

is not at all surprising to also find that MexicanAmerican children speak significantly less in theclassroom than Anglolchildren.

The disparities in teacher interaction with Angloand Mexican American children documented inthis report are cause for serious concern. In es-sence they reflect the failure of the educationalsystem to adequately adapt its programs to theneeds of Mexican American students. In orderto insure that no group of students is excludedfrom full participation, changes are needed bothin preparing the individuals who will teach Mexi-can American students and in making the educa-tional program more relevant to those students.We urge your consideration of the facts presentedand the use of your good offices in helping toeffect the changes necessary to enable all Ameri-cans to pat ticipate equally in the Nation's educa-tional tradition.

Respectfully yours,

Stephen Horn, Vice Ch i anFrankie M. FreemanMaurice B. MitchellRobert S. RankinManuel Ruiz, Jr.

John A. Buggs,Maff Director

TABLE OF CONTENTS

Page

Acknowledgements .

Preface

1-ntroduction, F.IFT .. 44liegiliee:

The Importance of Classroom

Interaction

Systematic Observation and Evaluation

of Classroom Behavior . ,

Assessment of Teacher Behavior with

Students of Different Ethnic Groups

Chapter I. Data Collection and Analysis

A, Description of Flanders Interaction.

Analysis Categories ,

B. Modification of Flanders Interaction

Analysis System to Speci

Ethnicity , , .

C. Training of Classroom Observers

D. The Sample . F.441:1, Fltf. I F

E. Data Collection . FF4 ... FFI44FeFF

F. Data Preparation .

iF

Page

3 Chapter III. A Discussion of Six Significant

5Disparities 21

7A. Teacher Praise and Encouragement . 21

B. Teacher Acceptance and Use of

Student Ideas a

C. Positive Teacher Response 32

a Teacher Questioning .. 34E. All Noncriticizing Talk by the

Teacher 1711111. 1II44!I 38

F. All Student Speaking ....} 39

7

8

9

11

11

13

14

14

14

15

Chapter IL Overall Disparities in Teacher-

Pupil Interaction; An AngioBias 16

Chapter IV. Summary and Conclusions . 43

Appendices .. V i 41 e

A. Sampling Procedures

++I*+ .

B. Data Collection 477, 4.4 7. it! t

C. Data Preparation and Variables for

. .

D. The Analysis , 14114

E. Sources of Possible Bias ...... 59

F. Teacher Characteristics and Classroom

Information Form Ill a, I 11.111 64

G. Socioeconomic Status Information

Form...... iFIFfeFfifi4344 4EFF 66

H. Display Tables . 14,444144 7 ii4i 67

45

45

48

51

57

Figure

Figure

Figure

1ICT OF FIGURES

Page

1. Modified Flanders Interaction Form 13

2Average Measures of Per Pupil Inter-action for Individual Mexican Ameri-can and Anglo Students 7

Average Amount of Praise or Encour-

aement Per Pupil Given by Teachers

to Individual Mexican American andAnglo Students 21

Figure 4. Average Amount of Praise or Encour-

agement Per Pupil Given to Individual

Mexican American and Anglo Students

by Teathers Using Specified Criterion

for Assignment of Seats 25

Figure 5. Average Amount of Praise-or Encour-agement Per Pupil Givdn to Individual

Mexican American and Anglo Studentsby Mexican American and AngloTeachers 27

Students

Page

34

Figure 9. Average Amount of Positive Response

Per Pupil Given to Individual MexicanAmerican and Anglo Students by

Teachers Using the Specified Criterion

for Assignment of Seats

Figure 10. Average Amount of Positive Response

Per Pupil Given to Individual Mexican

American and Anglo Students by

Teachers in Schools of Different De-

grees of Ethnic Concentration Withinthe School

35

36

Figure 11, Average Amount of Teacher Question-

ing Per Pupil of Individual MexicanAmerican and Anglo Students 37

Figure 12. Average Amount of All NoncriticizingTeacher Talk Per Pupil to IndividualMexican American and Anglo Students 39

Average Amount of Speaking Per Pupil

by Mexican American and AngloStudents 41

Figure 6. Average Amount of Acceptance and Figure 13,

Use of Students Ideas Per Pupil Given

by Teachers to Individual MexicanAmerican and Anglo Students 29

Figure 7. Average Amount of Acceptance and

Use of Student Ideas Per Pupil Given

to Individual Mexican American andAnglo Students by Teachers in Schools

With Various Degrees of Ethnic

Concentration

Figure 8. Average Amount of Positive Response

Per Pupil Given by Teachers to In-

dividual Mexican American and Ar,glo

Figure 14. Summary of the Flanders Categoriesfor Interaction Analysis 49

Figure 15, Classroom Interaction Behavior In-

dices Used In the Analysis 53

Definitions of Prime Seats in Class-room Seating Patterns 55

Figure 16.

31

Figure 17. Category Cutoff Points for Interval

Measures of School, Student, andClassroom Characteristics 56

ACKNOWLEDGEMENTS

The Commission is indebted to ecilia E.

Cosca and Gregg B. Jackson who wrc.e this re-port, under the direction of Francis G. Knorr,Acting Chief of the Mexican American StudiesDivision. The report was prepared under theoverall supervision of Martin E. Sloane, AssistantStaff Director, Office of Program and Policy Re-view. The following staff members and forinerstaff members were also instrumental in provid-ing the support necessary to make the prepara-tion of this report possible.

Marlene G. BlansittClaudette C. BrownCecilia P. BurciagaEdward J. CasavantesDoris a ChambersMagdalena G. DuranSally S. Knack

Diana LozanoBetty K. StradfordBeatrice W. TootleSusan N. UrangaPatsy 1. WashingtonThomas R. Watson

The 'Commission a, acknowledges the con-tributions of Henry M. Ramirez, former Chief ofthe Mexican American Studies Division and nowChairman of the Cabinet Committee on Oppor-tunities for Spanish Speaking People.

Two chief consultants, Dr. Uvaldo Palomaresand Dr. Thomas P. Carter have provided im-portant assistance for the design and develop-ment of this report. Dr. Palomares is President ofthe Institute for Personal Effectiveness in Children(IPEC) in San Diego, and co-author of Methodsin Human Development. Dr. Carter is Dean of theSchool of Education, California State University,Sacramento and author of Mexican Americans inSchool: A History of Educational Neglect.

Dr. Monroe K. Rowland, Professor of Educationat California State University, San Diego devel-oped some of the instruments used to obtaindata for this report. Dr. James Retson, Consultantto the Supplementary Education Center of theSan Diego County Board of Education, trainedthe classroom observers.

From the beginning of the Mexican AmericanEducation Study the Commission staff has beencounseled by an. Advisory Committee composedof the following persons:

Reverend Henry J. CassnDirector of BilingualBicultural EducationSchool of EducationUniversity of MassachusettsAmherst, Massachusetts

Dr. Ernest F. GarciaAssociate Professor of EducationCalifornia State UniversitySan Bernardino, California

Mr. Adalberto GuerreroProfessor of LanguageUniversity of ArizonaTucson, Arizona

Dr. Irwin KatzProfessor of PsychologyGraduate DivisionCity University of New YorkNew York, New York

Mr. Frank MaganaCommunity Representativefor Head Start, Region VIDepartment of Health,Education and WelfareKansas City, Missouri

Mr. Theodore MartinezDirector of Student UnionUniversity of New MexicoVice PresidentBoard of EducationAlbuquerque Public SchoolsAlbitrwerque, New Mexico

Mr. Jesus Jose RubioVice PresidentUrban Research Group, Inc.Austin, Texas

Ms. Vilma Martinez SingerAttorneyCahill, Gordon, Sonnett,Reindell, and OhlNew York, New York

r, ChrIes Tofql

jIrodar

[girl Ameko Reseorch

Service Agenq

Demicr, Colorado

Dr,

Superinferldelit

East \Nhtter C4 School DIstrict

Whittier, CR ii0M13

The Athikoly Committee h3S beep of gut

ossignee i providing g-uidlice in the plInimg

nd natio of [ills repot

klvi other persons indudi Nir WO dice

triCi suprifitederits, prineipals,, tailors, .1nd

other school persormol cooperged dully in proi

vidiq informgion pms$4 to ie this 50.

iiccess. Fever, .se responsibk for acE

cmq, of the dt, validity of the all/SeSi 3id

the souriduss of the coodusios rests with the

Commissim,

PREFACE

This report is the fifth in a series on MexicanAmerica& education in the Southwest by theU.S. Commission on Civil Rights. The series pro-vides a comprehensive assessment of the educa-tional opportunities available to Mexican Ameri-cans in the public schools of the Southwest. Oneof its principal objectives is to inform educators,parents, legislators, and community leaders of thepolicies and practices of schools attended byMexican Americans and of the extent and qualityof education received by Mexican Americanstudents.

This report focuses on teacher-pupil verbalbehavior in the classroom. The study attemptedto find out if there are important differences inthe verbal interactions of teachers toward. theirMexican American and Anglo2 pupils.

Sources of InformatiOn

The information for the Mexican AmericanEducation Study is drawn from several sources.The principal sources are (1) the Commission'sSpring 1969 mail survey of Mexican AmericanEducation in schools and districts throughoutthe five Southwestern States; (2) HEW's Fall 1968elementary and secondary school survey of theseStates; and (3) the Commission's field study ofschools in California, Texas, and New Mexicoduring the 19771 school year. The first fourreports of the series were based primarily ondata obtained from the HEW and the Commissionmail surveys. This report is derived primarily frominformation gathered in the field.

Field study data reported in this volume con-

' The term Mexican American refers to persons who wereborn in Mexico and nOw hold United States citizenship orwhose parents or more remote ancestors immigrated to theUnited States from Mexico. it also refers to persons who tracetheir lineage to Hispanic or Indo- Hispanic forebears who re-sided within Spanish or Mexican American territory that is nowpart of the Southwestern united States.

Chicano is another term used to identify members of theMexican American community in the Southwest: The term hasin recent years gained wide acceptance among Mexican Ameri-cans especially the youth. It also receives wide currency in themass media.

? The term Anglo refer's to all white persons who are notMexican American or members of other Spanish surnamedgroups.

sist mostly of the observations by Commissionstaff of teacher-pupil verbal interaction in theclassroom. Interactions were coded on theFlanders Interaction Analysis form, an instrumentespecially designed to collect such information.Other information was obtained on teacher, class-room, and school characteristics associated witheach observed classroom.

The field study staff received 4 days of inten-sive training. Observers, with the help of a trainer,first thoroughly familiarized themselves with thestandard Flanders coding procedures, practicedcoding with audid and video tapes of instructionalsessions, and then utilized the instrument in regu-lar classrooms. Periodically, the reliability of theobservations of each of the trainees was checkedto determine the degree to which their observa-tions and those of the instructor were in agree-ment. By the end of the training period, thecoders had reached a high level of reliability. Inaddition, at the beginning of data collection ineach State,a the observers' reliability was checkedby the trainer. In each instance it was high.

Publications

The results of the Mexican American EducationStudy are being published in a series of reports.Those previously published include:

Report I: Ethnic Isolation of Mexican Americansin the Public Schools of the SouthwestThe ex-tent to which Mexican American students areisolated from Anglo students by school is themain focus of this report. In addition, this studyalso documents the underrepresentation ofChicanos as teachers, principals, other administra-tive personnel, and school board members.

Report II: The Unfinished Education: Outcomesfor Minorities in the Five Southwestern StatesThe second report analyzes the performance ofschools in the SouthweSt in terms of outcomesof education for students of various ethnic back-grounds, as measured by school holding power,

Collection of data by the Commission staff started in NewMexico in October 1970; California was visited in November,and the staff finished in .exas in February 1971.

reading achievement,achievement, grade repetition, overage-ness, and participation in extracurricular activities.

Report III: The Excluded Student: EducationalPractices Affecting Mexican Americans in theSouthwestThis report examines the way theeducational sy,tein looks at the unique linguisticand cultural background of the Mexican Ameri-can student. It also examines p;ograms used bysome of the schools in attempting to adjust tothis background and the school's relationship tothe Mexican American community.

Report IV: Mexican American Education in Texas:A Function of WealthThis report focuses onschool finance in Texas as it affects the educa-

tional opportunity of Chicano students. It ex-amines: (1) State aid to local school districts,particularly State apportionments granted underthe Minimum Foundation Program; (2) propertyvaluations upon which districts rely for locallyraised revenue; (3) property tax efforts of schooldistricts; and (4) the .relative economic burdenof property taxes on differing income groups. Asa corollary to Report IV the Texas State Com-mittee to the U.S. Commission on Civil Rightsissued a report in which it offered recommenda-tions for school finance reform in Texas.

A forthcoming report will identify educationalproblems confronting Mexican Americans in theschool systems of the Southwest and make recom-mendations concerning possible solutions.

,Str"

_

INTRODUCTION

The Importance of Classroom Interaction

The heart of the educational process is in theinteraction between teacher and student. Itthrough this interaction that the school systemmakes its major impact upon the child. The waythe teacher interacts with the student is a majordeterminant of the quality of education the childreceives.

Information on what actually happens in theclassroom is thus very important in assessing thequality of educational opportunity. Some of themost significant aspects of the teaching-learningprocess can be identified only by observing theactual classroom interaction: The teachers' skills ininstructing, guiding, and encouraging students aredemonstrated by what the teacher does and saysin the classrooM. The extent to which studentsare being actively involved in the learning processis shown by their participation in the classroomactivities. The way teachers and students feelabout each other is evidenced in the way teacherand students react to one another.

Numerous other factors play a role in the ed-ucational experience of a child. Factors such as

the quality of school facilities, the types of text-books, and the training and experience of teachersaffect the quality of schooling because they pro-vide the setting for learning. But educational op-portunity is primarily affected by what goes on inthe classroom: For example, additional trainingon the part of the teacher will not necessarilyaffect the skills of the teacher; however, if thetraining improves the way the teacher interactswith students in the classroom, then the qualityof schooling is improved.

The effects of classroom interaction have beenomitted from the major studies on the equality'of educational opportunity afforded to minoritystudents. The most comprehensive of thesestudies is the 1966 U.S. Office of Education studyentitled Equality of Educational Opportunity:'This national survey collected extensive data onthe facilities, services, and curriculum of the

James S. Coleman, et. at, equality of Educational Opportu-nity, U.S. Department of Health, Education, and Welfare, Officeof Education, Washington: U.S. Government Printing office, 196G.

7

school, on the social, personal, and educationalcharacteristics of the school personnel, and on theacademic achievement, attitudes, and home back-ground of students. However, neither the .U.S.Office of Education Study nor similar studies haveassessed what actually goes on in the classroom)

Equality of Educational Opportunity found thatdifferences in student achievement from oneschool to another were not strongly related 'tothose- characteristics of schools and school per-sonnel which were measured: Of all the factorsassessed in that study, the different in achieve-ment were most strongly related to the student'sown social and economic background. Some edu-cators and laymen have interpreted these findingsto mean that there is not much the schools can doto improve the educational achievement ofminority students. However, because the U.S,Office of Education study did not assess actualclassroom processes it cannot validly be con-cluded from its findings that the quality of schoolscan have little influence on student achievement.'

This report focuses on differences in the wayteachers interact with Mexican American andAnglo students in the classroom, Comparisonsare made of several types of verbal teacher be-haviors, such as praising and questioning of stu-dents. In addition, the verbal participation ofAnglo students is compared with that of MexicanAmericans.

The record of educational achievement showsthat Mexican Americans in the SouthwesternUnited States are not receiving the benefits of a

Fur a comprehensive review of the major studies until 1969see James Guthrie, et, al,, Schools and inequality, Cambridge:MIT, Press, 1971, ell: 4, Three major !,tudles since then are:George Mayeske, A Study of Our Nation's kkoalsA WorkingPaper, U.S, Department of Health, Education, and Welfare, Officeof Education, Washington: U.S. Government Printing Office, 1971;

Frederick mosteller and Daniel Moynihan, eds., On the Equalityof Education, NeW 'fork: Vintage Books, 1972; Christoplwr Jencksci. al, Schou/5 and Inequality, New York; Basic BOOK 1072

4A second reason why the conclusion that schools can havelittle influence on student achievement 15 not supported bythe findings of the U.S. Office of Education study is that onlyexisting 011(1111o:is in schools were assessed. It is not possibleto know the potential impact on minority ,,tudent achievementof future changes in school conditions and characteristics ofschool personnel,

publiic education to the same extent as are Anglopupils. A previous report in this series, entitledThe Unfinish Education, showed that fully 40percent of the Chicano students in the Southwestnever complete high school; in contrast only15 percent of all Anglo pupils in the region failto complete their schooling. The proportion ofChicano students reading below grade level is

generally twice that of Anglos. Of those Chicanoswho do complete high school, two out of threeread below their grade level and one out of fouris unable to read above the ninth grade level:This report examines one important aspect of theprocess of educational neglect, the way teachersrelate to the two groups of students in the classroom.

Systematic Observation and Evaluation of Class-room Behavior.. ...

The assessment of the quality of the classroomteaching process is a complex matter, The teachercomes to the classroom with a given set of at-titudes, expectations, skills, and goals. The teacherseeks to instruct, guide, and encourage between20 to 40 students, each of whom, in turn, comesto the classroom with his or her own individualattitudes, expectations, skills, and goals, A multi-tude of different types of exchanges occur be-tween teacher and students and among studentsduring a normal class session.

Traditionally, the assessment of teaching hasbeen conducted primarily on a subjective, or evenintuitive, basis by occasional visits from principalsor professors. In the last two decades a more ob-jective approach to observing and evaluatingteaching has been developed, known as interac-tion analysis. Interaction analysis involves syste-mally observing of classroom behavior and

relating this behavior to characteristiLs of teachers,

students, and schools, or to the achievementlevels of students. Classroom observation is con.dticted by coding the behavior of teachers andstudents according to a system of categoriesdesigned with a ,pecific orientation or focus, Thisinformation is then systematically compiled toobtain a picture of the actual teaching-learningprocesses,

One of the most widely used classroom inter-action observation systems is that developed byDr. Ned Flanders. The Commission chose theFlanders system of Interaction Analysis becausethis system focuses on forms of teacher behaviorwhich are most directly related to encouragingand involving the student in the learning process.The Flanders system codes the predominantclassroom behavior once every three secondsaccording to the most appropriate of the follow-ing 10 categories: 1) teacher accepts student'sfeelings; 2) teacher praises student; 3) teacheraccepts or uses student's ideas; 4) teacher asks aquestion: 5) teacher lectures; 6) teacher givesstudent directions; 7) teacher criticizes student;8) student speaks in response to teacher's ques-tions or directions; 9) student speaks on his owninitiative; 10) no one is speaking or confusionprevails.'

On the basis of a decade of classroom inter-action research, some forms of teaching behaviorhave been identified which appedr to have apositive affect on pupil attitudes and achievement.They are behaviors which involve the acceptanceand use of student ideas, some forms of praise orexpression of appreciation of a student's con-tribution, and behaviors which involve question-ing of students. These forms of behavior do not

Edmund 1, Amidon and Ned A. Flanders, The Role of theTeacher In the Classroom: A Manual For Understanding andimproving Teachers' Classroom LIehavior, Minneapolis: Paul S.Amidon Associates, 1963, pp, 6-11,

Barak, Rosenshine, -Teaching Behavior Related to Pupil

Achievement, Review of Research," Research mto ClassroomProcesses: Recent Development; and Next Steps, ed. an West-bury and Arno Bellack, New York: Teachers Colleiie Press, 1971,pp, W-91 Rosenshine reviews aseries of studies and draws con-clusions from the relative frequency with which different find-ings are replicated, He qualifies his conclusions because abouthalf the studies do not show significant positive effects for Lkbehaviors mentioned in the above text, although very few showsignificant negative effects. His qualifications are probably morerestrictive than justified by the pattern of results. This Is be=cause for inferences to be validly based directly on the frequencyof a given statistically significant finding across minuted studies,the studies must have approximately equivalent probabilities offinding significant results when they actually exist (statistical

power), Rosenshine does not indicate that this is so, and It Is

quite likely it was not the case. Some of the studies with non-significant results probably had such low statistical power thatthey had little chance of indicating significant differences exceptwhere there were huge differences in the sample.

invariably increase student achievement or favor-ably affect attitudes, but the evidence suggeststhat they generally do."

For example, one study found that the studentswho showed the greatest improvement on stand-ardized tests of verbal and quantitative skills werein classrooms where the teachers used a greatdeal of praise and encouragement and acceptedand used the students' ideas.' A second study.found that teacher trainees who frequently ac-cepted or used their students' ideas were moreeffective in teaching specific course content thanteacher trainees who did not.'" In another studythe frequency of teacher questioning was foundto be positively reldted to the amount of stu-dent learning in vocabulary, reading, and mathe-matical skills." Numerous other studies havefound similar effects."

Assessment of Teacher Behavior with Studentsof Different Ethnic Groups

Classroom interaction analysis can be used todetermine how patterns of classroom behaviorsvary with different types of teachers and stu-dents. Classroom interaction patterns can bestudied in relationship to teacher characteristics,such as age, sex, and training, and studentcharacteristics, such as age, ability, and socialclass background, Despite the large number ofsuch studies there is a paucity of research in-vestigating teacher-pupil interactions among stu-dents of different ethnic groups. Only threestudies with this focus have come to the atter,tide of the Commission. None involved MexicanAmericans, Furthermore, no broad generalizations

'Betty Morrison, The Relations of Internal and (sternalikon tai Patterns of Teacher Behavior. Unpublished doctoral dis-sertation, University of Michigan, 1%6.:

iu Jimmie Fortune, A Study of the Generalities of PresentingBehaviors in Teaching, Project Report to U.S. Office of Education,Memphis: Memphis State University, 1967,

%Wrens, Petalionships fiet%iven Teacher Charader-Atics and Student Behavior! Part 3, Project Report No. 26211 toU.S. Office of tducation, Salt take City; University of Utah,TM.

12 Ned A. Handers; Analyzing Teaching Behavior, Menlo Park,California: Addison-Wesley Publishing Company, 197(), pp, 309-

425,

can be .Made from the results of these three

studies because they were limited to very small

geographic areas and all were conducted in

segregated school systems:"

The Commission's study was designed to in-

vestigate possible disparities in the way teachers

treat. Mexican Americans and Anglo s within the

same. classroom, Although the main focus of the

study was on .Mexican American and Anglo stu-

dents, it was also intended to compare the _inter..

action of teachers with black and other minority

students to the extent possible. However, the

number of students of these ethnic groups which

were observed was too small for analysis:.

The ability of the teacher* to involve and

encourage students is important to the educa-

tional achievement of all students, but is particu-.

lady crucial to the education of minority students.

As the Commission has previously pointed out,

schools in the Southwest have generally failed to

adapt their curriculums and prograMs to .. the

interest..skills,and language with which Mexican

See Evan Powell and William White, Learning Climate Corre.

Wes in Black and White Rural khools, Athens: R&D Center in

Education Stimulation, University of Georgia, 1970; here Brophy

and Thomas Good, Dyadic Teacher-Child interaction: Variation

Acrn$5 Social Class and Racial Groups, Paper presented to 1971

American Educational Mg)datin Annual Meeting, New York:

The Wang, 1971; Bruce Biddle and Marvin Loflin, Verbal

Behavior in lilack.Chetto and White--Suburhan Classrooms, An

Overview Paper presented to 1971 American Educational Research

Association Annual Meeting, New York: The Meeting, 1971,

' U S . COMMiSSiCin on Civil Rights, The Excluded Student,

Washington: U.S. Government Printing Office, 1972.

American students enter school," This failure

is a serious barrier tolweducational opportunity

of Chicano pupils, Despite. this barrier, there

much that teachers can do to facilitate the learn-

ing of Chicano students; Teachers can demon-

strate respect by inviting the students to share

their culture, personal feelings, and values They

can encourage Chicano participation by:accept-

ing and building upon their contributions;

Teachers can also provide Mexican American

students with the necessary assistance to help

them overcome difficulties, These forms of be-

havior are likely.. to improve motivation and raise

the level of academic performance of Mexican

American students,

When the.teacher treats one group of students

more favorably than another; the damage done

to the child of the second group is twofold. If

the teacher seldom praises and encourages the

Chicano student, for example, this is likely to

lower his motivation and hinder his academic

performance, If, at the same time, the Chicano

student is aware that the teacher expresses much

more praise and encouragement toward Anglo

students, he or she will come. to feel that the

teacher does not like him or does not consider

his education to be important, This will further

impair his motivation and achievement, Con-

sequently, whenever teachers relate more favor.

ably toward Anglo students than to Mexican

American students in the classroom, thedif-

ferential treatment can be extremely harmful to

the education of the Chicano student

CHAPTER I: DATA COLLECTION AND ANALYSIS

A. Description of Flanders Interaction An-alysis Categories

Of the 10 categories of verbal behavior utilizedin the Flanders Interaction Analysis System, seveninvolve "Teacher Talk," two involve "StudentTalk", and one involves silence and confusion.The category system is totally inclusive of all possi=ble events. This means that all classroom behavior

can be classified in one of the 10 categories, Adefinition of each of the 10 categories and a de-tailed explanation to help the reader distinguish

among them follow,"

TEACHER TALK

1, Accepts FeelingsCategory 1 consists of teacher behavior whichaccepts or clarifies the expressed feelings ofthe students in a nonthreatening manner,whether the feelings are positive or negative,

i'l)efinitions and explanations for each of thu categories arearlapiat frrinn Edwina j, Amidtm and No.1 Fiaotivrs, op. cit.,pp. 1.)=21.

11

Statements that predict or recall feelings areincluded. Statements classified as Category 1do not express teacher evaluation or approval.Thus, in response to a student's statementthat he did not like arithmetic, the teachermight say "I know what you mean, John."

2. Praises or EncouragesPraise and encouragement are statementswhich carry the value judgment of approvalof student acjon, or behavior. Examples ofCategory 2 statements are: "That was verywell done,'Martin, keep up the good work",or simply: "1 like what you are doing, Linda".One word statements such ass-right-, "good",are included. Also encompassed in this cate-gory are jokes that release tension, providedthey are not at the expense of another indi-vidual.

3. Accepts or Uses Ideas of StudentCategory 3 consists of statements which clar-ify, build, or develop student ideas or sugges-tions. When the teacher paraphrases, restates,or summarizes something a student has said,it is considered this type of interaction. Inaddition, this category also includes simpleacknowledgements of a student's contribu-tion such as: "Well, that's an interesting pointof view. I see what you mean."

4, Asks Questions

Questions asked by the teacher with the in-tent that a student will answer comprise thiscategory. Questions not intended to be an-swered by the students, slich' as ,rhetoricalquestions or those meant to give directions,are not included in this category. Thus, aquestion such as: "What is the capital ofOhio?" is coded in this category, whereasthe statement: "Torn, will you please closethe door?" is not.

Lecture

Category S consists of statements in whichthe teacher gives facts, opinions, or ideasabout content or procedure. These can bebrief information-giving statements, as wellas extended explanations or discussions on

12

the part of the teacher. Rhetorical questionsare included within this category.

6. Giving DirectionsTalk by the teacher which directs, orders, orcommands the student to comply is codedin Category 6. Examples of this behavior arestatements such as: -Class, will you take outyour workbooks now?" and "For tomorrow,I want you to answer the questions in Lesson6", Whenever extensive directions are inter-spersed with the giving of information (Cate-gory 5), the appropriate category is used foreach statement. For example, if the teacherbegins by giving the students directions onthe next day's assignment and in the processgives factual information on the content ofthe lesson, both Category 6 and Category 5are coded.

7. Criticizing or Justifying AuthorityStatements of criticism are those which aredesigned to change student behavior fromnonacceptable to acceptable. A statementsuch as: "I don't like the way you have beendoing your work. Do it another way." is con-sidered criticism. Statements justifying au-thority are those in which the teacher isstating why he is doing what he is doing or'defending himself against the student. Anexample of such a statement is: "You can'tdo that now because I say so and I'm theteacher".

STUDENT TALK

8. Student TalkResponseCategory 8 consists of verbal behavior ofstudents in direct response to a statement bythe teacher. This type of "student talk" gen-erally follows directly from questions or di-rections of the teacher and does not involvecontributions initiated by the student,

9. Student TalkinitiationStudent statements which are initiated by thestudent rather than by the teacher comprisethis category. In general, when a studentraises his hand to make a statement or to

ask a question when he is not prompted bythe teacher, his subsequent speaking falls intothis category. Also included is talk by thestudent when he volunteers additional infor-mation or elaboration after a direct responseto a question by the teacher. In this situationthe student's direct response would be classi-lied as a Category 8 statement, while thefurther elaboration Would be classified as aCategory 9 statement.

OTHER BEHAVIOR

10. Silence or ConfusionThis category includes all behavior not in-cluded in the other categories. Periods ofsilence or confusion in communication, whenit is difficult to determine who is talking, areclassified in this category,

B. Modification of Flanders Interaction An-alysis System to Specify Ethnicity

Previous research using the Flanders system hasbeen conducted almost exclusively on a wholeclass basis, i.e., no distinction was made betweenthe interaction of the teacher with individual classmembers and with the class as a whole. Becausethe Commission sought to compare teacher inter-action with students of different ethnic groups,the Flanders system was modified so that eachbehavior was coded with reference to the ethri-city of the student with whom it was associated.Provision was made for coding each communica-tion event according to whether it involved anindividual Mexican American, Anglo, black, a stu-dent of another ethnic background, several students simultaneously, or all of the class." A copyof the sheet used to code the classroom behavioris presented in Figure 1. Once every three secondsthe observer marked a tally in the box which mostappropriately indicated (1) the behavior that wasoccurring (rows 1-10 indicate Flanders categories.

lb When iwo or more students were speaking simultaneouslyin a formalized manner, such as in classroom recitation, or ifthe teacher was speaking to two or more students at once, suchas when addressing the total class, the behavior was coded under"C" as being associated with the class as a whole.

1-10) and (2) the ethnicity of the student withwhom it was associated (columns marked MA,A, B, 0, and C refer t© Mexican American,Anglo, Black, Other, and class as a whole,respectively.)

Figure 1

Modified Flanders Interaction Form *

School and District Code No. Date-

Access No Classroom No.

District Name Grade Period

School Name Subject

Teacher Name

Observer's Initials

Ability group level

MA 0 C TOTAL

1

2

3

4

5

6

7

8

9

10

Behavior was coded according to the type of interaction andaccording to the ethnicity of the student involved. Rows 1.10indicate the 10 categories of the Flanders Interaction AnalysisSystem. Column headings MA, A, B, 0, C indicate MexicanAmerican, Anglo, Black, Other, and Class as a whole respec.lively, The form is shown in reduced size.

13

C. Training of Classroom Observers

Five Commission staff members received inten-sive training in the use of the modified FlandersInteraction Analysis System. The training sessions,consisted of 2 days of practice and preparationwith the use of videotape recordings of classroominteraction and 2 days of practice in live classroomsettings)! In order to assure conAstency and valid-ity of the data, the coding of each observer waschecked for reliability at the conclusion of thetraining session and at the beginning of the ob-servations in each of the three States.'

D. The Sample

Classroom observation was conducted in schoolsin California, New Mexico, and Texas.lv Withineach State, geographical areas were selected thatincluded rural, urban, and suburban schools inwhich large numbers of Mexican American stu-dents were enrollecL2° From these areas a sampleof schools was drawn t© be representative of theschools attended by most Mexican Americans inthe geographic regions!' Fowever, a number ofschools had to be eliminated because they werebeing or about to be investigated by Federal CivilRights Agencies and, consequently, expected tobe unreceptive to classroom observation. Because

!' The training sessions were conducted by Dr Monroe K.Rowland and Dr: lames Reison of-the Department of Educationat San Diego State College.

tl A discussion of the procedures used to check for observerreliability and of possible sources of bias is found in AppendixE, which begins on p. 59.

Under i ideal circumstances, the Commission would havemade a random sample of schools throughout the Southwest:Flowever, to use existing resources most effectively, it wasdecided to limit the sample area to the three States with thelargest Chicano populations: California, Texas, and New Mexico.These States contain 60 percent of the Spanish origin studentsin the United Stales and about 90 percent of the total numberin the Southwest.

'The areas selected w1. California: Santa Clara County including the city of Sanlose,

2. Texan Ow metropolitan areas of San Antonio and CorpusChristi, the area between these two population centers, andthe area 30 miles south of Corpus Christi,3. New Mexico: the Albuquerque area and the south centralpart of the State near El Paso, Texas.

2' The details of the sampling procedures are indicated InAppendix A, which begins on p. 45,

14

of the large number of schools excluded from thesample on this basis, the sample probably doesnot adequately _represent those schools wherethere are likely to be the greatest_ disparities inteacher behaviors toward Mexican Ameritan andAnglo Students,"

Fifty-w;o schools were randomly sampled fromthe eligible schools in the selected regions. Fourhundred and ninety-four elasSes in which English.was being taught at fourth, eighth, lath, and 12thgrades were observed," Interaction data whichwere adequate for analysis were 'available. from429 of the visited classrooms24

E. Data Collection

Teachers were notified beforehand that theirclassrooms were to be observed by staff membersof the U.S, Commission on Civil Rights. They weretold that the information was to be used for astudy of classroom interaction although the exactnature ?rid purpose of the study was not dis-closed." The Commission observer generally saton the side or in the back of the room for ap-

" One hundred and ninety-four schools of the original Marlwere eliminated for this reason: All of these were in Texas.

21 In a few, cases social studies classes were substituted forEnglish classes. This was done when the school had an ungradedEnglish program or did not require students to take English atthe grade level sampled: About 7 percent of the classrooms weresocial studies classes.

" The major cause of inadequate data was the failure of someclassrooms to have at least one Mexican American and at leastone Anglo student: this was a prerequisite 'for the types ofanalysis used in the study,

It is likely dm the presence of observers had some effecton 'the behaviors of teachers and students, although the extentof this effect is not known, It is expected that under observation,teachers were, in general, more likely to try to draw studentsOut and less likely to criticize them. This would tend to increasethe frequency of occurrence of certain types on teacher be-haviors while reducing others. If this were the r,nly effect ofthe presence of observers it would not affect the difference inteacher interaction with Mexican American and Anglo students.However, because teachers were aware that the observers werefrom the U.S. Commission on Civil Rights, it is likely that theytended to relate more positively to the Mexican American stu-dents than they would do under taormal circumstances, If thiswere so, the Commis$Ion's estimates of the differences inteacher interaction with the two groups of students are actuallyconservative, i.e., the differences which exist under normal cir-cumstances would be greater than those found'by the Commis.sion,

proximately 1 hour.- For a 10- minute period the-observer coded the classroom behavior at 3-sec-ond intervals according to the modified Flanderscategory system.26 The remainder of the time wasspent observing other .aspects of the classroomteacher and students.

F. Data Preparation

The Commission sought to determine if therewas differential treatment in the manner inwhich teachers interacted with Chicano and Anglopupils in the classroom.27 In order to do this, itwas necessary to obtain a measure of the occur-rence of each type of interaction for each ethnicgroup, adjusted for the number of students ofeach of the groups in a given classroom. In otherwords, it was not sufficient to compare the num-ber of times during the 10-minute observation

"A detailed description of the procedure used to code class-room behavior is found in Appendix Q, %vbich will begin onI,,

21 i was originally intended to include a comparison of teacherinteractions with black and other minority S t u den ts: However.since the number of black and other minority students observedwas too small it was not possible to analyze these data.

period that the teacher interacted with Chicanoand Anglo- students without knowing the numberof Chicano and Anglo students in the classroom.For this purpose, a "per pupil measure" was cal-culated for each type of behavior to representthe number of times the average studenrof eachof the two ethnic groups was involved in a speci-fied interaction.

Per pupil measures: of each behavior were ob-tained for Anglo and Mexican American studentsusing the following procedure: The number oftimes each behavior occurred for Mexican.Amen-cans was divided by the number of Mexican Amer-icans in the classroom and a similar calculationwas made for the Anglo pupils. The difference inthe way the teacher interacts with,Anglo and Chi-cano pupils is measured by the disparities in theAnglo and Chicano per pupil measures.

The sections- which follow discuss the Com-mission's findings with regard to the differences inthe types and- quantity of teacher interaction withAnglo and Chicano pupils."

n The statistical procedures used in the various analyses areoutlined in Appendix D. which begins on p. 57

*4.

An=CHAPTER OVERALL DISPARITIES.

IN TEACHER-PUPIL INTERACTION.: AN ANGLO BIAS

The findings of the Commission's study onteacher interaction with Anglo and Chicano pupilsare shown in Figure 2. The figure presents theaverage frequency of teacher-pupil behaviors ac-cording to the Flanders categories. For each typeof behavior, the interaction of teachers with Mexi-can. American students is compared with Anglostudents by a measure of disparity. The per pupilinteraction measures are given for the seven typesof "teacher talk" and the two types of "studenttalk". In addition, three composite measures ofbehavior are reported: amount of positive teacherresponse to individual students (Categories 1, 2,

16

and 3); .amount of noncriticizing teacher talk(Categories 1 through 6); and total student speak-ing (Categories 8 + 9).

The data in Figure 2 show that there are import-ant differences in teacher interaction with MexicanAmerican students and Anglo students as evi-denced by the size of the disparities in the perp_ upil interaction measures.

Disparities in six of the 12 categories are statis-tically significant. Essentially, this means that forthese six disparities it is likely that similar dispar-ites would be found in most of the classrooms inthe survey area from which the sample was

Figure 2

Average Measures of Per Pupil Interaction for Individual Mexican American and Anglo Students29

AverageMexicanAmerican

AverageAnglo Disparity'

PercentIncreasein Anglo

over M.A.

Teacher Behavior1, Acceptance of student's feelings .004 .008 +...'134 100.0%

PRAISING OR ENCOURAGING :137 .186 -' 35.8*3. ACCEPTANCE OR USE OF STUDENT IDEAS .156 .219 +PP 40,4*4. QUESTIONING .525 .638 +.111 21,15. Lecturing .584 :710 +.126 21,66. Giving Directions .146 .141 :005 3,47. Criticizing or JUstifying Authority .055 .052 .003 5.5 --

Student Behavior8: Student TalkResponse .771 .948 +.177 23.09. Student TalkInitiation ..790 1.034 +.238 29.9

Composite Measures of Behavior*POSITIVE.TEACHER RESPONSE (1-3) .296 :413 +.117- 39.5*ALL NONCRITICIZING TEACHER TALK (1 =6) -1,551 1.901 +.350 22.6*ALL STUDENT SPEAKING (8+9) 1.567 1.982 +.415 26,5

Disparities between Anglo and Mexican American are statistically significant at pm-M. This means that for these disparities thereis only one chance in 100 that-corresponding disparities would not be found in the population from which the sample (vas drawn.

" Per pupil interaction measures represent the number of times during a 10-minute observation period that the average pupil of eachethnic group was involved in interaction rtf each type. The figures were obtained by the following method: 1) for each classroomobserved the number of tallies associated with students of each ethnic group was divided by the number of students of that ethnicgroup in the classroom, 2) these per pupil measures for each classroom were added and then divided by the total number of class.rooms to obtain the average per pupil interaction measures for the sample:

3° The standard deviations of each disparity measure given in Appendix H, Table 2 on p.68

drawn. 3t In other words, these differences arerepresentative of the schools attended by mostChicano students in the three geographic areasstudied. For the remaining six disparities, it is pos-sible that they are similarly representative of theschools attended by most Chicano students inthese areas. However, because these six disparitiesdid not prove statistically significant, this state-ment cannot be made with a high degree of cer-tainty. Nonetheless, they provide an accurate pic-ture of interaction patterns in the 429 observedclassrooms. This report focuses primarily on thosedisparities which are statistically significant be-

31 The phrase ' area" means the total classrooms in thethree geographic regions from which the sample of 494 class-rooms was drawn,

cause of the substantial certainty of their occurTench over wide geographical areas.

The six. categories in which the disparities arestatistically significant are:

Praising or EncouragingAcceptance or Use of Student IdeasQuestioningPositive Teacher ResponseAll Noncriticizing Teacher TalkAll Student Speaking

Mexican American pupils in the survey area re-ceive considerably less of some of the most edu-cationally beneficial forms of teacher behaviorthan do Angles in the same classrooms. Mexican

17

Americans receive significantly less praise and en-couragement from the teacher and less often hearthe teacher accept or use the ideas they express.Teachers also spend significantly less time in ask-ing questions of Chicano pupils than of Anglopupils, On the composite measure of positiveresponse from the teacher, which includes ac-ceptance of student feelings, praise or encourage-ment, and acceptance of student ideas, MexicanAmericans receive significantly less than Anglos.Futhermore, teachers address significantly more oftheir total noncriticizing talk to Anglo pupils thanto Chicanos. Closely related to the differences inteacher behavior with students of each ethnicgroup is the finding that Mexican American stu-dents speak significantly less in class than doAnglos. In total, the six statistically significant dis-parities in classroom interaction all favor Anglo-pupils over Chicano pupils. The implication ofeach of these disparities will be discussed ingreater detail in subsequent sections of this report.

The six categories. in which the disparities arenot statistically significant are:

Acceptance of Student's FeelingsLecturingGiving DirectionsCriticizing or Justifying AuthorityStudent TalkResponseStudent TalkInitiation

These disparities also indicate patterns of inter-action favoring Anglos over Chicanos, Althoughit cannot be known with a high degree of certaintywhether or not these six disparities exist through-out the survey area, their occurrence in the largesample of 429 classrooms visited is of considerableimportance.

Teachers expressed very little acceptance of thefeelings of any students, but they did express ac-ceptance twice as often for Anglos as for MexicanAmericans, Teachers also spent more time relat-ing information to Anglo pupils than to Chicanopupils. The average Anglo pupil received 20 per-cent more of this "teacher talk" classified as lec-turing than did the average Chicano pupil. Thisis important because more of the teacher's timewas spent in giving information, or lecturing, than

18

in any other type of behavior, About 36 percentof the time the teacher spent speaking to individ-ual students, she was relating information to them.

Only two of the 12 measures Of teacher be-havior involved Chicano pupils more than Anglopupils. These were directions and criticism, thetwo which appear to be the least favorable of allthe measures, of teacher behavior. The averageMexican American student received slightly moreof both directions and criticism from the teacherthan did the average Anglo. Although the differ-ences in direction and criticism are small they areimportant as part of the total pattern of classroominteractiona pattern in which Chicano pupilsconsistently are 'encouraged less and discouragedmore than their Anglo counterparts.

The results also indicate that the average Mexi-can American verbally participated less in theclassroom, both in response to the teacher andon his own initiative than the average Ang10,32The average Anglo pupil observed talked about23 percent more in response to the teacher thanthe average Chicano pupil. He also spent approxi-mately 30 percent more time talking on his owninitiative than the average Chicano pupil. Thissecond type of "student talk" is especially im-portant because it indicates the extent to whichstudents feel confident that what they have to sayis worthwhile and that the teacher will welcometheir ideas and opinions.

The total picture of classroom interaction pat-terns presented by the 12 disparities in Figure 2is that of a teaching process which is failing toinvolve the Mexican American student to the sameextent as the Anglo pupil, both in terms of quan-tity and quality of interaction. Teachers speak lessoften, and less favorably, to Mexican Americansthan to Anglos. At the same time, Chicano pupilsgenerally speak out less in class than do Anglopupils. In view of the central importance of inter-

"Although the disparity in the composite measure of studenttalk was statistically significant, the disparities in the two indi-vidual measures of student speaking were not, The reason forthis is that the calculation of statistical signkcance is based notonly on the absolute difference, but also on the number ofcases involved and the degree to which these cases vary amongthemselves.

action to learning, it is evident that Chicano pupilsan not receiving the same quality of education inthe classroom as are Anglo pupils.

The remainder of this report focuses on thosedifferences between Mexican American and Anglopupils which, with a substantial degree of cer-

tainty, represent corresponding differences in theentire survey area. The importance of each of thesix statistically significant disparities, and their re-lationship to certain teacher, student, school, andclassroom characteristics, are discussed separatelyin the section that follows.

CHAPTER III. A DISCUSSION OF SIXSIGNIFICANT DISPARITIES

A. Disparities. in Teacher Praise andEncouragement

Praising and encouraging pupils- is one of theimportant ways for a teacher to build confidenceand lead students to participate more in classroomactivities. Although it is possible to overuse praise

and it may not always be beneficial," it is usuallya positive type of teaching technique by whichthe teacher expresses confidence in, and approvalof, a student.

When used judiciously, praise and encourage-ment can -have a positive effect on theself-esteem and on his or her attitude .toward theteacher, the school, and learning in general. As a

result he is more apt to want to contribute to theclassroom discussion and thus become more ac-tively involved in the teaching-learning process..Through praise and encouragement, the child alsois likely to apply himself- more vigorously andperform better in school.

OBSERVATION BY COMMISSION STAFF:"

During a period of oral questioning from theteacher,. one student sat quietly, not volun-teering any answers. Then.the teacher singledhim out for praise, noting that he had an-swered 68 questions correctly on a 69 ques-don written test administered earlier. Follow--ing the teacher's praise, this student beganvolunteering to answer her oral questions.Often, his hand was the first to be raised.

In the Commission's study, teachers make spar-ing use of praise and encouragement generally.'5But the average Anglo received about 36 percentmore praise or encouragement than the averageMexican American pupil in the same classroom,(See Figure 3.) This substantial disparity.. in praisevery adverse effeCts on the motivationand academic performance of Chicano pupils, The

George J, Mouly, Psychology for Effective Teaching, NewYork: Holt, Rinehart & Winston, Inc., 1968, pp. 350-352.

" The observations presented in this report occurred while a

Commission observer was in the classroom. Following the class-room visits, the observers recorded those incidents which theythought to be significant,

35 Only 2.6 percent of all class time observed was spent Onteacher praise of individuals and of the class as a whole,

direct consequence of the disparity is that MexicanAmericans receive less of the educational benefitsof praise and encouragement than do Anglos. Inaddition, the disparity may damage the academicself-esteem and motivation of Chicano pupils be-cause students tend to evaluate themselves on thebasis of the teacher's treatment of them in com-parison to other students.0

A number of possible explanations could begiven for this disparity, but none can justify thispattern of unequal treatment. There is no reasonwhy Mexican American students, as a group,should receive less praise and encouragement

Figure 3

Average Amount of Praise ar EnCouragementGiven by Teachers to Individual. Mexican Ameri-can and Anglo Students

MexicanAmerican

Anglo

'6 Thomas Good and Jere Brophy, Analyzing Ciassroom Inter-action: A More Powo11.11 Alternative, Report Series No. : 26,Austin, Tex.: The Research and Development Center for TeacherEducation, University of Texas, 1969, p, 7,

than Anglo students if the educational programis adequately oriented toward the needs of allstudents:

One explanation that might be given for thedisparity is that Chicanos speak less often thanAnglos and thus make fewer contributions whichthe teaches can praise or encourage. In fact, otherfindings in this report indicate that Chicanos dospeak less often than Anglos in the classroom.However, this is not a justification for the dispar-ity in praise or encouragement for several reasons.First, the Commission also found that the teachersask Mexican Americans fewer questions than Ang-los and thus provide the Mexican Americans withfewer opportunities to speak in class. Second,praise and encouragement are coded by the Fland-ers system as a single category. Students whoseldom speak may provide the teacher with fewopportunities to praise their verbal contributions.However these same students are generally theones who need the most encouragement fromthe teacher. Third, the schools' neglect of thelanguage and cultural background ofi MexicanAmerican students contributes to the reluctanceof some Mexican Ame'ricans to participate verballyin classroom activities.

A large proportion of Chicano pupils enterschool speaking very little English or with seriousdifficulties in using the language." Yet few schoolshave adequate language programs to meet thec,eneeds." This language difficulty prevents or in-hibits many Chicano students from speaking inclass in the elementary grades. In addition, thecontent of the curriculum in most classrooms isnot designed to be relevant to the home back-ground and familiar experiences of.Spanish speak-ing children." Likewise, teachers are not generallytrained to incorporate the unique interests andexperiences of Chicano pupils into the class

8r U.S. Commission on Civil Rights, The Excluded Student,p. 14.

" /bid., pp_ 21.29."For a discussion of methods to develop and implement a

bicultural elementary school curriculum, see Feliciano Riveraand hector Cordova, 'curriculum and Materials for BilingualBicultural Education" in The 14ationa/ Eiernonlary Principal issueon Education for the Spanish Speaking, Vol L, Norther 2,November 1970.

22

r-

room discussion. These practices operate from the,earliest years of school to exclude the Chicanochild from active participation in the educationalprogram. Through this neglect is begun the cycleof low participation, lack of interest, pooriachieve-ment, and low self=esteem which characterizes theeducational experience of many Chicano pupils.

A second poss.ble explanation for the disparityin praise or encouragement is that Chicanos maymake proportionately fewer contributions worthy_of commendation. The Commission's study didnot address itself to the question of whether ornot the contributions of Chicano students are lessdeserving of praise than those of Anglos. Heretoo, however, this explanation, if true, does notjustify the disparity in praise. As noted above, thefailure of the school to adapt to the different lan-guage and cultural heritage of Mexican Americanstudents is a significant contributing factor in thecycle which results in lower participation andachievement on the part of Chicano pupils,Furthermore, praise or encouragement from theteacher is a technique as important in stimulatinghigher performance as in rewarding performancealready achieved. It is the responsibility of teach-ers to provide students with opportunities to suc-ceed rather than fail, and to encourage and rewardstudents for their efforts as well as their "successes.

A third possible explanation is that teachers,because of bias or other personal predilections,may praise contributions of Chicanos less oftenthan similar contributions of Anglos. The Com-mission's data do not provide documentation ofthe extent to which this is true. One source ofbias in a teacher's treatment of students may bedifferential teacher expectations of student per-formance. In the view of a number of teachersobserved by CommisSion staff, Mexican Americanchildren are not expected to perform as well asother children,

OBSERVATION BY COMMISSION STAFF;

One teacher, working in a predominantlyMexican American school complained toCOMIniSSI011 staff of the problem she laced:

am a good teacher, I think. And if i had a

normal hunch of kids 1 could teach. But thiscertainly is not a normal bunch of kids,"

Evidence from other studies suggests that manyteachers actually believe that Chicano pupils arenot as capable of learning as other children.Teachers in one elementary school in a pre-dominantly Mexican American town reported thatshe thought that Chicano students vere basically"dull" or "slow witted.' Another study reportedteachers to have said: "Look, so many SpanishAmerican children have to repeat the first gradetwo to three times. They just can't learn as fast asAnglo American children. If you don't believe mecheck their test scores."' These types of negativeattitudes are undoubtedly reflected in the wayteachers treat Mexican American students in theclassroom. Many hard working teachers fail to seethat their own lower expectations and resultingbehayior are part of the cycle of educationalfailure. Thus, one California junior high schoolteacher stated: "We will keep trying . but thereis nothing yo can do with these kids, they can'tdiscuss, they can't talk, all you can do is give themseatwork to keep them under control."'

Teacher expectations of students are likely toinfluence their, reactions to student behavior.'The relationship of expectations to teacher's praiseof students was demonstrated by one study whichfound that at least some teachers praised a higherpercent of the correct responses of students theyperceived as high achievers than they did theequally correct responses of students they per-ceived as loWer achievers."

Thus, the disparity in praise or encouragement

Theodore Parsons, Jr., "Ethnic Cleavage in a CaliforniaSchool," unpublished PhD, dissertation, Department of Educa-tion, Stanford University, 1965, p. 188.

Clark S. Knowlton, "Bilingualism a Prob:em or Asset," Ad-dress delivered to the meeting of faculty and staff of AnthonySchool District, Dec, 8, 1965, Anthony, N. Mex.: Mimeographed,1965.

°Thomas P. Carter, Mexican Americans in School: A Historyof Ed'cational Neg/ect, New York: College Entrance ExaminationBoard, 1970, p, 114,

For a discussion of the effects of teacher expectations onteacher and student behavior, see R. Rosenthal and L. Jacobsen,Pygmalion in the Classroom: Teacher Expectation and Pupils'Intellectual Development, New York: Holt, Rinehart, and Win-ston, Inc., 1968, ells, 1-4.

given to Anglo and Mexican American pupils can-not be justified as resulting totally from diltcrcncesin the two groups of students. There is evidencethat the disparity results, at least in part, fromattributes of the schools and the teachers:

In an effort to refine further 1110 data on dsparities in praise or encouragement, the Commis-sion compared vat-in0115 in the amount tit dis-parity found with different school, classroom, andteacher characteristics. Such questions were askedas: "Does the disparity increase or decrease withgrade level?" "Do Mexican American teachersshow less disparity in the amounts of praise givento students of each ethnic group than do Angloteachers?" "Is the difference in praise or encour-agement smaller in tracked than in untrackedclassrooms?" "Is the disparity larger or smaller inpredominantly Mexican American schools?" Toanswer questions such as these, a total of 22characteristics of the school, the classroom, andthe teacher were studied for their relationship tothe disparity in praise or encouragement. Thesecharacteristics are listed below:45

School characteristics

1) Mexican American percentage of schoolrollment

2) Anglo percentage of school enrollment3) Degree of ethnic concentration within school

[measured by the variance of the Anglo percentcomposition of the classrooms]4) Average socioeconomic status (SES) of Mexi-

can Americans in the school [principal's estimate]5) Average SES of Anglos in the school [princi-

pal's estimate]

6) Difference in Mexican American and Angloaverage SES

7) Average HS of the school [weighted averageof Mexican American and Anglo SES]8) State in which school is located

Classroom characteristics

t. Thomas Good and Jere Brophy, Teachers' Communicationof Differential Expectations for Children's C/assroorn Perform-ance, Austin: University of Texas R&D Center fur Teacher Edu-cation, 1969.

A detailed description of the calculation of each of thesemeasures is found in Appendix C, which begins on p. 51.

23

9) Grade level of class10) Track level of class [as reported by teacher]"11) Subject matter of course12) Criterion used to seat students [as reportedby the teacher]13) Physical location of Mexican Americans inthe classroom [based on observed seating posi-tions]

14) Physical location of Anglos in the classroom[calculated from observed seating positions]15) Mexican American percentage of enrollmentin the class

16) Anglo percentage of enrollment in the class17) Total number of students in the class

Teacher characteristics18) Extent of teache'r's education

19) Teacher attendance at any inservice trainingsessions related to teaching Mexican Americans[as reported by the teacher]20) Teacher's ethnicity21) Teacher's age22) Teacher's sex

The disparity in praise or encouragement givenMexican Americans and Anglos was found to varysignificantly" among categories of only two char-acteristics: the seating criteria used by the teacherand the ethnicity of the teacher. This means thatthe disparity in praise was larger under some stu-dent seating patterns than others, and that teachersof one ethnicity showed a greater disparity inpraise or encouragement than those of anotherethnicity. The fact that the disparity did not varyacross differences in any of the other character-istics means that the difference in the amount ofpraise given Anglo and Chicano pupils was foundto be of the same magnitude in many types ofschools and classrooms and with different typesof teachers. Thus, the disparity was about equallylarge in predominantly Anglo and predominantly

" Tracking is an educational practice by which students areseparated in classes according to measures of their achievementor -ability", or a combination of both these Thus, the lowachieving or low "ability- students are placed in certain classes,and the high achieving or high "ability" students are placed inother classes.

1' Statistically siitnificant at 0Students are seated in the classroom together with other

students perceived by the teacher to be of the same ability:

24

Mexican American schools, in schools of varyingconcentrations of students from low socioeco-nomic backgrounds, at all grade levels studied,in classrooms with young or old teachers, andso forth.

The disparity in praise or encouragement givenChicano and Anglo students varied according tothe criteria used by teachers to seat students, Inclasses where teachers seated students homogene-ously by ability" there was a significantly greaterdisparity in praise than in classes where studentswere seated by any other method, The findingsin Figure 4 show the average amount of praiseor encouragement given to Mexican American andAnglo students by teachers using five differenttypes of seating arrangements. Teachers who as-signed their students to seats to form homogene-ous ability seating groups exhibited far greaterdisparities in praise toward Mexican American andAnglo students than did teachers who used anyother seating criterion. Thus, teachers who seatedstudents homogeneously gave about as muchpraise to Chicanos as did teachers who used otherseating arrangements, but they gave considerablymore praise to Anglos than did the other teachers.

As illustrated in Figure 4, teachers using thehomogeneous seating criterion give Anglos morethan four times the amount of praise or encour-agement they give Mexican Americans,- The Com-mission's data indicate that the use of this seatingmethod is much more prevalent in the early gradesthan in the later grades." It is the early years ofschooling which most influence a student's atti-tudes toward school and his patterns of academicachievement. Therefore, the practice is mostwidely used when it has the potential for thegreatest harm,

The Commission's data offer no direct explana-tion for the increased disparity in praise or en-couragement associated with the homogeneousseating pattern. However, the explanation may insome way be related to the fact that homagene-

" The percent of all classes of each grade lq'vel using thehomogeneous seatinK method were as follows: 1&percent of allfourth grade classrooms, 2 percent of the eighth grade class-rooms, 1 percent of the 10th grade classrooms, and none ofthe 12th grade classrooms.

Figure 4

Average Amount of Praise or Encouragement Given Per Pupil te 'Individual Mexican Americanand Anglo Students by Teachers Using the Specified Criterion for Assignments of Seats

StudentEthnicity

CriterionFor As-

signmentOf Seating

No.

Classrooms

.559

.179.198

.183

.161

.121 .125

.179

.115 117

MA A MA

TeacherChoice

72

Alpha-betical

53

MA A

Homoge-neous

By

Ability

13

MA

StudentChoice

244

MA A

Student ChoiceWith TeacherModification

47

ous seating is likely to result in a greater degreeof physical separation of Chicanos and Angloswithin a classroom. The reason for this is thatstudent ability is usually judged by achievement.At present, in our school system, Mexican Ameri-.can students are disproportionately representedamong the lower-achieving students: Therefore,

this seating criterion tends to separate MexicanAmerican and Anglo_students physically within theclassroom. The physical separation may in someway accentuate the disparity in praise Or encour-agement which, under any seating arrangement,consistently favors Anglo students.

It should be noted that only 3 percent of the

25

classes sampled used homogeneous ability as theseating criterion. While the disparity in praise orencouragement was much larger when this seat-ing criterion was used, there was some disparityunder the other types of seating criteria as well.Therefore, homogeneous ability seating alone can-not account for the overall disparity. Nonetheless,the magnitude of the disparity raises serious ques-tions concerning the use of this seating method.

The disparity in the amount of praise given to

Mexican .American and Anglo students was alsofound to vary significantly with the ethnicity ofthe teacher. Figure 5 shows that Mexican Americanand Anglo teachers give similar amounts of praiseor encouragement to Chicano pupils. However,Mexican American teachers praise Anglo pupilsconsiderably more than their Anglo colleagues.This results in a larger disparity in praise or en-couragement from the Mexican American teachersin favor of Anglo students.

Figure 5

Average Amount of Per Pupil Praise or Encouragement Given to Individual Mexican American andAnglo Students by Mexican American and Anglo Teachers

CIJ

J

26

.136 .138

.172

Student MA A MA AEthnicity

Teacher MexicanEthnicity American Anglo

No. OfTeachers 36 386

27

it is rapt known precisely why Mexican Ameri-can teaci,rrs give more praise to Anglo studentsthan do At Wo teat ers: 7i -5 nornenon maybe the result of a combination of factors involv.ing characteristics of the Mexican American teach-ers, the children they teach, and the schools inwhich they teach.

There is a tendency for Mexican Afnericanteachers to be found in classrooms with a highproportion of Chicano pupils and in schools withhigh Chicano enrollments. Half of all Chicanoteachers who were observed taught classes in

which 60 percent or more of the students wereMexican American and most taught in predomi-nantly Mexican American schools.5° Thus, in com-parison to Anglo teachers, Chicano teachers gen-erally teach where there are fewer Anglo pupils.This may in some way be related to the greateramounts of praise they give to the average Angloin their classes,

One possible interpretation is that MexicanAmerican teachers may tend to use the relativelyfew Anglos in their classrooms to emphasize themiddle class Anglo culture and values to the Chi-cano pupils. it is possible that, to a large extent,many Mexican American teachers operate underthe philosophy that success for Chicano pupilslies in acquiring Anglo traits." If this is the case,they would be likely to give more praise and en-couragement to the average Anglo pupil than tothe average Chicano pupil in their classroom. Thefact that Mexican American teachers generallyhave relatively few Anglos in their classroomS.would tend to accentuate the difference in theamount of praise per pupil they give to Anglo andChicano students.

This same phenomenon could be occurring on

Seven of every lit Chicano teachers taught in schools wherethe enrollment was 90 percent or more Mexican American. Inaddition, most of the Mexican Anierican teachers taught inschools in which the ethnic composition of the individual class -rooms varied a groat deal front the ethnic composition of thetotal school. The average Mexican American leacher observedby Commission staff taught in a predominantly Mexican Amer',can !AMA!, in a classroom which had:a consitlerahly higherproporlion of Mexican Americans than the school as a whole.

' This phenomenon has been found to be characteristic ofsome minority teachers in particular older teachers trained inmore traditional methods.

a more subconscious level not directly related toa desire to help the Chicano pupils, Asa resultof having gone through an educational 5ystemdominated by the Anglo culture and working ina school system lirected by Anglo administrators,it is possible that many Mexican American teachersseek to identify with the culture and values of thedominant society, This also could be reflected ingiving greater amounts of approval to 1110 Anglostudents in their classrooms,

It should be noted that Mexican Americanteachers made up only 9 percent of the totalteachers observed and that Anglo, as well as Chi,cano teachers, gave somewhat more praise ofencouragement to Anglo pupils than to Chicanopupils. Therefore, the finding that the disparity inpraise is greater among Mexican American teach-ers by no means accounts for the overall d isparityin praise.

Whatever the specific explanation for the in-creased amount of praise or encouragement givento Anglo pupils, this finding suggests that MexicanAmerican teachers, as well as Anglo teachers, needtraining in dealing with the two groups of stu-dents, This is an important area of concern forteacher training programs, as well as inservicetraining for experienced teachers.

B. Acceptance and Use of Student Ideas

One of the most effective means of encourag-ing students to participate in classroom discussionand of motivating them to learn is through the useof behavior characterized as the "acceptance anduse of student ideas." When a teacher makes 5tatements acknowledging or incorporating a student'scontribution, the teacher is both commending thepupil for the idea and informing the class that thestudent's contribution is worth listening to. This isan important way of giving positive reinforcernentto the studentmore effective, for example, thanmerely expressing approval without repeating orreferring to the points which were made 2 In

repetition or rephrasing, the teacher is in effectsaying that the student's contribution is sufficientlyworthwhile to be taken note of by the entire class,

62 Rosenshine, op. cit., (r) 71,

while at the same time indicating to the studentthat. she took the time to listen and try to under-stand what he was saying.

Further, teacher discourse using student contri-butions generally reflects the extent to which theteacher is drawing upon the interests and experi-ences of the child to interest him in the contentmaterial, One of the most important tasks inteaching is to help the student see the relevance ofwhat he is learning. As one educator has pointedout, the key to motivating the child lies in bridgingthe gap between those things he should learn andthose things he wants to learn." One way ofbridging this gap is to encourage students to ex-press their interests and ideas in the classroomand then to build on these col itributions by relat-ing them to the content to be learned.

Another reason why the use of student ideas isbeneficial to learning lies in the very repetition ofa worthwhile idea or piece of information. When-ever a teacher restates or summarizes a corloctanswer or interpretation given by one of the stu-dents it is potentially beneficial for the learningand recall of all the students in the class.

The importance of the acceptance and use ofstudent ideas to learning has been documentedby research. This research has shown that the moreteachers make use of this behavior, the better theachievement and attitudes of those students. Arecent review of the research on teacher effective-ness reached the following conclusion:

it can now be stated with fairly high con-fidence that the percentage of teacher state-ments that make use of ideas and opinionspreviously expressed by pupils is directly re-lated to average class scores on attitude scalesof teacher attractiveness, liking the class, etc.,as well as to average achievement scores ad-justed for initial ability.54

in fact, the evidence from the research to dateindicates that of all types of teacher approval,none is as strongly and consistently related tohigher pupil achievement than the use of student

mouly, op. cit., pp. 333-41" Ned A. Flanders, "Teacher Effectiveness," Encyclopedia of

Educational Research, 1969.

ideas,"In view of the importance of this type of teacher

behavior, the Commission's finding on the dispar-ity in the teacher's use of the contributions ofAnglo and Chicano pupils is a disturbing one. Theaverage Anglo pupil in the survey area hears theteacher repeat, or refer to, an idea he or she hasexpressed about 40 percent more than does theaverage Chicano pupil (See Figure 6). This dis-parity is potentially even more damaging to theChicano pupils than is the disparity in praise orencouragement, for it reflects a disparity both inteacher approval and in the extent to which thestudent's own interests and experiences are

Figure 6

Average Amount of Acceptance and Use of Stu-dent Ideas Per Pupil Given by Teachers to In-dividual Mexican American and Anglo Students

.3

MexicanAmerican

Rosenshine, op, cit., p. 71.

Anglo

deemed worthy of incorporation into the class-room discussion. The Chicano student's percep-tion of the worth of his own ideas and experi-ences, in comparison to those of Anglo students,necessarily is influenced by the teacher's use ofthe contributions of members of each group. Thedisparity in favor of the Anglos is likely to resultin lower self-esteem, decreased interest, andpoorer academic performance for the MexicanAmerican pupils.

The possible explanations for this disparity inthe acceptance and use of student ideas are essen-tially the same as the three previously discussedin reference to the disparities in praise. it may bethat Chicanos are speaking less in the classroomand thus making fewer contributions which theteacher can accept or use. Secondly, Chicanosmay make proportionately fewer contributionswhich are considered sufficiently worthwhile forthe teacher to use, Finally, teachers may expressacceptance or use of contributions by Chicanosess often than similar ones by Anglos. As dis-

cussed in the previous section, these may, in part,explain the disparity, but they do not justify it.

As with the disparity in praise, the Commissionsought to determine if disparity in the teacher'suse of student ideas varied by characteristics ofschool, classroom, and teacher. It was foundthat the disparity in the use of student's ideasvaried significantly arnong categories of only oneof the 22 conditions: the degree of ethnic con-centration within a school, as measured by thevariation in the ethnic composition of classroomsin the school." The fact that the disparity did notvary across differences in any of the other 22 char-acteristics means that the f.fisparity was equallylarge for all the other different types of schools,classrooms, and teachers studied.

Figure 7 illustrates the extent of disparity in theuse of student ideas found in three different typesof schools: (1) low ethnic concentration within a

" See pp. 23 & 24 for a list of characteristics.The degree of ethnic concentration within a school was

measured by an index of variance in class percent Ang!o. Thisindex measures the extent to which the percentage of Ang!opupils in each of the classrooms visited varies from the averagepercent Anglo of the total classrooms. visited in the school.See Appendix C, p. 51.

30

school, where Chicano and Anglo pupils appearedto be evenly distributed among the classrooms-,"(2) medium eth-lic concentration, where there wasa tendency for Mexican American pupils to befound in some classrooms more than others; and(3) high ethnic concentration, where MexicanAmerican .and Anglo pupils were definitely sep-arated in different classrooms." In all three typesof schools, -..eachers used the ideas expressed byAnglos more than they used those expressed byChicanos. However, in schools with low ethnicconcentration, the teachers exhibited a signifi-cantly larger disparity in the amount of acceptanceand use of student ideas than in schools wherethere was a moderate degree of concentration."

in schools with a low degree of ethnic concen-tration, the average Anglo heard the teacher ac--cept or repeat a contribution he had expressednearly 21/2 times more often than did the averageChicano pupil, This is an extremely large disparityon such an important type of teacher behavior:In addition, it affects a large proportion of Mexi-can American pupils, since over one-fifth of allthe classrooms visited were in schools with lowethnic concentration.

The results indicate that a sizable part of theoverall disparity in the acceptance and use ofstudents' ideas is accounted for by the disparitythat exists in the schools with low ethnic con-centration, This is because these classrooms ac-counted for a sizable proportion of the total andbecause the magnitude of the disparity found in

sA Although the index of ethnic concentration was calculatedonly from those classrooms visited in a school, it is assumedto be reasonably representative of the relative degree of ethnicisu!ation in the whole school. See Appendix

" The term "ethnic concentration" is not to be confused withthe term "ethnic composition". The ethnic composition of aschool refers to the percentages of the total school enrollmentcomprised by each ethnic group in that school. The level ofethnic concentration is a measure of the extent to which theethnic composition of each clasi§,in the school is similar tothe ethnic composition of the whole school. For example,schools with a low percent composition of Mexican Americanstudents can still have a high level of ethnic concentration ifall or most of the Chicano students are found in a few class=rooms.

The disparity in the low ethnic concentration schools wasalso substantially larger than that in the high ethnic concentra-tion schools, but the difference was not statistically significant.

Figure 7

Average Amount of Acceptance and Use of Student Ideas Per Pupil Given to IndividualMexican Americ.an and Anglo Students by Teachers in Schools

with Various Degrees of Ethnic Concentration

GJ

Student MA A MA A MAEthnicity

Degree OfEthnic Con-centrationWithin TheSchool *

No. OfClassrooms

Low

96

Medium High

163 174

* The degree of ethnic concentration within the school represents the extent to which there is variation in the ethniccomposition of classrooms within a school. In low ethnic concentration schools, Mexican American students appear to bedistributed evenly among the classrooms. In medium ethnic concentration schools, there is a tendency for Mexican Ameri-cans to be found in some classrooms more than others. In high ethnic concentration schools, Mexican Americans are definitelyisolated in specific classrooms. For a discussion of the specific cutoff points used to differentiate low, medium, and high, seeFigure 17 in Appendix C.

the other classrooms was relatively small.'The schools with a low level of ethnic concen-

tration are schools which have little or no ethnicsegregation of students by classes within theschool. The finding that these schools exhibiteda large disparity between Anglo and Chicanopupils in the teacher's acceptance and use of stu-dents' ideas should not be interpreted to meanthat this is a necessary result of mixing studentsof different ethnic groups in the classroom. Rather,it shows that for integration to work, schools mustdo more than mix majority and minority studentsin the same classrooms. Teachers must be trainedto interact equitably with all students in integratedclassroom situations and schools must adapt theircurriculum to the varying backgrounds of thestudents. The Commission's finding indicates thatthese steps have not been adequately carried outin integrated schools.

C. Positive Teacher Response

The Category "Positive Teacher Response"represents a composite of all teacher behaviorswhich express approval, support, or acceptance ofstudent behavior. In this study it is measured bythe sum of three types of teacher behavior:

The Commission attempted to determine possible reasonswhy this disparity was accentuated in schools of low ethnicconcentration. One possible reason that was explored was therelationship of this disparity to "tracking", an educational prac-tice by which students are separated in classes according tomeasures of their achievement or "ability", or a combinationof both of these. The data show that tracking occurred muchless frequently in schools of low ethnic concentration, as com-pared to other schools. Eighty-two percent of the classes inschools with low ethnic concentration were reported to beuntracked whereas 45 percent and 29 percent of the classes inschools with medium and high levels of ethnic concentration:respectively, were reported to be untracked. On the basis ofthis association, it was hypothesized that the disparity mightoccur most frequently in untracked classrooms where there is

a wide range of academic performance among students: How-ever, this was not confirmed by the Commission's data. Dis-parities in the acceptance and use of student ideas betweenMexican Americans and Antos did not vary significantly be-tween tracked and untracked classrooms. Although the differ-ences were in the expected direction, the calculated F statisticwas less than one-tenth of the value needed in order to bestatistically significant. The magnitude of the disparity in accept-ance and use of ideas of Mee can American and Anglo studentswas :079 in untracked classes, .050 in low track classes, .055 inthose of medium track, and .022 in that of high track.

acceptance of student feelings, praise and en-couragement, and acceptance or use of studentideas. This combination measure reflects theoverall degree of teacher warmth, approval, andencouragement, and, as such, is an indicator ofthe emotional tone of teacher-student inter-actions. The greater the amount of positive re-sponse from the teacher, the more the student islikely to feel that he is capable of achieving andthat his contributions are worthwhile.

The importance of this measure in the teaching-learning process lies in the significance of eachof the teacher behaviors of which it is comprised.As discussed in previous sections, praise and en-couragement, and the acceptance or use of stu-dent ideas are both key types of behavior whichhave a significant impact on student attitudes andachievement, Teacher acceptance of student feel-ings also is important because it indicates howmuch the teacher allows or encourages individualexpression of emotion in the classroom setting.Although the Commission data show that teachersseldom exhibit this behavior, even an occasionalexnression of acceptance or feeling is likely toaffect the emotional climate of the classroom.

The Commission's findings regarding positiveteacher response to Anglo and Mexican Americanstudents show that the average Anglo pupil re-ceives about 40 percent more positive responsefrom the teacher than does the average Chicanopupil. (See Figure 8). This is one of the largestdisparities found in this study.

This disparity essentially reflects the large dis-parities in teacher praise or encouragement andteacher acceptance of students' ideas." Con-sequently, it raises concerns similar to thoseraised by the disparities in those behaviors. inaddition, because positive teacher response rep-resents overall warmth and approval, this disparityis also indicative of differenCes in the emotionaltone of teacher relationships with Anglo andChicano pupils.

That is, 'categories 1-3 of the Flanders Interaction AnalysisSystem.

h' See Fig. 2, p. 17, supra.6' This is because the disparity in acceptance of student feel=

ings occurred very infrequently and was found not to be statis-tically significant.

The Commission sought to determine if theinequality in positive teacher response variedamong differences in any of the 22 characteristicsof the school, the classroom, and the teacher. Itwas found that the magnitude of the disparityvaried significantly among categories of only twoof the 22 conditions: the classroom seating criteriaused and the degree of ethnic concentration, asmeasured by the variation in the ethnic composi-tion of classrooms within a school. The disparitywas equally large across differences of the remain-ing 20 categories.

The manner in which the disparity in positiveteacher response varied across different methodsof seating was very similar to the pattern foundfor the disparity in praise or encouragement."

Figure 8

Average Amount of Positive Response Per PupilGiven By Teachers to Individual Mexican Ameri-can and Anglo Students

.296

Al 3

MexicanAmerican

34

Anglo

As illustrated in Figure 9, teachers who seat stu-dents by ability exhibit a much greater disparityin giving positive response to Anglo and Chicanopupils than do teachers who use any other criteriafor seating. These teachers give Anglos more thanthree times as much positive response than theygive Chicanos.

It is not possible to explain this association fromthe data collected. However, as noted earlier, itis likely that in classes where the homogeneousseating criterion is used, Chicanos are morephysically isolated from Anglos than in classeswhere other seating criteria are used. This physi-cal separation may accentuate the disparity inpositive teacher response between the two ethnicgroups.

The disparity in positive response was alsofound to differ among schools of varying levelsof ethnic concentration within the school. Again,the pattern of variation was very similar to thatexhibited for disparities in the use of studentideas." As Figure 10 illustrates, the difference inthe amount of positive response given to Chicanoand Anglo students is significantly greater in lowethnic concentration schools than in those witha moderate amount of ethnic concentrationacross classrooms. The greater inequity in positiveteacher response to Chicano pupils found in theschools with the least amount of segregationwithin the schciol highlights the need for teachertraining and curriculum adaptation in integratedsettings.

D. Teacher QuestioningThe relative amount of teacher questioning" is

"The pattern found for the disparity in praise across seatingmethods was also found for the disparity in the acceptance ofstudent ideas, although the latter was not statistically significant.The pattern found for disparity in positive feedback results fromthe association of seating criteria with disparities in both of thebehaviors which are the two main components of positive feed-back.

'See fig, 7, p. 31, supra.The manner in which this teacher behavior was recorded

does not differentiate between questions directed at studentswho were volunteering to speak and those directed at studentswho were not volunteering. Regardless of whether the teacherwas requesting a student's response or permitting a .student'svolunteered response, the. amount of questioning is a measure©f the extent to which teachers directly solicited contributions,

Figure 9

Average Amount of Positive .Response Per Pupil Given to Indiviual Mexican American and AngloStudents by Teachers Using the Specified Criterion for Assignment of Seats

1.0

0

1.062

.422

315

7

30 .284

.391

.284

.380

StudentEthnicity

CriterionFor

AssignmentOf Seating

No, OfClassrooms

MA MA A MA MA A MA A

Teacher Alpha- Homo- Student Student ChoiceChoice betical geneous Choice With Teacher

Modification

72 13 244 47

35

Figure 10

Average Amount of Positive Response Per Pupil Given to IndividualMexican American and Anglo Students by Teachers in Schools of Different Degrees

Ethnic Concentration Within the Schools.*

U)1.6

Ce

U)

tU

4'

0

aJti

.569

50 .362.374

Student MA A MA A MA AEthnicity

Degree OfEthnic Con-centrationWithin TheSchool

Low Medium High

No. Of 96 163Classrooms

174

The degree of ethnic concentration withir the school represents the extent to which there is variation in the elhni^:composition of classrooms within a school. In low ethnic concentration schools, Mexiran American students appear idistributed evenly among the classrooms: in medium ethnic concentration schools, there is a tendency for Mexican Amur-cans to be found in some classrooms more than others. In high ethnic concentration schools, Mexican Americans are definitelyisolated in specific classrooms. For a discussion of the specific cutoff points used to differentiate low, medium, and high, seeAppendix C.

Figure 11

Average Amount of Teacher Questioning Per Pupilof Individual Mexican American and AngloStudents

:7

MexicanAmerican

important because it indicates the extent to whichstudents of each ethnic group are asked orallowed to speak by the teacher. The value ofteacher questioning lies in the basic importanceof student involvement in the teaching-learningprocess. If the teacher is to. make the contentmaterial relevant to the class, he or she must en-courage students to express their own interestsand experiences in the classroom situation. In ad-dition, if students are to be more than passiverecipients of the information given by the teacher,they must be encouraged to think independentlyand express their insights and conclusions in class-room discussion. Teacher questioning is also im-portant in providing students with the practiceneeded in certain phases of the learning process.Furthermore, student participation prov'cles ameans by which the teacher is informed of thestudent's progress in such a way that he canprovide immediate direction and reinforcement.

Not all teacher questioning is necessarily bene-ficial to learning. It is possible fo questioning tobe overused, and different types of questions aremore appropriate to different learning situations.Nonetheless, questioning is generally considereda positive type of teacher behavior. Researchconducted to date indicates that, at least in theprimary grades, the frequency of asking questionsis associated with increased pupil achievement.°8

The amount of teacher questioning is particu-larly significant with regard to those studentswho, for one reason or another, are reluctantto speak in class. Instead of ignoring these stu-dents in classroom interaction, the teachers canattempt to draw them out through the use ofquestions aimed at their special interests or per-formance levels.

According to the Commission's data, the aver-age Anglo pupil in the survey area receives about21 percent more questioning from the teacherthan the average Chicano pupil. (See Figure 11).This difference shows the extent to which Angloshave more direct opportunity to speak out in classthan Chicanos.

One possible explanation for the fact that

Anglo6, Rosenshine, op, cit.,

37

Mexican American pupils are questioned less fre-quently than Anglos may be that they raise theirhands less often to volunteer an answer or makea comment, Even so, this would not justify thedisparity in questioning. Teachers are responsiblefor encouraging all students to participate, notmerely those who are eager to do so. The surestway to encourage students to participate is toask them questions. For example, students canbe questioned on content material with whichthey are familiar and in which they are interestedand, therefore, can have a fair chance of successand satisfaction. Thus the teacher, calling onstudents reluctant to assert themselves, can givethem the opportunity to succeed and encouragetheir more active class participation,

OBSERVATION BY COMMISSION STAFF:

The teacher called on a Mexican Americanboy and, at first, the student hesitated aboutreading his story before the class. .The en-tire class urged him on they seemed toknow that he would have an interestingcomposition. He stood up and read bothcompositions while the entire class roaredwith laughter, He had composed two veryhumorous situations. The teacher did notneed to reinforce very muchthe class hadalready done it for her. By merely callingon the student, she had given him an op-portunity that he felt ((good" about.

By the same token, the teacher, by failing tocall on a student, can dampen his enthusiasmand lessen his class participation even more,

OBSERVATION BY COMMISSION STAFF;

There were several Chicanos who kept rais-ing their hands eagerly at every question.Mrs. C. would repeatedly look right overtheir heads, and called on some of the sameAngle students over and over. 'n some casesshe would call on the Chicanos only becausethe Anglos stopped raising their hands. Aftera while the Mexican American childrenstopped raising their hands.

The disparity in teacher questioning of Anglo

38

and Mexican American pupils is the most directevidence of the failure of the schools to involveChicano pupils adequately in the classroominteraction process. It is possible that the dis-parities in praise, use of student ideas, and over-all positive teacher response are partly a resultof the fact that Chicano pupils are less verballyactive in class. However, at least part of the reasonwhy they speak out less can be attributed to thefact that teachers ask fewer questions of them.

The disparity in teacher questioning did notvary significantly across differences in any of the22 characteristics of schools, classrooms, orteachers. This means that the disparity in teacherquestioning was found to be equally large in alltypes of classrooms observed, in all types ofschools, and regardless of the background of theteacher studied. Thus, the disparity was essen-tially the same in predominantly Mexican Ameri-can or predominantly Anglo schools, in trackedor untracked classes, in classes where the teacherwas old or young, and so forth.

E. All Noncriticizing Talk By The Teacher

The amount of time a teacher spends talkingto individual students is indicative of the overallindividual attention he or she gives to them,both positive and negative. When criticism is

excluded from the sum of teacher talk, the re-sulting measure represents all teacher attentiondirected to individual students which is not dis-approving or critical. In the Commission's studyall noncriticizing teacher talk is a combinationmeasure of positive response, questioning, lectur-ing and direction giving. Although these be-haviors, taken individually, may not always proveto be beneficial to the student, when taken to-gether they are an indication of the relativeconcern a teacher shows toward individualstudents.

It is possible for a teacher to spend too muchtime talking to individual students, just as it is

possible for a teacher" to spend too much timetalking to the whole class. Either of these canbe detrimental if a teacher spends so much timetalking that the students have little opportunity

to speak. However, in the Commission's study,the amount of teacher talk directed to individualsdid not appear to discourage student speaking.On the contrary, the amount the teacher spoketo individual students was associated with morestudent speaking rather than less. Teachers talkedmore to Anglo pupils than to Chicano pupils andAnglos spoke more in class than did MexicanAmericans, Thus the quantity of noncriticizingtalk by the teacher was positively related to stu-dent participation.

Figure 12 illustrates the comparative amountsof all noncriticizing teacher talk received by theaverage Anglo and the average Mexican Americanstudent. Teachers spend 23 percent more timein all nondisapproving talk with Anglo than withChicano pupils. The amount of this disparity re-flects differences in attention and concern showntoward the two groups of students by teachersin the survey area.

For the average Chicano child sitting in a class-room in the Southwest and experiencing this dis-parity in treatment, it is likely to mean, not onlythat he or she does not receive needed assistancearid encouragement, but also may lead him toconclude that, somehow, his education is not asimportant as that of the Anglo pupils in his class-room. If in comparison to Anglo pupils, theteacher neglects the Chicano pupils, the message

.conveyed to the Chicanos is likely to be that theyare not as important as their more privilegedclassmates.

OBSERVATION BY COMMISSION STAFF:

Mrs, M. was leading a class discussion onunions, but all the interaction was betweenthe teacher and three Anglos sitting in thefront of the class. They were very eager, butthe rest of the class was bored. Mrs. M.finally said: "The same hands, I always seethe same hands,"

The disparity in all noncriticizing teacher talkto Chicano and Anglo students did not varysignificantly across differences in any of the 22characteristics of the school, the teacher, or theclassroom. This means that approximately similar

.disparities were found among all those types ofclassrooms, schools, and teachers studied.

F. All Student Speaking

The amount of student speaking in a classroomis one impOrtant indicator of the extent to whichpupils are being actively involved in the teaching,.learning -process,- Student participation is impor-tant in the learning process for numerous reasons.Students must be able to express their interests

Figure 12

Average Amount of All Noncriticizing TeacherTalk Per Pupil Given to individual MexicanAmerican and Anglo Students

1,5

1:0

1.551

1.901

MexicanAmerican Anglo

39

-

p.

. -

and discuss their experiences at certain points inclassroom discuSsion so that the teacher canutilize these familiar elements to build motiva-tion. To encourage independent thinkinr studentsmust be given the opportunity to express theirown ideas and conclusions. At certain points inthe learning process verbal practice and revieware useful in aiding recall. Student participationalso serves to inform the teacher of the student'sprogress in a situation where na can provide im-mediate direction and feedback.

In the classrooms in the survey area, MexicanAmerican pupils speak significantly less often inthe classroom than do Anglos. (Figure 13) Theaverage Ang/o student spends about 27 percentmore time speaking in the classroom than theaverage Chicano student. This finding demon-strates that Mexican American students are muchless actively involved in the teaching-learningprocess than are Anglo.pupils.

The disparity in participation of Anglos andChicanos results from the interrelationship ofmany factors, some in the student's background,some in the characteristics and behavior of theteachers, and some in the educational systemitself. A detailed discussion of all of these factorsis beyond the scope of this report. However, otherfindings presented in this report give evidencethat the differential behavior of teachers towardstudents is at least a part of the reason Chicanopupils do not participate as much as Anglos.

Teachers ask significantly fewer questions ofMexican American students and thus give themfewer direct opportunities to participate. Thedisparities found in praise or encouragement, inacceptance of ideas, and in overall attention alsoare likely to have an influence on the amount ofspeaking done by the two groups of students. IfChicano students are less often made to feelthat what they have to say is worthwhile or thatit merits the teacher's attention, then this, inturn, will make them want to participate less.

OBSERVATION BY COMMISSION STAFF:

One Chicano sat toward the back in a cornerand volunteered several answers. At one

point the teacher did not even acknowledge,much less reinforce, his answer. At anothertime he volunteered an answer which wasperfectly suitable. Yet the teacher stated:

yes, uh huh, but can anyone elseput it in different terms?" The teacher thencaned on an Anglo boy who gave the samebasic response with very little paraphrasing.

Figure 13

Average Amount of Speaking Per Pupil by Mexi-can American and Anglo Students

2.0

1.5

MexicanAmerican

Anglo

The teacher then beamed and exclaimed:"Yes, that's it exactly."

In order to determine whether the disparityin the amounts of speaking done by Chicano andAnglo students was any more prevalent. underdifferent conditions, the 22 characteristics of theschool, the, classroom, and the teacher59 werestudied for their -effect on the disparity. It wasfound that . there was no significant differencein the amount of the among all cate-gories of the 22 conditions: Thus, the averageChicano student speaks . much less than- theaverage Anglo student, regardless of the -type ofschool or classroom, or the characteristics ofthe teacher.

"See pp. 23 and 24 or a list of characteristics.

41

CHAPTER IV: SUMMARY AND CONCLUSIONS

The basic finding of this report is that theschools of the Southwest are failing to involveMexican American children as active participantsin the classroom to the same extent as Anglochildren. On most of the measures of verbal inter-action between teacher and student, there aregross disparities in favor of Anglos.

Thus teachers praise or encourage Anglo chil-dren 36 percent more often than Mexican Ameri-cans. They use or build upon the contributionsof Anglo pupils fully 40 percent more frequentlythan those of Chicano pupils. Combining all typesof approving or accepting teacher behavior, theteachers respond positively to Anglos about 40percent more than they do to Chicano students.Teachers also direct questions to Anglo students21 percent more often then they direct them toMexican Americans. In addition, Mexican Ameri-can pupils receive significantly less overall atten-tion from the teacher, measured by the extent towhich teachers address their students in a non-critical way. In light of these findings, it is notsurprising to have also found that Mexican Ameri-can children participate less in class than do An-gips; they speak less frequently both in responseto the teacher and on theieqpwn initiative. Thetotal picture that emerges from this study of class-room interaction is one in which Mexican Ameri-can students are ignored compared to their Anglocounterparts.

The classroom is the setting in which a child'sschooling takes place and the interaction betweenteacher and students is the heart of the educa-tional process. The importance of certain typesof interaction for student learning has been docu-mented in previous studies. It has also been ex-plained how all elements of this interaction, takentogether, create a climate of learning which di-rectly affects educational opportunity. Conse-quently, the discovered disparities in teacher be-havior toward Mexican Americans and Anglos arelikely to hinder seriously the educational oppor-tunities and achievement of Chicano pupils. Thesefindings raise disturbing questions concerning theability of our schools to meet the educationalneeds of all students adequately.

Some ,would argue that the schools and teach-

ers are not responsible for these disparities inteachers' behavior toward Mexican American andAnglo students. They would argue that these dis-parities are a result of characteristics of Chicanopupils, such as differences in language and culture,attitudes toward school, and academic achievement levels.

As a group, Chicano pupils do differ from Anglopupils in language, culture, and economic back-ground. A large, proportion of Chicano pupilsenter school speaking very little English or withserious difficulties in using the language." In ad-dition, the culture, values, and familiar experi-ences of Chicano students often differ substan-tially from those of Anglo students and those onwhich the school program is based. The differ-ences between the background characteristics ofChicano students and the language. and cultureof the schools are major obstacles to the educa-tional progress of Chicano pupils. These discrep-ancies between the school and the home are oneof the main causes of the lower participation andachievement levels of Chicano pupils in school.

The differences in language and culture maypartly explain but cannot justify the disparities inclasSroom interaction documented in this report.It is the responsibility Of the school and theteacher to accept the child as he comes to schooland to orient the program to his cultural andlinguistic needs, This, the schools of the Southwest have failed to do.

Only a very small percentage of schools in theSouthwest have implemented language programsto remedy the English language deficiencies ofMexican American students: The content of thecurriculum in most Classrooms is designed to berelevant almost exclusively to the middle classchild of the dominant society. The textbooks andsource materials rarely make use of the skills andexperiences which are familiar to children ofSpanish speaking backgrounds. Similarly, teachersare seldom trained to incorporate the interests

'According to principals' estimates in the Commission's 1969survey of schools and districts, 47 percent of Mexican Americanfirst graders do not speak English as well as the average Anglofirst grader. See U.S. Commission on Civil Rights, The ExcludedStudent, op. cit., p.147

43

and experiences of Chicano children into class-room discussions. In effect, the language and cul-tural background of Mexican American studentsis virtually excluded from the school programs inthe Southwest!'

This exclusion takes its toll on the attitudes andachievement of Chicano pupils. Without the bene-fit of adequate language programs, many MexicanAmericans fall behind academically in the earlyschool years and are never able to catch up. Theomission of their culture, values, and familiar ex-periences from the design of the educational pro-gram causes many Mexican American pupils tofeel that the school is an alien environment withlittle relevance to them. These early school ex-periences of Chicanos thus set in motion the cycleof lowered interest, decreased participation, pooracademic performance, and lowered self-esteemwhich is so difficult to break in the later schoolyears. The schools bear major responsibility forthis cycle of educational failure.

The failure of many schools in the Southwestto create settings conducive to the education ofMexican Americans invariably makes the teacher'sjob more difficult. However, there is still muchthat -the teacher can do to encourage and helpthe Chicano student. The teacher can demon-

71 MS. Commission on Civil Rights, The Excluded Student, op.

strate respect for the Chicano student by incor-porating the culture and personal experiences ofChicano pup_ iis into the classroom lessons anddiscussions. The teacher can encourage the stu-dent's participation by accepting and buildingupon his contributions and can try to providehim with the help needed to keep up with theacademic material. However, the disparities inteacher behavior toward Anglo and Chicanopupils documented in this report indicate thatChicanos are not receiving the benefits of thesetypes of teacher instruction in the classroom. In-stead, the pattern of teacher-student interactiononly mirrors the educational neglect of MegicanAmerican students found throughout the educa-tional system.

It is the schools and teachers of the Southwest,not the children, who are failing. They are failingin meeting their most basic responsibilitythat ofproviding each child the opportunity to gain themaximum benefit of education and develop hiscapabilities to the fullest extent. In the Cornrnis-sion's view, the schools of the Southwest willcontinue to fail until fundamental changes aremade. Changes are needed in the way teachersare trained and in the standards by which they arejudged, and changes are needed in educationalprograms and curriculums so that all children raybe reached.

APPENDIX A.

SAMPLING PROCEDURE

The Commission limited data collection to class-rooms in selected regions of the three SouthwestStates with the largest number of Mexican Ameri-cans: California, New Mexico, and Texas. Geo-graphical areas within each of these States wereselected that included rural, urban, and suburbanschools in which large numbers of Mexican Amer-ican students were enrolled. The areas selectedwere: 1) California; Santa Clara County includingthe city of San Jose 2) Texas: the metropolitanareas of San Antonio and Corpus Christi, the areabetween these two population centers, and thearea 30 miles south of Corpus Christi 3) NewMexico: the Albuquerue area and the south cen-

tral part of the State near El Paso, Texas. In all,there were 968 schools in these selected areas.

The sampling base included all of the schoolsin these areas which met the following six criteria;(1) The schools were located, as of 1968, in adistrict with more than 300 students. This wasnecessary because ethnic data used in the secondcriteria were not available for districts with lessthan 300 students.72 (2) The schools were located,as of 1968, in a district with at least 10 percentMexican American enrollment; this helped to

The source used was the rail 19G8 Elementary and SecondarySchool survey ref the Department of Health, Education, andWelfare

focus on schools where there were substantialnumbers of Mexican Americans. (3) The schnnlshoused at :east one of the grades upon whichthe Commission chose to focus its observations(fourth, eighth, 10th; or 12th). (4) The schoolshad an enrollment of at least 700. if they housed10th or 12th grades, at least 500 students if theydid not house 10th or 12th grade but did housean eighth grade, and at least 200 students if theydid not house eighth, 10th, or 12th grades. Theseenrollment limitations were necessary to ensurethat at least two classrooms would be eligible forobservation in all selected schools. (5) The schoolsdid not have more non-Mexican American minor-ity students than Anglo students and did not ex-clusively serve military installations. This elimi-nated schools which are quite untypical of thoseusually attended by Mexican Americans. (6) Theschools were not in districts recently investigatedor about to be investigated by Federal civil rightsagencies; this eliminated schools which were ex-pected to be particularly apprehensive aboutbeing observed by a civil rights agency and thuswould be unlikely to exhibit their normal class-room procedures.

There were 430 schools that met all of theseconditions and thus were included in the sampl-ing base. Schools were then stratified by State,grade housed ( 4, 8, 10, or 12); and percent ofthe enrollment that was Mexican American(0-24.9, 25-49.9, and 50-100). One school fromNew Mexico was randomly selected from eachstrata and two schools from California and Texasfrom each strata. A few high schools listed for twodifferent grade level strata were selected twice,once from each of the two strata. In all, 52

wseparate schools ere selected for classroomobservations, 10 from New Mexico, 22 from Cali-fornia, and 20 from Texas.

The criteria used to define the sampling baseprecludes the sample from being consideredrepresentative of all schools in the three Statessurveyed. The extent to which the defined sampl-ing base would differ from one representative ofthe whole of the three States cannot be knownwith certainty. The criteria that eliminated most

46

of the schools from the sampling base were therequirements that district enrollment composi-tion be at least 10 percent Mexican. American,and the requirement that schools not be locatedin districts which had recently been investigatedor were scheduled to be investigated by a FederalCivil Rights Agency. The first criteria biases thesampling base by including more schools witha high percentage of Mexican American enroll-ment than is the average for the States. Thesecond criteria biases the sampling base by ex-cluding numerous schools that are in districtswhere there has recently been strong evidence ofillegal segregation or ethnic disparities in schoolconditions and practices. As a result, the samplingbase is more typical of schools attended by mostMexican Americans than would be a samplingbase that includes all schools in each State.Moreover, the sampling base has fewer schoolswhich might be expected to have the greatestdisparities between Mexican Americans andAnglos.

A number of criteria were used in samplingclassrooms from the 430 schools. It was desiredto limit the observation of classes to one subjectarea. English classes were chosen because theywere expected to have more teacher-pupil inter-action than classes. with subjects such as mathor science; these latter classes were thought toinvolve mostly lectures and work by students attheir desks." Furthermore, the English languagearts were considered to be the single most im-portant-subject area for Mexican Americans, be-cause success in all other subjects is predicatedon developing skills in speech, reading, andwriting.

Observations of social studies classes weresubstituted for those of English classes in schoolswhich'had an ungraded English program or whichdid not require students to take English at thegrade level to be sampled.74 The choice of. social

"Tine literature suggests this conjecture was not correct;math claFies tervi to have the most student participation.

"An ungraded _English program is one in which individualclasses have no standard grade designation. Classes are likelyto contain pupils of different ages who frequently are identifiedaccording to level of performance in English language artsrather than grade or age level.

studies classes in these cases depended upon:1) the proportion of the students who took Englishwhen it was not required; 2) whether the schooladministration could identify a sufficient numberof English classes whose enrollment was pre-dominantly of children at the grade level to besampled; and 3) whether the organization ofthe social studies program lent itself to obtaininga more representative sample of the student bodyat the appropriate grade level. Approximately 7percent of the classes visited were social studiesclaisses.

All fourth and eighth grade English or SocialStudies classes in elementary and intermediateschools were visited. At least half of the 10thand 12th grade English or Social Studies classeswere randomly sampled in the high schools. Inhigh schools which practiced some form of "ho-mogeneous ability" grouping of the students, thesampling design called for a random selection ofat least half of the classes at each "ability" level.However, it was discovered that most schools hadfew high or low "ability" classes, but many middle-ability- classes. Consequently, nearly all high andlow "ability" classes were included in the sample,whereas only about half of the middle "ability"level classes were selected. Provisions were madeto observe each teacher with at least one eligibleclass. As a result of the above sampling proce-

dures, some teachers were observed more thanonce. No teachers refused to be observed.

A total of 494 classrooms (80 in New Mexico,198 in California, and 216 in Texas) were observedby Commission staff. Of these, data suitable formost analyses were obtained for 429 classes in-cluding 70 in New Mexico, 171 in California, and188 in Texas. Observed classes were eliminatedfrom the analyses if (1) they were mistakenlyselected; (2) the amount of collected data wasinsufficient for analysis; or (3) the class did nothave at least one Mexican American and oneAnglo student. Most classes mistakenly selectedwere mixed grade classes where many of thestudents proved not to be at an eligible gradelevel. Classes were eliminated because of insuffi-cient data if the observer spent less than approxi-mately 5 minutes coding verbal activity. Assign-ment of students solely to desk work for the entirebbservation period was another cause of insuffi-cient data. When this occurred observers re-quested the teachers to interact with the class forat least 10 minutes. A few teachers failed to com-ply. Since a major part of the analyses of class-room interaction was devoted to the examinationof possible disparities in the interactions of teach-ers with Mexican American and Anglo students,a classroom was not included unless there wasat least one student from each group.

APPENDIX B.

DATA COLLECTION

Instruments

The data on the teacher and pupil interactionswere coded with the Flanders Interaction AnalysisCategories. The Flanders system was chosen be-cause it focuses on teacher behaviors most di-rectly related to encouraging and involving stu-dents in the learning process and because it is thesingle most widely used classroom observationsystem in education research. It includes 10 cate-

, gones of behavior which are summarized in Fig-ure 14.

Data were collected on a number of teacher,classroom, and school characteristics associated

48

with each observed classroom. This informationwas gathered so that an analysis could be madeattempting to relate these characteristics to possi-ble disparities in teacher behavior toward studentsof different ethnic groups. Most of the character-istics were chosen for analysis because of theiralleged importance to the quality of instruction,to student achievement, or to human interaction.

Data on classroom characteristics were collectedby observers' visual inspection. A brief (less than5 minutes) interview with the teacher at the endof the instructions! period Provided additionalinformation. The forms used to collect these dataare shown in Appendix F which begins on page

Figure 14

Summary- of the Flanders Categories for Interaction Analysis

1.

2.

INDIRECTINFLUENCE

3.

4.TEACHER

TALK

DIRECT 6,INFLUENCE

7.

.

STUDENTTALK

9.

ACCEPTS FEELING: accepts and .clarifies the feeling tone of thestudents in a nonthreatening manner. Feelings may be positive ornegative. Predicting and recalling feelings are included.

PRAISES OR ENCOURAGES: praises or encourages -student actionOF behavior., Jokes that release tension, not at the expense of an-other individual, nodding head or saying "uhhuh?" or "go on" areincluded,

ACCEPTS OR USES .IDEAS OF STUDENT: clarifying, building, ordeveloping ideas or suggestions by a student. As teacher brings-more- of his own ideas into play, shift to category five._

ASKS QUESTIONS: asking a question- about content or procedurewith-the intent that a student answer.

LECTURING: giving facts or opinions about content or procedure;expressing his own idea; asking rhetorical questions.

GIVING DIRECTIONS: directions, commands, or orders with whicha student is expected to comply.

CRITICIZING OR JUSTIFYING AUTHORITY: statements intendedto change student behavior from nonacceptable to acceptable pat-tern; bawling someone out; stating why the teacher is doing whathe is doing, extreme self-reference.

STUDENT TALK-RESPONSE: talk by students in response to teacher.Teacher initiates the contract or solicits student statement.

STUDENT TALK-INITIATION: talk by students, which they initiate.If "calling on" student is only to indicate who may talk-Tlext, ob-server must decide whether student wanted to talk. If he did, usethis category.

10. SILENCE OR CONFUSION: pauses, short periods of silence, andperiods of confusion in which communication cannot be under-stood by the observer.

Source: Edmund 1. Arnidon and Ned Flanders, The Role of the Teacher in the Classroom: A Manual for Understan g and ir,,r rovingTeachers' Classroom Behavior, Minneapolis: Paul S. Arnidon Associates, 'l%3, p, 12.

64. Data on the ethnic composition of the schoolswere collected from an interview with the princi-pal; often school records were consulted to as-sure accuracy. Information on the socioeconomicstatus (SES) of the parents of the school's studentbody was collected by .requesting the informa-tion from the principal in a short questionnairemailed to the principal after the classroom ob-servation had been completed." (See AppendixG which begins on page 66.)

Coding Procedures

The coding procedures were as follows: Theobserver entered the room at the beginning of aninstructional period and took a seat, usually in theback of the classroom, but occasionally to oneside. The first several minutes were used to fillin a seating chart indicating the location of allchairs in the classroom and the ethnicity of eachoccupant 76 If the class started with organizationaltasks such as roll calls, the handing in or returningof papers, or similar tasks, the observer used thattime to code a number of characteristics of theteacher and classroom which could be ascertainedvisually. If the class started with instructional inter-actions between the teacher and the students, theobserver started Coding those interactions as soonas the seating chart Was completed, and coded theother characteristics later.

Ten minutes were used to code the classroominteraction. At a fixed rate of once every 3 secondsthe observer marked a tally on the coding formunder the most appropriate behavior categoryand participant designation. An additional tallywas also marked whenever a change in the typeof classroom interaction occurred in the middleof the 3-second coding interval." Also, when a

's The decision to collect this data was made after the ob-servers had returned from the Southwest.

When the first few schools were observed, the observerchecked their perceptions of the students' ethnicity with theteacher's perceptions. In a few cases light haired Mexican Ameri-cans appeared to be Anglo and some American Indians appearedto the observers to be Mexican Americans, but usually the ob-servers could make the proper distinctions if the child spokeduring the observation period: Consequently the procedure wasrevised so that observers only checked their perceptions of thestudents' ethnicity with the teacher when thby were not reason-ably confident of their accuracy,

50

teacher asked a question to a group of studentsor to the class as a whole and then called on anindividual student to respond, the question wasfirst coded as directed to the class as a whole,and then an additional tally in the questioningcategory was marked under the ethnicity of thestudent designated to respond.

In a few cases, before the 10 minutes of codingcould be completed there was a suspension ofthe normal.classroom interaction for such activitiesas listening to a recording or having the studentsdo a work assignment at their desks. When thishappened, the observer stopped coding and re-sumed when there was again some teacher-pupilinteraction.

This 10-minute segment of a classroom's instruc-tional process is not considered to be representa-tive of the instructional process in any individualclass. However, a sample of 10-minute observa-tions from a large number of classrooms is likelyto be representative of the interaction of class-rooms in the sampled population.

At the end of the class session the observerspoke with the teacher for a few minutes to col-lect information about the teacher's training andthe criterion used in assigning students to seatingpositions. Occasionally observers also checkedwith the teacher about their perception of a stu-dent's ethnicity.

There were five observers. An additional staffmember, using a course schedule obtained fromthe school or district, was responsible for assign-ing observers to classrooms. The course sched-ules included the following information on allcla5ses: name of teacher, grade, track level (iftracked), room number, course title, and time ofmeeting. In addition, at the fourth grade level,the schedules indicated the time of day Englishlanguage arts were taught. The staff memberscheduling observers tried to assure that eachobserver was assigned an equal proportion ofteachers by sex and ethnic background and anequal proportion of classr6oms by`track and gradelevel. Tests on the data indicated that this effortwas successful."

"This is a standard coding convention of the Flanders system,"Chi- square tests at the .01 level of alpha error.

APPENDIX CDATA PREPARATION ANDVARIABLES FOR ANALYSIS

The coding form for observing classroom inter-actions had 11 categories of interaction. Thesecategories were coded for five distinctions ofparticipants: individual Mexican American stu-dent; individual Anglo student; individual blackstudent; individual student of another ethnicity;and a group of students or the class as a whole.'The interactions associated with black studentsand students of -"other" ethnic groups were notanalyzed because of the extremely small numberof such students. The interactions associated witha group of students or with the class as a wholewere also not analyzed because this study wasintended to focus on possible disparities betweenstudents of different ethnic groups_

The nine measures for Chicano students andthe nine for Anglo students were modified in five-ways to create the indices of classroom interactionbehaviors which were to be analyzed. (1) Thebasic measures were corrected for variations inthe total number of tallies for each observationof a classroom. Observers were trained to codewith approximately one tally every 3 seconds for10 minutes, and they used stopwatches to guidetheir coding frequency and the duration of cod-ing. For a number of reasons, not- all coding ses-sions resulted in exactly 200 tallies. The previouslydiscussed coding procedures (on page 50) accountfor some of these reasons. Also, in some casesthe class period ended _before the observer couldcomplete 10 minutes of codipg. All classes forwhich observers had less than 100 tallies wereeliminated from the sample. All measures for theremaining classes were corrected to a standard200 tallies by multiplying each of the 18 basicmeasures by (200/total number of tallies for ob-senation of that class.) This eliminated the inter-classroom variation that was due to the total num-ber of tallies.

(2) Four compound measures were created for79 1) an p. 13.

Chicano and Anglo students from the basic ninecategories. A measure of total positive feedbackwas constructed by combining categories 1, 2,and 3 (teacher accepts students' feelings, teacherpraises or encourages students, -and teacher ac-cepts or uses students' ideas). This combines allthe behaviors which motivate a student to con-tinue or improve on what he is doing. A measureof total noncriticizing teacher talk was constructedby combining all categories of teacher talk exceptthe one of criticizing or justifying authority (Cate-gory 7). A measure of total student talk was madeby combining Categories 8 and 9 (student re-sponse talk and student initiated talk). And a meas-ure of the indirectness of teacher behavior wasconstructed by adding the first four teacher be-havior categories (acceptance of students' feelings,praise or encouragement, acceptance or use ofstudents' ideas, and questioning) and dividing thistotal by the sum of all teacher behaviors. The firstfour teacher behaviors are considered indirect be-cause they seek or encourage student initiative inthe learning process. The remaining three teacherbehaviors are considered direct, because theymanage or control the learning process for thestudent.

(3) The Mexican American interaction measuresand the Anglo interaction measures for each classwere converted to average per pupil measures bydividing the total number of tallies for each inter-action behavior for each ethnic group by thenumber of students of that ethnicity in that class,This was necessary because different classroomshad different numbers of Mexican Americans andAnglos. For example, in a class with a large pro-portion, of Chicano students, it is expected thatthere will be substantially more interactions di-rected by the teacher to Chicanos than in classeswhere there is a small proportion of that ethnicgroup. The only unit of analysis that is comparablefor all classrooms is the average student of eachethnic group..

The measures of indirectness of teacher be-havior toward the average Mexican American andof indirectness of teacher behavior toward theaverage Anglo were discovered to be invalid for anumber of classes, and were consequently dropped

52

from the analysis."(4) Mexican American and Anglo per pupil

measures were corrected for class size, The totalnumber of tallies should be unaffected by classsize, but per pupil measures will be. This is be-cause teachers have a relatively fixed amount oftime to interact with individual students, so themore students there are in the class, the less timethe teacher will have to interact with each one.The purpose of controlling for class size is not todeny its importance on the per pupil measures.The correction for class size was made by multi-plying each per pupil interaction measure by classsize divided by 25." This standardized the perpupil 'measures to the equivalents for a class sizeof 25, which was the mean class size for thesample. This control does not prevent studyingthe relationship which class size might have onthe standardized per pupil measures.

(5) A final major modification of fhe data wasto construct difference scores for each classroomfrom each of the 12 per pupil measures for Mexi-can Americans and the corresponding per pupilmeasures for Anglos." This allowed the elimina-tion of student ethnicity as a separate factor whenstudying the differences in per pupil measures ofMexican Americans and Anglos in relationship toschool, classroom, and teacher, characteristics.

A total of 36 interaction variables are developedfrom the above process. They are listed in Figure15.

Data on teacher, classroom, and school char-acteristics were converted into 22 indices of thesecharacteristics, (listed on pp. 23 & 24 of the text).The construction of these indices and the reasonsfor examining them in this study are discussed be-low.

"The calculation of this measure resulted in division by zerofor a number of cases.

Control could have been induced by either correcting theper pupil measure for class size or conducting all analyses withclass size as a crossed factor. lo have controlled by use of sizeas a crossed factor would have required t.47) way analysis ofvariance with unequal cell size. A seldom mer tIoned characterislic of the popular computer programs which du tuck analysesis that they will sometimes yield highly erroneous results if thecell are not approximately equal in size

"The per pupil measures for Ankles were used in this studyjust to construct these difference scores; they were not directlyanalyzed.

Figure 15

Classroom Interaction Behavior Indices Used inthe .Analyses..

Average Mexican American Per Pupil Measures

1. Average teacher acceptance of MexicanAmerican students' expressed feelingsAverage teacher praise or encouragementof individual Mexican American studentsAverage teacher acceptance or use of indi-vidual Mexican American students' ideas

4. Average teacher questioning of individualMexican American students

5. Average teacher lecturing to individualMexican American students

6. Average teacher giving directions to indi-vidual Mexican American students

7. Average teacher criticizing of individualMexican American students

a. Average response talk by individual Mexi-can American students

9. Average initiated talk by individual Mexi-can American students

10. Average teacher positive feedback to indi-vidual Mexican American students

11. Average noncriticizing teacher talk to indi-vidual Mexican American students

12. Average of all talk by individual MexicanAmerican students

Average Anglo Per Pupil Measures

1. Average teacher acceptance of Anglo stu----dents' expressed feelings

2, Average _teacher praise or encouragementof individual Anglo students

3. Average teacher acceptance or use of indi-vidual Angle students' ideas

4 Average teacher qtrektioning of individualAnglo students

5. Average teacher lecturing to individualAnglo students

6. Average teacher giving directions to indi-Vidual Anglo students

7. Average teacher criticizing of individualAnglo students

8. Average response talk by individual Anglostudents

9. Average initiated talk by individual Anglostudents

10. Average teacher positive feedback to indi-vidual Anglo students

11. Average noncriticizing teacher talk to indi-vidual Anglo students

12. Average of all talk by individual Anglo stu-dents

Measures of Difference Between Average Mexi-can American Per Pupil Measures and AverageAnglo Per Pupil Measures

1. Difference in teacher acceptance of Mexi-can American and Anglo students' feelingsDifference in teacher praise of MexicanAmerican and Anglo studentsDifference in teacher acceptance or use ofideas of Mexican American and Anglo stu-dents

4. Difference in teacher questioning of Mexi-can American and Anglo students

5. Difference in teacher lecturing of MexicanAmerican and Anglo students

6. Difference in teacher giving directions toMexican American and Anglo students

7. Difference in teacher criticizing or justify-ing authority to Mexican American andAnglo studentsDifference in response talk by. MexicanAmerican and Anglo studentsDifference in initiated talk by MexicanAmerican and Anglo students

10. Difference in teacher positive feedback toMexican American and Anglo students

11. Difference in noncriticizing teacher talk toMexican American and Anglo students

12. Difference in all talk by Mexican Americanand Anglo students

Data on the State in which the classrooms werelocated were used to test for possible differencesbetween geographical areas. References to Stateare actually references to the limited geographicalareas from which schools were sampled in eachof the three States: Grade and subject were in-cluded because they are often considered impor-tant to pedagogical techniques and it was of inter-est to discover whether they affect the actualteacher-pupil interaction.

Characteristics of the teacher, such as educa-tional attainment, inservice training, ethnicity, age,

and sex, were included because of their possiblerelations to differences in teacher behavior towardpupils. Personality or attitudinal characteristics ofthe teacher would also have been desirable vari-ables for investigation, but the short time availablefor interviewing each teacher precluded validmeasures of these characteristics.

The school enrollment figures for each ethnicgroup of students were converted to measures ofthe Mexican American percentage of school en-rollment and Anglo percentage of school enroll-ment. Class enrollment figures for each ethnic'group were converted to measures of MexicanAmerican percent somposition of the classroomand Anglo percent composition of the class-room.83 Previous research indicates that the mi-nority ethnic composition of schools is related toa number of characteristics of the school, includ-ing teacher morale, student attitudes, and studentachievement.84 Furthermore, it is well known thatethnic background of people in small groups af-fects their interactions in those groups. Conse-quently; it seemed likely that the' ethnic composi-tion of the school and classroom would be relatedto possible disparities in teacher-pupil behaviorsassociated with Mexican American and Anglo stu-dents.

Data were collected as to whether each ob-served classroom was tracked; and if so, whetherit was a low, medium, or high track. Tracking is

" During the analysis it became apparent that the measureswere almost complementary fur the classrooms and schools inthe Commission sample.

a' Guthrie, et. I. Ch. 4.Coded at the top of the classroom Protocol form (Appendix

n on the line following "AG". See p. 64,

54

the practice of assigning students to classrooms soas to make class enrollments more homogeneousin respect to some purported measure of the stuxdents' ability or performance. It was of interest tosee how the practice of tracking, which manyschool officials claim is for the benefit of all stu-dents, affects teacher-pupil interactions amongMexican American and Anglo students.

A measure of the degree of ethnic concentra-tion was constructed by computing the varianceof the Anglo ethnic composition of all classes

observed in a given school. This indicates theextent to which Anglo and minority students ina given school are separated into different class-rooms. Racial or ethnic segregation betweenschools has historically been associated with un-equal distribution of educational resources. It wasthought that the segregation of students withinschools might also be associated with inequitiesin teaching.

During the classroom period, observers filledout a seating chart (See p.6.5) which indicated theposition of each seat in the classroom and forthose seats that were occupied, the ethnicity of its

occupant. Data on the seating placement of stu-dents of each ethnic group were used to constructseparate seating indices for Mexican American andAnglo students for each class. Seating positionswere defined in reference to the primary locationof the teacher and the arrangement of the stu-dents' seats. Usually the teacher was in the frontcenter of the classroom, but was occasionally inone corner or to a side. Prime seats were definedas those in a "T" pattern with the teacher locatedin front of the intersection of the two bars (SeeFigure 16). The standard "r' pattern consisted ofa column of three seats intersecting the first rowin front of the teacher's primary position, with onefront row seat on each side of the point of inter-section. If the teacher's primary position was centered between two columns of seats an expanded"1" pattern was used, consisting of a double col-umn three seats deep and with one front rowseat on either side of the column. When the ex-panded "T- was used, each seat counted only 5/8in order to be equivalent in weight to the seats inthe standard tiT"-pattern, If the teacher's primary

A(T)

X x x X X

X X x x xX x x xx X x x x

Standard

X

X

Figure 16

Definitions of Prime Seats in ClassroomSeating Patterns

x

x

(T)=Teacher

B

Expanded "T"

!position was in a corner of the room a diagonal"T" similar to the standard but focusing ona corner, was used to define the prime seats.

The back row was considered the last row ofoccupied seats. If the last row of seats had morethan one seat and at least one occupied seat, itsseats were coded as back row seats. But if thelast row had just one seat and it was occupied,it was coded as an isolate seat, and the next rowforward with at least one occupied seat was codedas the back row. Isolate seats also included anyoccupied seats with no occupied seats in front,behind, or on either side of them. In a few caseswhere a circular seating arrangement was used,no students were coded as being in the "T" or inthe back row.

The index was constructed by assigning a valueof '1.0 to prime seats, 10 to back row seats, 4 0to "isolate" seats, and 2.0 to all other seats. TheMexican American seating index equaled the sumof the number of Mexican Americans in prim eseats times 1.0, plus the number of Mexican Amer-icans in back row seats times 10, plus the numberof Mexican Americans in isolated seats times 4.0,plus the number of all other Mexican Americanstimes 2, all divided by the number of MexicanAmericans in the classroom. The Anglo seatingindex was constructed in a similar manner.

This index is a measure of seating priority for.Mexican Americans and Anglos in a classroom.Previous research indicates that the students inprime seats tend to be involved in about 50 per-

C

Diagonal "T"

cent of all classroom teacher-pupil interaction,which is a large percentage considering that primeseats are defined as being five seats of a normal20-30 seats." Consequently, this appeared to bea potentially important variable when analyzinginteraction data.

Data on the criteria by which students are as-signed or allowed to choose their seating positionwere also collected. Seating criteria were dividedinto five categories: student choice, student choicewith teacher modification (teacher modificationwas usually to correct discipline problems), alpha-betical order, homogeneous "ability" grouping,and other methods of teacher choice. It wasthought that these criteria would reflect the teach-er's attitudes toward the students and influencestudent expectations, both of which may be mani-fested in the teacher-pupil interactions.

Data were collected from each school on thepercent of students from each ethnic group whocame from families with an annual income of lessthan $3,000 and greater than $10,000, and fromfamilies where the head of the household has hadeight or less years of education, a high schooleducation, and a college education. These vari-ables were used to construct an index of the socio-economic status (SES) of the Mexican Americansin each school, an index of the SFS of Anglos ineach school, an index of the difference betweenthe Mexican and Anglo SES, and a school average

16 Raymond Adams and Bruce Biddle, Realities of Teaching,New York: Holt, Rinehart and Winston, Inc., 1970. pp. 49 -51.

SES index. The average of the percent low educa-tion and the percent low income was used tomake the SES indices. The school average indexis a weighted average of the Mexican Americanand Knglo indices." the SES indices arei'ell c; measures of low iES, (i.e., the higher theindex the larger the percentage of families of lowSES). The socioeconomic status of a student's par-ents has repeatedly been proven to be the singlecharacteristic most strongly associated with stu-dent academic performance in schools as they arenow constituted in this country. Consequently, it

was of interest to determine if SES is also associ-ated with differences in teacher-pupil interactions.

Twelve of the above indicated 22 measures of.school, classroom, and teacher characteristics areinterval measures. For the purpose of most of theanalyses they were converted to categorical meas-ures. Figure 17 shows the cutoff points usedwhen this was done.

" Because the schools in the sample had few students of otherethnic groups th index is a good estimate for the whole school.

Figure 17

Category Cut-Off Points For Interval Measures ofSchool, Student, and Classroom Characteristics

Teacher Age:

School Percent Mexican American:School Percent Anglo:

Mexican American Seating Index:Anglo Seating Index:

Class Percent Mexican American:Class Percent Anglo:

Class Size:

Ethnic Concentration In School:School Mexican American SES:

School Anglo SES:

School Average SES:

Difference Between the Mexican American SES and the Anglo SES: 0-14.0, 14.1-22.5, 22.5-100

20-29, 30-39, 40-49, 50-100

.0-29.9, 30-49.9, 50-100

0-44.9, 45-67.9, 68-100

1-1.92, 1.93-109, 2.10-4.0

1-1.92, 1.93-2.09, 2.10-4.0

0-22.1, 22.2-36.9, 37-59.9, 60-100

0-32.9, 33-58.9, 59-75,9, 76-100

0-17, 18-22, 23-27, 28-41

0-84.9, 85-214.9, 215-1000

0-22,5, 22.6-38,8, 38.9-100

0-6.2, 6.3-118, 119-100

0'11.0, 11.1-30.0, 30.1-100

APPENDIX D. THE ANALYSES

The disparities in the classroom behaviors be-tween the average individual Anglo and the aver-age individual Mexican American in each classwere studied with the matched sample t test. Thedifference scores were constructed by subtractingthe average Anglo per pupil measure of a givenbehavior in each classroom from the correspond-ing Mexican American per pupil measure_ Therewere 12 such tests, one for each of the 12 perpupil behaviors_ This was done because thematched sample t test was expected to providea more powerful test than would have been pro-vided by one-way analysis of variance using stu-dent ethnicity as the classifying factor." Further-

more, the investigation of the relation of theteacher, classroom, and school characteristics tothe disparities between students of the two ethnicgroups was made conceptually more simple bythe use of the difference scores than it would havebeen by testing the interaction effect in a two-way analysis of variance model.

"The matched sample t test is more powerful than the twosample t test only when there is at least moderate covariancebetween the- measures for which the difference scores are cal-culated. This seemed likely because all students in a classroomdo have the same teacher, but it was not empirically knownprior to the data analysis. Therefore, all 12 of, the matched sam-ple t tests were replicated using the two sample t tests for inde-pendent samples; the results indicated that the matched samplet test was in fact the more powerful one for these data.

Significant disparities in teacher behavior ex-

hibited toward Anglo and Chicano pupils werealso studied for possible relationships to the 22teacher, classroom, and school characteristics. The

research question was whether certain levels orcategories of any of the characteristics might ac-

count for most of the total disparity in a givenclassroom behavior." There is no simple test ofthis queqion. The extent to which the disparityin any one category or level of a characteristiccan account for the total disparity in a classroombehavior is dependent on a number of factors.These include: the mean value of the disparity,the standard deviation of the disparity, the numberof the cases in that category or level, and the valueof these statistics relative to those of the othercategories or levels. One possible approach ofanalysis is to do independent tests for the signifi-cance of the disparity in each category or levelof each characteristic and then make deductionsbased on the results and the relative number ofcases in each category or level. This approachmanages to combine the bad futures of highalpha error (because of the large number of tests)and low power (because of the relatively smalldegree of freedom due. to making each test ononly a small part of the sample). The alternativeapproach of analysis is to test for significant differ-ences in the magnitude of a disparity within class-rooms across levels or categories of each. char-acteristic and then make deductions based on theresults and the relative Ns. The latter approachwas used in this study because it has higher powerand lower alpha error than the former one.

" The term "overall disparity" will be used to refer to asignificant difference between Mexican Americans and Angleswithin classrooms for all 429 classrooms taken together.

A number of additional analyses were con-ducted to help interpret the results in the aboveoutlined tests. First, all the pairwise associationsbetween the 22 conditions were investigated bychi-square tests. (See Appendix H, Display 1, p.

67.) Second, all of' the analyses of the possible

relations of significant disparities within the class-room to the 22 teacher, classroom, and schoolcharacteristics, were replicated using MexicanAmerican per pupil scores as criteria in place ofthe Mexican American-Anglo difference scores.

This provided insight into whether discoveredvariations in the per pupil difference scores across

categories or levels of a given condition are theresult of variations in the Mexican American per

pupil measures, the result of variations in theAnglo per pupil measures, or the result of bothtypes of variations. Third, the associations between

the classroom coder and the 22 conditions were

tested by chi-square to Investigate possible biases

in the assignments of coders to classrooms. Fourth,

each of the 12 difference scores were tested fordifferences between coders by one-way analysisof variance with coder as the classifying factorand each behavior measure as a criterion, Fifth,

Dunn's pairwise multiple contrast and Scheffe'slinear trend contrasts were used as post hoc pro-cedures for the analysis of variance tests. Sixth,all pair-wise correlations between the variouscategories of intenction were calculated to show

the interrelationships between these measures.

(See Appendix H-3 on p. 68). All statistical-testswere conducted at the .01 level."

" The total disparities in interaction were appriori hypothe-sized to be positive except for directions and critici5m; theywere analyzed with one.tail tests.

fj

APPENDIX E.

SOURCES OF POSSIBLE BIAS

A bias is a tendency to err in a given manner.There are three major sources of possible biasesin classroom observation studies: coding biasesof observers, obtrusive biases caused by the ob-server's presence, and sampling biases. Both genteral and specific problems of this study concerningeach of these possible sources of bias are dis-cussed in the following sections.

Possible Coding Biases

Coding biases can be intentional or uninten-tional. It is usually presumed that the researchersare not intentionally biased, but sometimes suchdeceptions are discovered. The more commonconcern is that researchers are unconsciouslybiased in their observations due to their expecta-tions or other subconscious needs. An observerwho wants to discover a certain pattern of be-havior may operate with perceptions biased in

favor of finding that pattern. Or a .person withserious personal problems in handling a certaintype of interaction may not correctly perceiveinstances of that type of interaction.

All observers in a study may be biased in thesame manner, all may be biased in different ways,or some may be biased and others unbiased. Ob-servers may be biaSed in only one dimension ofthe coding scheme, in more than one dimension,or in a compound manner, as in a statistical inter-action of two or more variables. An example ofthe latter would be if a coder tended to codeCategory 3 behaviors of the Flanders form asCategory 2 behaviors for Anglo students but didnot code the behavior of Chicano students in thismanner.

Observer biases may be what shall be calledspecific or nonspecific and reciprocal or non-reciprocal. A specific bias is the result of misper-

ceptions in a consistent direction, such as thetendency to code Category 3 behaviors as Cate-gory 2, It is the direction that must have ;omeconsistency. There need not be consistent mis-perceptions for a coding bias to be specific. Anonspecific coding bias is one where the codermarks too few or too many tallies in a given cate-gory, but not as a result of misperceptions in aconsistent manner, For instance, the observer maynot tally as many Category 3 behaviors as he orshe should, but when doing so does not overtally in just one or two other categories. A recip-rocal bias is one in which two categories areincorrectly tallied, but in a manner that tends tocancel out errors in the sum of the tallies. Forinstance, an observer may easily tend to confusebehaviors in Categories 2 and 3; if so, he willsometimes code Category 3 behaviors as Category2 behaviors, but he will also sometimes code Cate-gory 2 behaviors as Category 3 behaviors. For theerrors to completely cancel the product of thefrequency of the behavior times the frequency ofthe error must be equal for both categories. Anonreciprocal observer bias is one in which onebehavior is sometimes incorrectly coded as a sec-ond behavior, but the second behavior is seldom,if ever, incorrectly coded as the first behavior.

Most observational category systems are ipsi-tive; that is, if one category is overtallied, someother category or categories have to be under-tallied. This creates further complications in theeffects of the above discussed forms of observerbias. However, ipsitivity should be thought of interms of degree rather than as an absolute, andthe degree is inversely proportional to the numberof options in the bounded set. Consequently,observational systems which have a substantialnumber of categories do not have serious prob-lems with ipsitivity.

The last important characteristic of observer'bias is that biases among observers may corn-pound or cancel out one another. Tiliy tend tocancel out each other when all of the followitngconditions are present: (1) Some observers arebiased in an inverse manner and equal extent assome of the other observers; (2) observers areassigned to classes at random; and (3) each ob-

60

server codes a substantial number of classes.Under these circumstances, observer biases mayhave little effect on statistics of the collecteddata."

Coding- biases can be controlled in two ways.Coding systems and training can be designed tominimize the likelihood of bias, and checks canbe used to detect coding biases during the actualobservations so that biased observations can becorrected, eliminated, or taken into account wheninterpreting the results.

Coding systems which have categories that aremutually exclusive, all inclusive, and have explicitcoding conventions will help minimize the likeli-hood of unintentional coding biases. Thoroughtraining with constant .feedback and analysis ofthe source of error also is important for minimiz-ing coding biases.

Both the category system and training used inthis study should have been highly conducive tominimizing coding bias. The Flanders system is asystematic coding procedure. Its categories aremutually- exclusive and taken as a whole are allinclusive. Only a few coding conventions areneeded to code most classroom interaction 's -withthis observation system.

The person who trained the Commission ob-servers was an Anglo who had not conductedany studies on ethnic disparities in the classroominteractions, and had no reason to think that theCommission would again in the foreseeable fut-ure conduct studies which would offer him con-sulting opportunities, He played no part in thestudy other than to train and check the coders'proficiency. Consequently, there is no apparentreason to suspect that he trained the observersto code in a biased fashion.

The trainer provided substantial feedback tothe observers during all phases of their training.Before starting the data collection,. all test codingsof the five_observers showed a minimum of .85reliability with the trainer's coding, except in onecase which the trainer thought involved a class

"" Such biases will always have some effect on the inferential"statistic because they increase the error variance which reducesthe chance of finding- statistically significant differences whentesting hypotheses.

session particularly difficult to code." Most testsessions yielded reliability coefficients exceeding.90. On two occasions during the course of theactual field work, the five observers did additionaltest coding to calculate their reliability with thetrainer. In all but one instance the reliability coeffi-cients exceeded .90, and again, in that one casethe trainer felt the classroom session to be un-usually difficult to code

Reliability was measured by Scott's Pi .coeffi-cient. The Scott's Pi coefficient is used to deter-mine the degree of correspondence between twocoders; a value of 1.0 indicates perfect correspon-dence and a value of 0.0 indicates no correspon-dence. These checks of reliability yielded resultsranging from .8 to .9, which is unusually high forobservers using the Flanders system. These highresults are probably due partly to the intensivetraining, and partly to the fact that coder reliabilitywas calculated between each observer and thetrainer rather than between observers as is usuallythe case and which is likely to yield somewhatlower reliability estimates.

The Scotts Pi reliability coefficient was designedto calculate Intercbcler reliability without studentethnicity distinctions in the coding. When thecoder reliability checks were conducted for thisstudy, the data for each behavior were summedacross the different student ethnicitieS and theclass as a whole distinctions. This procedure isinadequate for testing the possibility of coderbiases related to student ethnicity. It was usedonly to calculate a_ summeric statistic of overallcoder reliability.

In each reliability check session the trainer ex-amined the difference between the, number oftallies in each cell of his coding sheet and thenumber of tallies of the corresponding cell of theobserver's coding sheet. The trainer. reported thathe did look for biases associated with studentethnicity, but that none were apparent from theexamination of the coding sheets. Unfortunately,these coding sheets were lost; therefore, statisticscould not be calculated to confirm the trainer'sreport. However, a post hoc analysis of the col-

92 Reliability or this s--ion vita .78.

lected data suggests that there were no seriouscoding biases related to student ethnicity. (Seep.62 of this appendix).

The best way to check for coding biases duringactual observations is to have an objective stand-ard simultaneously code classroom observers.There are two serious problems with this method.The first is that it is almost impossible to validlypresume that any observer can be an objectivestandard by which to judge the other observers.The best that can be done is to use a person asa standard who is thought least likely to be biased.The second problem is that if the person servingas an objective standard could accompany all theobservers during each of their coding sessions itwould be unnecessary to use any observers otherthan the objective standard. If the objective stand-ard only spot checks the observers, it should bedone in such a manner that the observers are notaware that they are being checked. This must bedone so that they will not s'bive to be particularlysharp or alert when checked, but less attentivewhen they know they are not being checked.However, for a person to complete such a spotcheck without the observer knowing he is beingchecked is very difficult. It requires specially de-signed observation rooms or the use of televisionand audio equipment in classrooms.

Another approach to checking for coding biasesduring actual observations is to put two observersin a classroom at the same time but in positionsso that they cannot see each other's codings. Theproblems with this method are that it. will notdetect biases which are being induced in a similarfashion by both observers and, when there aredifferences between observers, it is difficult todetermine which one of the two is biased, orwhether both are biased, but in different ways.Furthermore, two observers in a classroom aremore likely to make the teacher and studentsanxious than is one observer, thus :increasing thechance of obtrusive bias. (See pp. 62 & 63 in thiSappendix.)

A third approach is to have different observerscode the same class at different times. But 'sinceclassroom interactions vary a great deal from hourto 'hour, it would be necessary for-each observer

61

to obServe the same class many times. This wouldnot be feasible except for intensive or longitudinalstudies. This approach also has the problem thatwhen differences between observers are indicated,it is not known which of the observers is biased.in addition, differences between observers maybe due to differences in the obtrusive effects theyhave on the class rather than on coding differ-ences.

A fourth approach to use in checking for codingbiases among observers in actual classes is to as-sign observers to classrooms at random and haveeach observer code a large number of classes(but with each class observed by only one coder).Then each observer's average codings are corn-pared with the average codings of other observers.This approach has problems of interpretation thatarise both from comparing observers to each otherrather than to an objective standard, and frompossible differences caused by the obtrusive ef-fects of different observers. Also, as with anyrandom sampling, there is always a small chancethat the random assignment of observers to classeswill not result in each observer getting a com-parable sample of classes."

Despite problems with this approach, it pro-vided the best check for coding bias that couldbe made for this study. Observers were not as-signed to classrooms in a truly random manner,but the method was such that it could be expectedto have the same effect for the purposes of thischeck, All observers were assigned to classroomsby a team leader not involved in the observingwho tried to assure that all observers were as-signed an equal proportion of Spanish Surnamedand Anglo teachers, male and female teachers, anequal proportion of classrooms of varying trackcharacteristics (untracked, and low, medium, orhigh tracked), and an equal proportion of class-rooms at each grade level observed (4, 8, 10, and12). The only other information known when as-signing observers to classes was the time and roomnumber of the class and the name of the teacher.Statistical tests indicate that the efforts to assign

" This chance is minimized by having each observer rodelarge number of classrooms. In this study the smallest numberof classrooms observed by any coder was 53.

observers equal proportions of the above enumer-ated characteristics were successful."

The difference in tallies regarding the behaviorof Mexican Americans and Anglos in each class-room was compared among coders for each ofthe 12 behavior measures. There was a significantdifference among coders on only one behaviorpraise. It existed only between the observer whocoded the difference as being least favorable forMexican Americans and the observer who codedthe difference as being most_favorable for MexicanAmericans. Probably the hardest distinction tomake in the Flanders Coding System is betweenthe different categories of positive feedback. Asthere was no significant difference between ob-servers on the compound category of positivefeedback (suM of Categories 1-3), the significantdifference between the observers appears to bemainly due to reciprocal biases within the threecategories of positive feedback.

The lack of differences between observers inthe coding indicates that if there were codingbiases related to student ethnicity; they had tohaVe been made in same direction and to thesame extent by all five observers.This seems quiteunlikely, especially since the observers were fromdifferent ethnic or racial backgrounds and two ofthem had no knowledge of or experience withthe education of Chicanos before being assignedto work on this study.

Pt.13sible Obtrusive -Biases

Obtrusive biases are caused when the presenceof an observer in the classroom induces the teacheror students to their normal interactionpatterns. The nature of the obtrusive affects aredependentupon the characteristics of the observer,the teacher, and the students, Obtrusive biasesmay -laffeCt the teacher's behavior, the students'behavior, the behavior' of the teacher in inter-action with student characteristics (that is, astatis-.tical interaction) the teacher's behavior in inter-action with .the teacher's characteristics, or studentbehavior in interaction with the student's char-acteristics An examlitel.of the latter would be if

"_Chl7"SOuare.tests at .01 level,

Mexican American students, and only students of

this ethnicity, tended to initiate more speakingwhen being observed by a coder. There may also

be an interaction effect between the observers'characteristics and the characteristics of the teacher

or the students. An example of this would be ifMexican American students tend to initiate more

speaking when observed by a Mexican American

coder, but not when the observer is of anotherethnicity, while Anglo students do not initiatemore speaking when there is an observer in theclass, regardless of his ethnicity.

Obtrusive biases can be minimized by having

the observer's visit announced in advance, attempt-

ing to insure that the observer is not perceived

by the teacher or students as being biased against

them or having power over their welfare, and by

having the observer become familiar to the classbefore he starts coding.

The first condition was met; the teachers were

informed by the principal a day or two in advance

that they would be visited by an observer, It was

not possible to meet the other conditions ade-quately. It would have be most difficult to keep

the teachers from knowing that the observers werefrom the U.S. Commission on Civil Rights. Further-

more, the observers were of three different ethnic

groups. The ethnic identity and employment affil.

iation of the observers undoubtedly had the effect

of their being perceived negatively by some ofthe teachers, Moreover, the plan to visit a largenumber of classrooms precluded praeliminary visits

which would have allowed the observer to become familiar to the class.

Obtrusive biases are difficult to detect duringactual classroom observation. This is because, on

the one hand, they are confounded with codingbiases and, on the other hand, checks for them

usually can be expected to induce some obtrusiveeffects of their own.

No specific checks were made for obtrusivebiases: However, it is possible to make a reason-

able prediction of the direction of these biases.The general effect of any .classroom observer is

usually to put both the teacher and the students

on good behavior. The teacher tries to fulfill whatis generally considered characteristic of goodteachers: in control of the class, well organized,

involving the students, and supportive of theirefforts. The fact that the observers in this studywere known to the teacher to be from a civilrights agency and the fact that some of them were

obviously minority group members inust certainly

have induced some teachers to be\ particularlycareful about how they related to Mority stu-dents in the classroom. Consequently the observed

unfavorable disparities between Mexican Ameri-cans and Anglo students are probably soMewhat

less severe than the disparities that normall\pccur

in theclassroom,1

Possible Sampling Biases

There is no way to totally assure against sam-.

piing bias, but adherence to statistically sound\sampling procedures does allow one to limit theprobability of a significant sampling bias to a given \

value: The sampling procedure used in this study \has been discussed. (See Appendix A, pp. 45-47,)

These procedures followed or approximated thesampling procedures necessary for valid statistical

inference. The probability of sampling bias in this

stuj:. is one in a hundred for each tested hypoth-esis. It should be doted that this probability oferror is in reference only to inferences about the

specific population from which the sample was

drawn.

School and District No

Access No,

District Name

Teacher Name

APPENDIX F. TEACHER, STUDENT, AND CLASS-

ROOM CHARACTERISTICS INFORMATION

FORM

Classroom No.

1. Use the space on page T,2 as follows:a. On the seating chart on the following page,

record the seating patterns in the class-room, In each block record:

NENo seatNONot occupiedMMexican AmericanAAngloBBlack0Other

b. Draw the teacher's desk, if one is present.c. If 50% or more of the children face in one

direction, indicate the'focal point.d. Draw any seats occupied by isolated chil-

dren outside the main body, and indicateCM, A, B, 0) the ethnicity of the child inthat seat.

e. Draw any windows.1, if any of the items called for in directions

b - e are not ,sent, indicate this fact bydrawing a line through that direction.

2. Ask the teacher: Please tell me what collegedegree or degrees you hold, Record the an-

swer below.

Doctorate Masters Bachelors None

Ask the teacher: Have you had any in-service

courses dealing with the education of the

Date

Interviewers initials

School Name

AG

Grade Period

Mexican American child in relation to hiscultural and linguistic differences? If so, de-scribe.

4, Ask the teacher the following question, Re-cord the answers in the grid below:

Please help me to identify the ethnicity ofthe children in the room today. How manychildren in the room are Mexican American,Anglo, Black, or of other ethnic background?

You should use "other" for children of Ori-ental or American Indian background.

Your Classification Teacher's Classification

5, Ask the teacher: How do you decide where 7, indicate the age of the teacher in the spaceeach child will sit? Record the answer below: below.

6, Indicate the ethnicity of the teacher in thespace below,

B

20's 30's 40's 50's

8, Indicate the sex of the teacher in the space

below,

SEATING CHART

District:

School:

Address:

ContactPerson:

Address:

Phone No.:

APPENDIX G. SOCIOECONOMIC STATUS OF

FAMILIES BY ETHNIC GROUP COLLECTION

FORM

U.S. Commission on Civil Rights

Field Study Supplement

1. What percent of the Spanish Surnamed pupils

in this school come from families with a total

annual income of: (Estimate.)

A. Below $3,000?_ B. Over $10,000?

2. What percent of the Anglo pupils in this school

come from families with a total annual in-come of: (Estimate,)

A. Below $3,000?_ B. Over $10,000?____

3. What percent of Black pupils in this school

come from families with a total annual income of: (Estimate.)

A. Below $3,000. _ B. Over S10,000?

4. What percent of the Other pupils in this

school come from families with a total annualincome of (Estimate.)

A. Below $3,000? _ B. Over $10,000?--

5 What percent of the Spanish Surnamed pupils

in this school come from families in which

the highest educational attainment level of

the head of the household is: (Estimate.)

A. 0-8 years?

B. High School?

C. College?

D. Total 100 %

6. What percent of the Anglo pupils in this

school come from families in which the high-

est educational attainment level of the head

of the household is: (Estimate.)

A. 0=8 years? /0

B. High School?

C. College? %

D. Total 100 %

7. What percent of the Black pupils in this

school come from families in which the high-

est educational attainment level of the head

of the household is (Estimate.)

A. 0-8 years?

B. High School?

C. College?

D. Total

8. What percent of the Other pupil in this school

come from families in which the highest edu-

cational attainment level of the head of thehousehold is (Estimate.)

A. 0-8 years? %

B. High School? %

C. College?

D. Total 100 %

APPE

ND

IX H

. DISPL

AY

TA

BL

ES

H-1 Percent cat T

otal Variance of the B

ivariate Relationship

for those Pairs of Teacher, C

lassroom and Scho

which had a Statistically Significant R

elationship'.2II C

harac,

1. grade2. track

2.833. subject

3.57'4. ;teacher education.5. teacher inservice

trainiEng

6, teaclher ethnicity7. teacher age

2.998. ;teacher 'S

ex8.00

9. ;seating .criterion9.87

10. school '1%). M

A.

3,152.62

11.. school 1'0 A3.06

2.52'

12. MA

;seating index13. A

. seating; index'14. ;class '`I M

A.

4.5;3;

15. .class % A

.4.47

16. class size17. ethnic isolation in

;9.728.45

school18. schooll M

A. S

.ES

11.844.99

19. school A. 5E

5;4.18

2..0.72.77

29. s.chooll5E

5;14.26

21. stale4.24

3.37

12

3.4

5to

78

3.99

3.682.40

4.343.48

2.152.90.

4.532.9.3

3.10

2.53

3.492.35

1011

12

81..32

30.64;.33..87

32.25:39.19

6n_54

2.4756.55

4.646.07

6,02,2.49

5.1.5;

3.082.22,

4.632.3.54

26.59;10.95

12.289.43

1,782.13

4,3722.62

23.668.50

9.32'1.65

20.732.88

7.463.18

47.6347.69

111.2319.07

2,036.76

44.6928.59

2.733.44

7.493.04

1.881.84

13.1715.66

4.948.54

1314

1516

1718

1920

21

Tested by -C

hi So lest.

One ;of the 22 ch.aracteristics, Ilh.e difference hi the C

hicano andA

nglo averat; ew

ast. Included in this anahesis,..

H-2 Standard Deviations of Disparities in Be-!leviers Within: Classrooms Between theAverage Individual Mexican American -.andthe Average- Individual -Anglo

Teacher acceptance of students' feelingsTeacher praise or encouragement

Teacher acceptance or we of students'ideas

Teacher questioningTeacher lecturing

andardDeviation

StandardDeviation

.048 Teacher giving of directions .420

.344 Teacher criticizing or justifying authorityStudent response speaking

.2231.839

Student initiated speaking 2.6.393 Teacher giving of positive feedback.69491

.955 All noncriticizing teacher talk 2.4201.393 All student speaking 3.514

All Pairwise Correlations for the Measure Behavior Disparities Within Classrooms,

9 10 11

45.

6,

8,

9.

10.

Teacher acceptance of stu-dents' feelingsTeacher praise or encour- .05agementTeacher acceptance or use -.01 .54of students' ideasTeacher questioning .06 .59 .58Teacher lecturing .10 .21 .15TeaCher giving of direr-. -.03 .10 .09 .15 .21tionsTeacher criticizing .01 .01 -.04Student response speaking .04 .15 .19Student initiated speaking .04 .15 .14All positive feedback from .10 .86 .89teacherAll noncriticizing teacher .10 .60 .57 .77 .78 .38 .05 .28talk

/All student speaking .05 .20 . 22 27 .21 .12 .05

.02 .05 .19.30 .16 -.12 -.00.13 .15 .07 .06 .02.67 .21 .11 -.02 .20 .17

eported istics are Pcirson currefatlan ccetficicnts

.54 .24