REPORT RESUMESED 015 110 24 RE 000 998"MOTIVATED LEARNING" READING TREATMENT WITH ADDITIONALSUBJECTS AND INSTRUCTIONAL- TECHNICIANS.

STAATS, ARTHUR W. AND OTHERSWISCONSIN UNIV., MADISONREPORT NUMBER TR22 PUB DATE MAR 67REPORT NUMBER BR50216-111-22CONTRACT OEC510-154EDRS PRICE MF--$0.25 HC -$0.92 21P.

DESCRIPTORS... *REMEDIAL READING, *JUNIOR HIGH SCHOOL STUDENTS'*READING RESEARCH, *REINFORCEMENT, *SUBPROFESSIONALSI READINGACHIEVEMENT, WORD RECOGNITION, REINFORCERS, POSITIVEREINFORCEMENT' READING MATERIALS, UNIVERSITY OF WISCONSIN,MADISON,

A METHOD OF TREATING READING DEFICITS BASED UPON ANEXTRINSIC MOTIVATIONAL SYSTEM PREVIOUSLY EMPLOYEDSUCCESSFULLY WITH A SINGLE SUBJECT WAS EXTENDED TO 18ADDITIONAL SUBJECTS. THE JUNIOR HIGH AGE SUBJECTS INCLUDEDRETARDED CHILDREN IN SPECIAL CLASSES, SEVERAL EMOTIONALLYDISTURBED CHILDREN, AND CULTURALLY DEPRIVED CHILDREN. THEMETHOD OF TRAINING USED IN THE ORIGINAL STUDY WAS SIMPLE TOADMINISTER. IT WAS HYPOTHESIZED THAT SUBPROFESSIONALPERSONNEL COULD BE EMPLOYED TO ADMINISTER THE TREATMENT.ADULT VOLUNTEERS AND HIGH SCHOOL SENIORS WERE USED AS THEINSTRUCTIONAL TECHNICIANS. THE 18 SUBJECTS WERE GIVEN 38.2HOURS OF TRAINING IN DAILY HALF -HOUR SESSIONS DURING WHICHPERIOD THE AVERAGE REINFORCEMENT EARNED WAS $22.29. THE MEANNUMBER OF SINGLE WORD READING RESPONSES WAS 94,425. THE RATEOF READING ACCELERATED OVER THE PERIOD OF TRAINING, EVENTHOUGH THE READING MATERIAL BECAME MORE DIFFICULT. THISOCCURRED DURING A PERFOD WHEN PROGRESSIVELY LESSREINFORCEMENT WAS GIVEN PER READING RESPONSE. A MEAN OF 616.8NEW WORDS WAS LEARNED, AND 70.9 PERCENT OF THESE WAS RETAINEDIN A LONG TERM TEST. THE ATTENTION, ATTENDANCE, COOPERATION,AND DILIGENT WORK BEHAVIOR OF THE VARIOUS CHILDREN WEREMAINTAINED IN GOOD STRENGTH THROUGHOUT THE DURATION OF THESTUDY. (AUTHOR)

6 Ia f

a # II

a

e

a

I

a

a

a

U.S DEPARTMENTOE HEALTH.

EDUCATION & WELFAREOffICE OfEDUCATION

THIS DOCUMENTHAS BEEN

REPRODUCED EXACTLY ASRECEIVED IRON THE

PERSON ORORGANIZATION

ORIGINATING IT POINTS OF VIEWOR OPINIONS

STATED DO NOTNECESSARILY REPRESENT

OFFICIAL OFFICE Of EDUCATIONPOSITION OR POLICY.

WI

.I

.

k

s a

to lb.

a

Technical Report No. 22

"MOTIVATED LEARNING" READING TREATMENT

WITH ADDITIONAL SUBJECTS AND INSTRUCTIONAL-TECHNICIANS

Arthur W. Staats, Karl A. Minke,

William L. Goodwin, and Julie Landeen

Wisconsin Research and DevelopmentCenter for Cognitive LearningThe University of Wisconsin

Madison, Wisconsin

March 1967

The research reported herein was performed pursuant to a contract with the United StatesOffice of Education, Department of Health, Education, and Welfare, under the provisions of

the Cooperative Research Program.Center No. C-03 / Contract OE 5-10-154

This study was made possible by the help and cooperation of Robert D. Gilberts, Superinten-dent of Madison Public Schools, and the principals of the schools involved: Homer Winger,Douglas Muller, Jack R. Stickel s, George Maki, and Roger Cerutti.The first author is now on leave of absence at the University of Hawaii.

../*AMOIAIIMINGON,..

PREFACE

The Wisconsin R & D Center is designed to contribute to an understandingof, and the improvement of educational practices related to, cognitive learningby children and youth. Of primary concern in attaining the goal of the Centerare the learning of concepts and the nurturing of related cognitive skills. Con-ditions within the learner, such as motivation or cognitive organization, andconditions within the learning situation, such as the content and sequence ofinstruction, are also relevant areas of research.

In this technical report is described the successful extension of learningprinciples to complex human behavior, namely reading behaviors, of highlyretarded readers. The instructional program and procedures developed byStaats were administered to junior high school students by previously untrainedpersonnel. Implications for the treatment of behavior problems in clinicalpsychology as well as in education are drawn.

Herbert J. KlausmeierCo-Director for Research

iii

CONTENTS

page

List of Figures viiAbstract ix

I. Introduction 1

II. Method 3

Subjects 3Experimenters 3Instruments 3Reinforcer System 4Reading Materials 4

Vocabulary Words 5Oral Reading Materials 5Silent-Reading and Comprehensive-Question

Materials 5Procedure 5

Vocabulary Presentation 5Oral Reading 5Silent Reading and Comprehensive Questions 6Vocabulary Review 6

III. Results and Conclusions 7

IV. Discussion 12

References 15

v

LIST OF FIGURES

Figure page

1

2

3

Cumulative number of reading responses as a functionof the time in reading training for low, middle, andhigh initial reading ability groups

The ratio of words missed upon initial presentation tothe total number of words presented in the IndividualWord Phase as a function of the number of SRA storiesread for the three ability groups

Ratio of the monetary value of the tokens receiveddivided by the number of reading responses made asa function of time in reading

8

9

10

vii

ABSTRACT

A method of treating reading deficits based upon an extrinsic motivationalsystem previously employed successfully with a single subject was extended to18 additional subjects. The junior high age subjects included retarded childrenin special classes, several emotionally disturbed children, and culturally de-prived chilch.en. The method of training used in the original study was designedto be simple to administer and simple to record the performance of the child.Thus, was hypothesized that subprofessional personnel could be employed toadminister the treatment. Adult volunteers and high school seniors were usedas the instructional-technicians.

The 18 subjects were given 38.2 hours of training, in daily half-hour ses-sions, during which period the average reinforcement earned was $22. 29. Themean number of single word reading responses was 94,425. The rate of readingaccelerated over the period of training even though the reading material be-came more difficult. This occurred during a period when progressively less(about one-fourth as much) reinforcement was given per reading response. Amean of 616.8 new words were learned and 70.9% of these were retained in along term test. The attention, attendance, cooperation, and diligent work be-havior of the various children were maintained in good strength throughout theduration of the study.

ix

OE 6000 (REv. 9-66)

(TOP)

001

100101

102103

200

300310

320330

34.0

350400

500501

600601602603604605606

607

800801

802

803804805

806

807

808809

810811

812

813

814.

815

816

817

818 .

819

820

821

822

DEPARTMENT OF HEALTH. EDUCATION, AND WELFARE

ERIC ACCESSION NO. Uhrft.t. OF FAJOC.AIIUN

ERIC REPORT RESUME

Ii NO 13

IZI NO 0

CLEARINGHOUSEA C C E S S I O N NUMBER RE DATE.

05- ir 67

P, A. IS DOCUMENT COPYRIGHTED?T.A. I yes

ERIC REPRODUE TION RELEASE' YESTITLE

"Motivated learning" reading treatment with additional subjects and instruc-tional-technicians

PERSONAL AUTHOR(S)

Staats, Arthur W. and others

INSTITUTION (SOURCE) University of Wisconsin, Madison, Wisconsin, - el - , :,- . $ .1t 1- - ill'I " 4 ' $

SOURCE CODE

R E.: PORT/ Sattal tc1:agsbnisailleasutAsuuSOURCE CODEOTHER SOURCE

OTHER REPORT NO.OTHER SOURCE SOURCE CODE

OTHER REPORT NO./...........

PU Sq.. DATE te r 67 CONTRACT /GRANT NUMBER G.A13/enntrartM 5-in-16APAGINATION. ETC.

15 pp.

RETRIEVAL TERMS motivated learningremedial readingextrinsic reinforcementmonetary reward

IDENTI TIERS

ABSTRACT This report describes a method of treating reading deficits based upon anextrinsic motivational system. The Ss were 18 junior high students includingretarded children in special classes, several emotionally disturbed children, andculturally deprived children. The method of training was designed to be simpleenough that subprofessional personnel could be employed to administer the treatment.Ault volunteers and high school seniors were used as the instructional-technicians.Ss were given 38.2 hours of training, in daily half-hour sessions, during whichperiod the average reinforcement earned was $22.29. The mean number of singleword reading responses was 94,425. The rate of reading accelerated over the periodof training even though the reading material became more difficult. This occurredduring a period when progressively less (about one-fourth as much) reinforcementas given per reading response. A mean of 616.8 new word's were learned and 70.9%

of these were retained in a long term test. The attention, attendance, cooperation,and diligent work behavior of the various children were maintained in good strengththroughout the duration of the study.

DP)

0001

02

01

OF: 6000 'REV. !3-.31

acccssioN: tiUsii;.LFt ri -,A- r ; p. A, ; 17.Z7

; 1:R tr: A CCF.5S!ON NO.;

1

"Motivated learning" reading treatment with additional subjects and instruc-tional-technicians

35 r 67 i

p Avo-rm I:: L. 73-1 . A itcN, AND

FRI.0 R::PORT RESUME

IS DOCUMENT COPYRiGHTED,

ERIC. qriPiRODUCTiON Rrit.EASE? YCS 3Y ES 0

ts1 Ea Ndtp.1

03

DO

30

10

2030

4.0

50

)0

)0

)1

7

0a23

55

73

3

i.

5

7

3.*s

_pe:HsoNAL A,..rt-,0H,s

Staats, Arthur W. and others.............................--....:--i ,NI,r,-L,TIoN 50,:pcy.)

University of Wisconsin, Madison, Wisconsin ""c"°'.'LIU, ac Qnf.; ta_a_e_ s e,a.,:ja_ami_Dgskaapi. lull_ r;tiat..ex, ...j.42:sagaisauz_LearR;;PorT/SERMS NO. Technical Zeport No. 22oTNE:Ft SOURCC

SOUREL CoosI

0 ri-11,p Rcpw-fr ND........40.1...*4.OTHrri 3OURCL.

SOURCE: COCU;

! * ..........rw.........moo.vm...' Mb......00.....I.M.NOwlWort.......,2 2:4

I 141: P010 :. NC).

1'' PU (4.i... CAA Ti; -... mar. ,L.L5221.1122...ceiata. r......uadi_czaaLccur.

s I-., PAGo.A110N, c'c,i

I

15 p p .

I, L Tritt.VA:. .C.:-V.,5 motivated learningremedial reading

Iextrinsic reinforcement

Imonetary reward

i

1

,

We.W. LW.M1IpENTI PIERS

......................-.....---AEITRAET This report describes a method of treating reading deficits based upon an_extrinsic motivational system. The Ss were 18 junior high students includingretarded children in special classes, several emotionally disturbed children, andculturally deprived children. The method of training was designed to be simpleenough that 3ubprofessional personnel could be employed to administer the treatment.Adult volunteers and high school seniors were used as the instructional-technicians.iSs were given 38.2 hours of training, in daily half-hour sessions, during whichperiod the average reinforcement earned was $22.29. The mean number of singleivrd reading responses was 94,425. The rate of reading accelerated over the periodof training even though the reading material became more difficult. This occurred'during a period when progressively less (about one-fourth as much) reinforcementwas given per reading response. A mean of 616.8 new words were learned and 70.9%of these were retained in a long term test. The attention, attendance, cooperation,'and diligent work behavior of the various children were maintained in good strengththroughout the duration of the study.

DOCUMENT FILMED FR

I

IN1RODUCTION

The first author (see Staats, 1965a, 1965b,in press; Staats and Staats, 1963)has suggest-ed a general learning theory of human be-havior called an integrated-functional learningapproach. An essential feature of the approachis that its principles and methods must becapable of dealing with functional human be-haviors and contribute to the treatment of actu-al problems of human behavior. One aspectof this has been the formulation of a theory oflanguage learning and cognitive development(Staats, in press)including a theoretical andexperimental analysis of reading acquisition.This has involved laboratory studies of basiclearning principles in the context of the corn-plex behavior of reading acquisition (Staats,Staats, Schutz, and Wolf, 1962; Staats, Finley,Minke, and Wolf, 1964; and Staats, Minke,Finley, Wolf, and Brooks, 1964).

In addition, on the basis of the previousfindings and analyses, a method was devisedfor the treatment of children who because oflearning problems have not been able to readnormally. The method, which involved a sys-tem of extrinsic reinforcement employing to-kens, was tested in a study on a 14-year-old,culturally-d ep r iv ed, juvenile delinquent(Staats and Butterfield, 1965). The study,involving 40 hours of training conducted overa 4-1/2 month period, was extraordinarilysuccessful. The child, who was a severe be-havior problem in a traditional classroomsituation, worked and attended well in the ex-perimental training program, he made over65,000 word reading responses, he receivedspecial training on 761 words he did not know,he retained 430 of those words (57%), and hepassed all his courses in school. (the first timehe had ever passed a course in his whole his-tory), his misbehaviors in school fell off mark-edly, and his general attitudes towards schooland school work appeared to improve. Inaddition, achievement t e sts showed an in-crease from the 2.0 grade level in reading tothe 4.3 level.

One of the major points of the treatmentprocedure, however, was the fact that it wasdesigned to be easy to administer and record.This was done to solve the practical problemsinvolved in its wide-spread use for treatmentand for research. Thus, in the first studythe procedure was applIed by a person un-trained in teaching, a probation officer. Thishas many implications for education and clin-ical psychology.

The procedures are very specific andrelatively simple. Thus, it was not neces-sary tohave a person highly trained in ed-ucation to administer the training . . . .

It might be suggested that anyone with ahigh s c h o o 1 education and the ability toread could have administered the training.ihishas implications for the practical ap-plication of the present methods . . . [Staatsand Butterfield, 1965, p. 939].

It was also suggested that the procedurecould be applied in many settings, involvingvarious types of clinical, educational, andsocial problems. For examples it was sug-gested that helper,- of teacherswho will becalled instructional-technicians hereincouldapply the methods in special education train-ing programs.

This study was based upon the results ofa single subject, however. The next step inthis program of research was a more generaltest of the procedure. In addition, the presentstudy b:-gins to assess the suggestion of thefirst study that learn ing problems can betreated by sub-professionals; in this case lit-erate high school seniors and adult volunteerswith high school educations serve as the in-structional-technicians. The 18 children whoserved as experimental subjects includedunderachieving junior high school aged chil-dren with no special characteristics excepttheir learning problem, children considered

1

to be mentally retarded and in special claw2.s,and several children with emotional or be-havioral problems. Nine of the Ss each had

(

a

one of 9 high school seniors a s the instruc-tional-technician, and the other Ss each hadone of 9 adult volunteers as the instrr tional-technic ian.

METHOD

SUBJECTS

Eighteen experimental and 18 control Sswere selected for participation in the projectin the following manner: Lists of studentsdefined as poor readers on the basis of stand-ard achievement tests and enrolled in 7th or8th grade or in Special Classes were obtainedfrom five junior high schools in Madison,Wisconsin. These lists were then circulatedto the teacher s of the various schools with theinstructions to delete any names they felt werenot reading problems and to add any studentsto the list whom they felt had been omitted.Final selection of Ss for participation in thestudy was done on the basis of a 100-itemword recognition test developed from the read-ing material. The 36 Ss selected (28 maleand 8 female) were those most deficient onthis oral test, reading 81 or fewer of the wordscontained. Subjects were grouped in thirds,from low to high, on the basis of their scoreson the or al reading test, and subsequentlydivided into 18 pairs, matched on the basisof their score on this word recognition testand, where possible, on the basis of IQ scores.The 18 Ss for the experimental group wereobtained by randomly selecting one of the twostudents from each matched pair.

The average age at the start of the experi-ment was 14 years, 6 months; the average IQwas 74 (69 Verbal and 80 Non-verbal). Sevenof the experimental Ss and seven control Sswere enrolled in c l a s s e s for the educablementally retarded. A number of Ss wereclassified as educationally disadvantaged un-der provisions of the new rederal EducationAct.

EXPERIMENTERS

Two types of experimenters were utilizedin this study. Nine adult volunteers were ob-tained through the P. T. A.'s of the participatingschools and through the efforts of the indi-

vidual principals. These volunteers were paid$2.00 a session to cover transportation, baby-sitting fees, etc. The adult volunteers wereall housewives with an average of 4 childrenliving at home. Their average age was 42,and their families earned approximately $7,500per year.

In addition, nine high school seniors wereobtained from East Senior High School andCentral Senior High School. These studentswere selected by first circulating a list of thesenior class among the teachers at the highschool and having the teachers give the stu-dents a combined rating o f above average,average, or below average on the character-istic s of maturity, responsibility, and accept-able reading level. Students receiving a be-low average rating from any teacher wereeliminated, and the remainder served as thepool from which volunteers were sought.These Es were paid $1. 25 per 1/2-hour ses-sion, and participated in the study 5 timesper week during their regularly scheduledstudy halls. The seniors had a mean gradeaverage of B for their 4 years inhigh school.

The instructional- Technicians were trainedin the use of the procedures in two 1-1/2 hoursessions just pr ior to the beginning of theproject. This training included some experi-ence in actually administering the materials.In addition, a brief third training session washeld at the end of the second week to generallyinstruct the instructional-technicians on anerror in procedure committed by one of theadult volunteers and to indicate some minorchanges in procedure.

INSTRUMENTS

The primary test used, both for groupingpurposes and as a dependent variable, con-sisted of 100 words randomly selected fromthe reading materials used in the experimentalsessions consisting of 20 words selected fromeach of five grade levels (1.2, 1. 7, 2. 3, 3.0,

3

and 4.0). Subjects had to correctly pronouncethe word shown them on a flash card in orderto receive credit. The pretest was given be-fore assignment to treatment condition, andthe posttest was administered by persons un-familiar with the experiment.

Other tests, primarily of the standardizedvariety, were given during the screening ofsubjects, specifically the Iowa Test of BasicSkills, Grade 7, Form 2, and the Lorge-Thorndike Intelligence Test, Level D, Form1. It was planned to give alternate forms ofthese tests as posttests so the Ss' pre- andpost-experimental performance could be com-pared. Many subjects, however, as will bediscussed further on, appeared to be answer-ing some of their tests randomly, and theirchance scores substantiated this suspicion.

The same result was shown on the posttestwhere scores on the reading and vocabularysubtests of the Iowa Test of Basic Skills,Grade 7, Form 3, made it apparent thatmanyof the subjects were still operating at chanceexpectancy. Accordingly, four additionalsessions were utilized in which easier tests(the Lorge-Thorndike Intelligence Test, LevelC, Form 2; and the California ReadingAchievement Test, Elementary, Grades 4-6,Form W; and Forms 1 and X of the same twotests, respectively) were administered. Aspart of the attempt to salvage the test data,the extrinsic reinforcement method was ex-tended to the testing situation during the lasttwo sessions when the alternate forms weregiven. The alternate forms were thus givenwith the addition of the following reinforce-ment contingency instruction:

The test you are to take now involves a p ro-cedur e you should understand thoroughly.You will be able to earn money for doingthis testthatis, byworking hard, readingeach item, and not guessing on items youdo not know. You will get two cents forevery question that you answer correctlyso the more questions you answer correctlythe more you will make. However, youwill lose one penny for every incorrectanswer. Thus, if you do not read questions,and just guess at the answers, you will endup making nothing. Remember you get twocents for each question you answer cor-rectly, and lose one cent for each incor-rect answer.

Tests were scored, and students paid by en-velopes a few minutes after these two con-tingency test sessions ended.

4

REINFORCER SYSTEM

There were three types of token, distin-guished by color. The tokens were of diffe rentvalue in terms of the items for which the to-kens could be exchanged. Initially, a bluetoken was valued at 1/5 of one cent, a yellowtoken at 1/2 of a cent, and a red token at 1cent; however, at the end of the first week thetokens were devalued, and for the remainderof the study the value of the blue, yellow, andred tokens was 1/10,1/5, and 1/2 of one centrespectively as in the original study.

During th e eleventh week of the study abonus system was introduced for a few slowchildren in order to maintain a $. 20 minimumin terms of Ws daily earnings. From thispoint on, whenever S's earnings fell below$. 20 for any one session, he was given a bonusof $.05, $.10, or $.15, whichever amountwas needed in order tobring his earnings forthat day up to between $. 20 and $. 25. Whenthe bonus system was introduced, the childwas told that he could not move as rapidly asbefore because the material upon which he wasworking was difficult, but if he continued toworkhardhewouldbe given a bonus at the endof the session.

The child' s acquisition of tokens was plottedso that visual.evidence of the reinforcers wasavailable. The tokens could be used to pur-chase a variety of items. These items, chosenby the subject, could range in value from pen-nies to whatever the subject wished to workfor. Rec'rds were kept of the tokens earnedby S and of the manner in which the tokenswere used.

READING MATERIALS

The reading material, as in the earlierstudy (Staats and Butterfield, 1965), was takenfrom the Science Research Associates (SRA)reading-kit materials, Reading LaboratoriesIA, IB, IC, and IIA. The SRA kits consist ofstories developed for and grouped into gradelevels. For the purposes of this study, storieswere taken from the 1.2 (20 stories), 1.4 (60stories), 1.7 (20 stories), 2.0 132 stories),2.3 (12 stories), 2.6 (12 stories), 3.0 (32stories), 3.5 (32 stories), and 4.0 (10 stories)grade levels. Once a particular Reading Lab-oratory was selected for inclusion at a givengrade level, all the stories at that grade levelwere presented in sequential order (with theexception of the 4.0 grade level, where onlythe first 10 stories in Laboratc IIA were

presented). The different numbers of lessonspresented at each grade level were due to theuse of different numbers of Reading Labora-tories, in order to control somewhat the rateof introduction of new words.

Each story includes a series of questionswhich can be used to assess the reader'scomprehension of the story. The readingtraining program which again was that of theearlier study may be summarized as follows.

Vocabulary Words

A running list was made of the new wordsthat appeared in the series of stories. Eachdifferent form of a word was counted as a dif-ferent word for this purpose; thus, bring,brings, and bringing were all counted as dif-ferent words. The list finally included eachdifferent word that appeared in the stories, atotal of 4,253 words. From this list, the newvocabulary for each story was selected, andeach word was typed on a separate 3 x 5 card.

The average number of new words intro-duced in each story was 18.5, the least num-ber of new words being introduced in any onestory being 6, and the most being 50. Thefew stories presenting over 30 new wordsoccurred in the second half of the program,after the children had already been presenteda large number of lessons.

Oral Reading Materials

Each paragraph in the SRA stories wastyped on a 5 x 8 card. Each story could thusbe presented to S paragraph by paragraph.

Silent-Reading and Comprehensive-Question Materials

Each SRA story and its comprehensivequestions were typed on separate 8-1/2 x 11sheets of white paper.

PROCEDURE

Vocabulary Presentation

The procedure for each story in the seriescommenced with the presentation of the newwords introduced in that story. The wordswere presented individually on the cards, andS was asked to pronounce them. A correctresponse to a word-stimulus card was even-tually reinforced with a mide.value token. Aftera correct response to a word, the card wasdropped from the g r o up of cards yet to be

presented. The S was instructed to indicatewords that he did not know the meaning of, andthis information was provided in such cases.

When an inc orrect response to a wordstimulus occurred, or when S gave no re-sponse, the instructional-technician gave thecorrect response. The S then repeated theword while looking at the stimulus word. How-ever, the word card involved was returned tothe group of cards still to be presented. Acard was not dropped from the group until itwas read correctly without prompting. Afteran error on a word stimulus, only a low-valuetoken was given for finally reading the wordcorrectly without prompting. The vocabulary -presentation phase of the training was con-tinued until each word was read correctlywithout prompting.

Initially, the appropriate tokens were de-livered immediately contingent upon each un-prompted correct reading trial. However, bythe end of the second week it was apparentthat the instructional-technicians were havingdifficulty presenting materials, watching S,keeping data, and delivering tokens all at thesame time. So at the beginning of the thirdweek, all tokens were delivered at the end ofeach phase r ath e r than after each readingresponse.

Oral Reading

Upon completion of the vocabulary mate-rials, each paragraph was individually pre-sented to S in the order in which the paragraphoccurred in the story. When correct readingresponses were made to each word in theparagraph, a high-value token was given uponcompletion of the paragraph. When a para-graph contained errors, S was corrected, andhe repeated the word correctly while lookingat the word. The paragraph was put aside,and when the other paragraphs had been com-pleted, the paragraph containing errors wasagainpresented. The paragraph was repeateduntil it was read correctly in its entiretyat which time a mid-value token was pre-sented. When all paragraphs in a story hadbeen completed correctly, the next phase ofthe training was begun.

When a few of the Ss had some difficulty inreading the words in sequential order, eitherfailing to respond to one or more words in thesequence, or adding words for which no tex-tual stimuli were presented, the instructional -technician would point to each word in turn andS would then read them. It was generally pos-sible to later drop this additional procedure

5

without any loss in the c o nt r o 1 over Ss'performance.

Silent Reading and Comprehensive Questions

Following the oral reading S was given thesheet containing the story. He was instructedto read the story silently, and he was told thatit was important to read to understand thestory so that he could answer the questionswhich would be presented later.

Initially, S was presented reinforcementin the form of a yellow token on a 15-secondvariable interval schedule. That is, as longas S appropriately scanned the material hewas given a mid-value token an average ofevery 15 seconds. However, tilare was a greatdeal of variability in the abilities of the in-structional-technicians to administer this typeof schedule, and so at the beginning of thethird week the procedure was changed such thatfour yellow tokens were delivered at the endof the silent reading of the story, regardlessof the time needed to complete the task.

Upon completion of the story, S wrote hisanswers to the questions typed on a separatesheet and gave his answers to the instructional-technician. For each correct answer, S re-ceived a high-value token. For an answerwith a spelling error, S was reinforced witha mid-value token when he had corrected theanswer. For incorrect answers S had to re-

6

read the appropriate paragraph, correct hisanswer, and he then received a low-valuetoken.

Vocabulary Review

Some of the vocabulary words presented toS in the first phase of training were words healready could read. Many others, however,were words that the procedure was set up toteach. The oral-reading-phase performanceindicated the level of S's retention of the wordshe had learned and also provided furthertraining trials on the words not already learn-ed. A further assessment of S's retention ofthe words that he did not know in the vocabu-lary training was made after each 20 storiesof the SRA materials had been read. Thistest of individually presented words, for eachstory, was started immediately after com-pletion of the 20 stories and constituted fairlylong-term retention.

This test was also used as a review for S,and further training on the words was given.When S could not read a word, or missed one,he was prompted and had to correctly repeatthe name of the word while looking at the word.This word card was then put aside and pre-sented later, at which time S was reinforcedwith a low-value token if he read it correct-ly. If not, the procedure was repeated untila correct unprompted trial occurred.

III

RESULTS AND CONCLUSIONS

Each of the 18 subjects participated in thestudyover the period of the study with excel-lent cooperation, attention, and work behav-iors. In the original study, the juvenile de-linquent subject made about 65,000 singleword reading responses in 40 hours of train-ing. The mean number of hours of trainingin the present study was 38.2. During thisperiod the mean number of word reading re-sponses made by the subjects was 94, 425.

The present sample of subjects includedchildren with varying degrees of initial read-ing skill, as indicated by their ability to readthe 100 individual words taken from the SRAmaterials in the way that has been described.Thus, the percentage of these words that couldbe read by the children prior to the trainingvaried from a low of 22 percent to a high of81 percent. In order to better see how thetraining affected the subjects as a function oftheir initial reading ability some of the re-sults to be presented will be grouped accord-ing to the level of the children's ability. Thus,the 18 Ss were grouped into 3 subgroups. Thelowest had a mean percentage of reading theSRA words of 43. 2 prior to the study. Themean IQ of these chi l dr e n on the Lorge-Thorndike Intelligence Test was 71.9, rang-ing from 50 to 91. The middle group had amean percentage of 69. 2 on the SRA words,and a mean IQ of 74. 2, with a range from 53to 94. The high group had a mean percentageof 79.0 on the SRA words, and a mean IQ of79. 9, with a range of 66 to 92.

The total number of words the child readduring the training was tabulated. This in-cluded the words he read when singly pre-sented on cards, the words he read aloud inthe paragraph reading phase, the words heread in the silent reading phase, the wordsread in the comprehensive questions, as wellas the words read in the vocabulary review.The cumulative record of the words read overthe period of training for the three groups ofSs is given in Figure 1. In the original studyof Staats and Butterfield the record produced

indicated that the rate of reading acceleratedslightly during the period of trainingeventhough the reading materials became steadilymore difficult. A similar result is shown forthe Ss in the present study. (The later pointson the curves, where indicated, do not includeall the Ss in the group, since some Ss did notparticipate as many hours as others did. ) The6 Ss with the highest reading ability show thegreatest acceleration, the middle group thenext h igh e s t, and the low group the least.This acceleration takes place even though thechildren advance from 1. 2 grade reading ma-terials to 3.5 grade level materials. It doesappear, however, that the increase in diffi-culty does dampen the acceleration in readingrate of all the Ss particularly those with lowreading ability. In any event, as in the orig-inal study, the children's behavior of attend-ing to the task and making the appropriatereading responses did not diminish throughoutthe period of training. Thus, the token rein-forcement system employed was capable ofmaintaining the work behaviors of these chil-dren for a long period of time. During thistime the attentional and cooperative behaviorsinstigated resulted in many, many, learningtrialsa necessary aspect for the acquisitionof achievement in any skill.

Records were kept of the number of wordsthe children missed on first presentation, thenumber of these words which were then latermissed in the oral reading of the paragraphs,as well as the number of the words originallymissed that the child could not read on thereview test presented at the later time. On theoriginal study the juvenile delinquent missed761 words on first presentation, 585 (o rabout 77%) of these were retained in the oralreading phase, and 430 (about 57%) were re-tained in the review test when presented singly.The corresponding mean words missed on firstpresentation for the low ability group in thepresent study was 820. 3. In the oral reading,686. 5 (or 83. 7%) were retained. In the long-term retention test, where the words were

7

1 30,000

120,000 -

1 10,000 -

100,000 -

914000 -

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

41011) High Group

Midd I e Group

Low Group

,/,//

5

I

I0

I

I5

I

20

HOURS

1

25

1

30

t

35

I

40

Figure 1. Cumulative number of reading responses as a function of the time in reading train-ing for low, middle, and high initial reading ability groups. The letter (a) on eachcurve represents the last point in which all Ss in that group are represented.

presented sing 1 y, 433.3 (57. 4%) were re-tained. (Since the vocabulary review occurredat the end of each 20 lessons, most Ss werepresented with a number of words in the lastfew sessions which were not tested in thismanner. ) The performance of this group ofSs was thus roughly comparable to that of thefir st subject. The middle ability group missed535 words on first pr es en tatio n, retained483.8 or 90. 4%) on second presentation, andretained 306. 2 (or 69.1%) on the long-termreview test. The means for the high abilitygroup were 425.2 firstmissed, 360.3 (or 84%)retained on second test, and 291.0 (or 78. 7%)on review test.

Evidence indicated that the 100-word testtaken from the SRA reading materials was avalid indicator of th e Ss' reading skill withthe SRA materials. That Is, the low, middle,andhighgroups on this measure (with respec-tive reading percentages of 43.2, 69.2, and79. 0) performed in that order of ability on thereading task itself. On the measure of the

8

percentage of the words in the various storiesthat could be read on first presentation themeans of the three groups were 67. 7%, 82. 1%,and 87. 8 %. In addition, similar results wereshown on the measure of long term retention,the vocabulary review. The percentages re-tained correctly on this measure for the threegroups were 57. 4%, 69. 1 %, and 78.7%.

In the original study there seemed to besome evidence that as the training prog re s sedthe subject missed fewer of the words on firstpresentation. This type of evidence wouldtend to indicate that the subject had been learn-ing to sound out new words as a function ofthe training. Additional evidence (Staats,1965b)has shown that subjects can learn syl-labic reading unit s from training on wholeword reading tasks. To test the possibilitythat improvement in sounding out new wordsoccurred in th e present study, the ratio ofwords missed the first time to the total pre-sented was computed for each 20 stories pre-sented. The results for the 3 groups are

shown in Figure 2. The later points on eachcurve do not include all the subjects in thegroup, since some Ss did not read as manystories as others did. In any event, theredoes not appear to be evidence that the ratiodecreased, but rather that it was constantthroughout the training except for an increasein the ratio from the 1.2 grade level to the

.4 grade level of the reading material. Thequestion of whether phonetic reading skillsemerge from this type of training may be con-sidered to remain unanswered. More directtest will probably be necessary.

One of the important aspects of the pro-cedures involves the ratio of reinforcementfor the reading responses. The procedureswere designed t o progressively reduce theamount of reinforcement givenper reading re-sponse as the training progressedor con-versely to require more reading responsesper unit of reinforcement. As was indicatedin the original study, demonstration that thisis possible on a long-term training programhas a number of important implications. Sucha demonstration "is in part an answer to thequestion whether th e use of extrinsic rein-forcers in training will produce a child who is

.40

.30

.20

.10

110---.. High Group0- 413 Middle Group

.---. Low Group

dependent upon these reinforcers" (Staats andButterfield, 1965, p. 941).

Figure 3 supports the original demonstra-tionby showing that the ratio of the amount ofreinforcement earned divided by the numberof words read decreases as a function of num-ber of training sessions. By the tirne 60 train-ing sessions have been completed the childrenwere receiving about one-fourth as much re-inforcement for their reading behavioras theyhad at the beginning of the training. This wastrue of the 3 groups. This result is especiallyinteresting in view of the acceleration of rateof reading response shown in Figure 1. Thatis, in the period of time during which the re-inforcement p e r reading response is cut inquarter, the rate of reading increases. Asidefrom its theoretical and practical implicationsfor child learning, it is interesting to note thatthe result would be expected directly frombasic laboratory research with lower organ-isms, on the effects of reinforcement sched-ules. Thus, it is important to see that thebasic principles hold even in the context ofsuch a complex and uniquely human behavioras reading.

20

. .

40 60 80 100

NUMBER OF LESSONS

120 140 160 180 200

Figure 2. The ratio of words missed upon initial presentation to the total number of wordspresented in the Individual Word Phase as a function of the number ofSRA storiesread for the three ability groups. The letter (a) on each curve represents the lastpoint in which all Ss in that group are represented.

9

.08

.07

.06 -

.05 -

.04 -

.03

.02

.01

41----III High Group

D. - -a Middle Group

------. Low Group

1b a

I I I I I I 1 1

5 10 15 20 25 30 35 40

HOURS

Figure 3. Ratio of the monetary value of the tokens received divided by the number of readingresponses made as a function of time in reading. The letter (a) on each abilitygroup curve represents the last point in which all Ss are represented, and the letter(b) indicates the time block in which the bonus was introduced for most Ss.

As has been described, the test data con-cerning reading achievement appeared to beinvalidated by the test-taking behavior of thechildren. In addition to that, midway throughthe experiment special reading programs werebegun under the Elementary and SecondaryEducationAct Title I so that the control groupcould no longer be considered as such. Theexperiment, however, had been designed sothat analyses of covariance could be made ofthe data. The main effects were the experi-mental treatment (experimental versus con-trol) and the division into high, middle, andlow groups on the basis of the first oral wordreading (100 SRA words). The pretests wereused as the covariates for the various posttestdependent variables. Although a detailedpre-sentation of the design and analysis is not pro-ductive here, there are a few points that arerelevant, as well as a few analyses of addi-tional measures.

10

First, the analysis indicated that the testto be the main measure of reading achieve-ment, the Iowa Test of Basic Skills, was in-deed not producing reliable results. Therewas not even a significant relationship betweenthe Ss' pre- and posttest results, indicatingthat the Ss were randomly scoring their testitems. It is also interesting to note that theI. T. B.S. was not related to the oral readingof the SRA wordsa measure with consider-able face validity as well as empirical valid-ity in the study. There was a significant re-lationship between the pre- and posttest per-formance on the Lorge-Thorndike. However,there was no significant effect of the experi-mental treatment as compared to the controlgroup results. The California AchievementTest was given twice (alternate forms) in theposttest sessions in an effort to salvage thetest results. There was not a significant re-lationship to the pretest I. T.B. S. however,

when the teat was taken without reinforcersand only a significant relationship (at the . 01level) of the vocabulary part with the pretestI. T. B.S. when the test was taken with rein-iorcers. However, the analysis of covariancedid not reveal any effect on the C.A. T. of theexperimental treatment, when the pretest I.T. B.S. was used as the covariate. The C.A.T. vocabulary part taken with no reinforce-ment was significantly related (. 01 level) tothe oral reading measures (SRA words), how-ever, th e reading comprehension part wasnot. Under the reinforcement procedures bothparts of the C.A. T. were significantly relatedto the pretest oral reading measure (the vo-cabulary at the . 001 level and the reading partat the .05 level).

The individually administered oral readingmeasure compost:d of the 100 SRA words ap-peared to be a reliable measure to the extentthat the pre- and posttest measures were sig-nificantly related (at the . 001 level). Further-more, this measure did show a significant ef-fect (at the .01 level) of the experimental treat-ment. The experimental group subjects im-proved more on their ability to read the 100words than did the control group. The ex-perimental group had an original mean scoreof 63. 8 on this measure and a posttest scoreof 76. 1; corresponding scores for the controlgroup were 63.6 and 69.2.

Fourteen of the 18 control Ss improved up-on their oral readingprobably, at least inpart, as a result of the special reading pro-grams to which they also had been exposed,some of which used SRA materials° It is in-teresting to note that although each of the ex-perimental Ss improved in this measure ofreading ability, some by large amounts, thegreatest improvement was shown by the Sswho were originally lowest in the oral readingability. The mean improvement for the low,medium, and hi gh ability subgroups in theexperimental group was respectively 18. 3,8. 8, and 9.8. The mean improvement for allthe groups combined was 12. 2 words.

Another analysis of variance was comb ctedto analyze the effect of introducing the rein-forcement contingencies upon the test-taking

behaviors, and the test results, of the Ss forthe California Achievement Test (both vocab-ulary and reading sections), and the Lorge-Thorndike on the verbal, non-verbal, and totalscores. The main effects were the Experi-mental Treatment, Previous Or al ReadingAbility (that is, the 3 groups, low, middle,and high) and the Reinforcement ContingencyCondition. The results showed that the rein-forcement procedure for the test administra-tion produced s ignif i c an t effects upon theCalifornia Achievement Test. On the vocab-ulary the Ss advanced from a mean a 23. 56under no reinforcement to a mean of 26. 00with the reinforcement conting enc ies. Thisresult was significant at the .05 level as wasthe next result. On the reading comprehensionpart the Ss advanced from a mean of 35.53 to38. 50. The results for the Lorge-Thorndikewere: on the verbal an advance from a meanof 67.83 to 71. 36, on the non-verbal an advancefrom a mean of 80.89 to 85.89, on the totalIQ an advance from 74. 36 to 78.62. These 3results were significant at the .901 level.

This improvement occurred even though 7fewer items, on the average for the two tests,were answered by theSs under the reinforce-ment contingency conditions. Thus, the chil-dren's scores under the no-reinforcementconditions were spuriously inflated since someof those 7 additional items would have beencorrect by chance. At any rate, it may beconcluded that the children's test behaviorsand test results were affected, and in a posi-tive direction, when they were reinforced forcorrect test-taking behavior and punished forrandom responding.

Two types of instructional-t e c hn i c i an swere employed in the study and it is of someinterest to compare the results of their sub-jects for possible differences. As measuredby the pre-and posttest on the 100-word SRAoral reading measure, the mean gain in wordsread was 12. 56 for the adult-volunteer in-structional-technicians and 12. 11 for the highschool - student i n s t r u c ti ona I-technicians.Thus, neither group of instructional-techni-cians was differentially effective.

11

IV

DISCUSSION

The general procedure and the reinforcersystem appear to be generally functional. Theattention, attendance, cooperation, and dili-gent work behavior of t h e various childrenwas maintained in good strength throughoutthe length of the study. In fact these be-haviors appeared to be gaining strength. Thus,the methods have proved successful with anincorrigible juvenile delinquent and with thepresent group of 18 children which includedmentally retarded children and emotionallydisturbed children. In addition, the proce-dures have also been tested with the samesuccess in clinical work with several psychi-atrically referred children, one of whom wasdiagnosed as schizoid, as well as with ongoingresearch which will be mentioned further onwith culturally-deprived problem learners,emotionally disturbed children, and youngerchildren with learning problems.

It has been suggested (Staats, 1965a, 1965b,in press, in preparation; Staats and Staats,1963) that learning principles and proceduresmust be extendedin the realms of clinical,educational, and child psychologyto the studyand treatment of complex human behaviors inlong-term studies. The present methods in-dicated that this maybe done with detailed andobjective observations of the processwhilethe subjects are being treated for behaviorproblems. In this case the problem was a caseof cognitive deficit.

It should be stressed that this was possibleemploying sub-professionals as the instruc-tional- (or therapy-) technicians. In the firststudy a probation officer served in this role.In the pres ent case the technicians were house-wives and high school students. As was out-lined in the first ude of these procedures(Staats and Butterfield, 1965), this finding hasgeneral implications for various areas ofclinical psychology and education in which in-dividual treatment of behavior problems is toocostly when professionals are employed, butwhere the child does not adjust to the grouptraining situations. The following conclusions

12

made in the first study are thus supported bythe present results.

Thus, the procedures could be widely ap-plied or adapted by various professionals,for example, social workers, prison of-ficials, remedial teachers, [special edu-cation teachers], tutors, and so on. In aneven more economical application, helpersof professionals could be used to actuallyadminister the procedures; for example,selected delinquents (or prisoners) couldadminister the procedures to other delin-quents. Thus, the procedures could beutilized in various situations, such as set-tlement houses, homes for juvenile delin-quents, prison training programs, parts ofadult education, and so on. . . . [T ]heresults suggest possibilities for economicinnovations in education generally [Staatsand Butterfield, 19651.

Related to this is the matter of the cost ofthe treatment. In the first study the 4-1/2months of training cost $20. 31 in reinforcers.This was closely paralleled in the presentstudy. That is, the mean cost for the subjectsover a similar period was $22.29. In manycases this would be a very minimal price topay to insure that ordinarily untrainable chil-dren would be not only amenable to varioustypes of instruction, but would work diligentlyand learn well.

Moreover, not only was the behavior ofthese childrenvery appropriate in the learningsituation, but the behavioral measures indi-cated that the children covered a great deal ofmaterial and learned a number of new words.A mean of 616.8 words were presented thatthe children did not at first know. After train-ing, in the long-term retention measure, thechildren retained a mean of 70.9% of the words.This represents a straightforward indicationthat the children (including 7 children in c las s -es for the retarded) were learning the materialon which they were being trained. In addition,

the curves that indicate the rate at which thechildren were reading was accelerating, orremaining constant, even though the difficultyof the material was increasing.

These findings were supplemented by thedifferences between the experimental and con-trol subjects on their ability to read the 100SRA words when individually presented to eachchild. The experimental group increased inthis index of reading ability to an extent thatwas significantly greater than the increase ofthe control group. Thus, when the measure-ment of reading ability w a s controlled anddirectly observed, and when the reading stim-uli were among those on which the experimen-tal group had received training, the effect ofthe experimental treatment was seen over thewhole group as compared to the control group.This occurred even though the control grouphad begun to receive special training in read-ing during the study.

However, the standard test results did notshow a difference between the pre- and post-test measures, for the experimental over thecontrol group. Several alternative interpre-tations of this lack of effect on the tests areavailable. The training materials may notproduce general skills that transfer to othertasks, such as taking an achievement test. Ifthat was the case it would be an indictmentagainst the SRA materials at least when pre-sented in the present fashion. This interpre-tation is difficult to accept, however. Pro-vided the vocabulary included in the materialshas generality, learning the words should byitself improve the child's reading ability forany verbal material that included the words.Nevertheless, it is possible that more train-ing would be necessary to be reflected on thetests, that the reading materials used mighthave been too elementary for many of thesesubjects, and so on.

Another possibility is that the tests them-selves were insensitive dependent variablesin the present instance. For example, thetests might have been too difficult. It is in-teresting to note that the California Achieve-ment T e s t, but not the Iowa Test of BasicSkills, w a s significantly related to the oralreading scores. Perhaps if the more ele-mentary C.A. T. had been given in the pre-test instead of the I. T. B.S., the experimen-tal effect might have been shown. It is evenmore likely, however, that the test-taking be-havior of these children was at fault. Thechildren all had long histories of unsuccess-ful (which means unreinforced) test taking.It would be e xp e c t e d under those circum-

stances that th e behavior would deteriorateand be replaced by less effortful behaviorsthat is indiscriminant marking of items. Thisweakness would h av e been amplified by theuse of achievement tests that were too diffi-cult for the children, since it would preventthe children from obtaining reinforcement intaking the test. It should be noted that in theStaats and Butterfield study, where the testsdid reflect th e success of the training, thetesting was individually adminiotered ratherthan a group procedure, which guaranteedbetter test-taking behavior to some extent,and the tests were easier and contained feweritems (both of which would enhance the rein-forcement properties of the testing situation).

It should also be noted that the special train-ing given to the control group children prob-ably helped obscure the positive affects of theexperimental treatment. This, of course,would also have had an affect upon the resultsof the individually administered word readingmeasure.

It should be added, however, that these arematters for further research, some of whichis now underway. The first two authors havea study now in progress in which the tests andtesting circumstances h av e been improved.The study involves culturally-deprived chil-dren who are rnore homogeneously low in theirreading deficits. The instructional-technicianswill be drawn from a culturally-disadvantagedpopulation. Adrian Van Mondfrans and thefirst author are also extending the methodsand materials to younger children (4th, 5th,and 6th graders), with and without extrinsicreinforcement. And Judy McBurney and thefirst author have a study in which emotionallydisturbed children are being individually test-ed in greater detail during the course of thetreatment. Although these studies are not yetcompleted, it is again evident that the proce-dures and reinforcer system work universallywell with various children with various typesof problems.

The reinforcement, system used in the Staatsand Butterfield study as well as the presentstudy was designed so that the amount of rein-forcement given per unit of behavior decreasesover the period of the training. Thus, duringthe training th e child was gradually gettingless reinforcement for his reading behavioreven though his behavior was maintained in asgood or better strength. This result is in partan answer to the frequent question concerningthe u s e of extrinsic reinforcement in childtrainingthat is. that the child may becomedependent upon the reinfo7cers. It may be

13

suggested that sophisticated methods may bedesigned to produce the desirable behavior butwith no lasting dependence upon a,1 artificialsystem of reinforcement.

One more point will be mentioned here.The reinforcement procedure which was in-troduced to produce better test-taking behaviorhas general significance. The children whohad been randomly scoring items began toperuse each item more carefully under theextrinsic reinforcement procedure. The re-sults indicated that when the good test-takingbehavior was reinforced and poor test-takingbehavior punished the desired behavior in-creased in strength to a statistically signifi-cant extent. When reinforced the childrenscored better on the tests, even while answer-ing few' test items. This has theoreticalimplications; for differences in "motivation"are not usually controlled when testing dif-ferent groups of subjects, even though there

14

are manycategories of individuals whose mo-tivational circumstances would no doubt affecttheir test measurements. It may be suggestedthat any group comparisons on tests shouldtake into account likely differences in test-taking behavior as a function of motivationaldifferences. It is suggested that this could becontrolled in theoretical studies by insuringthe subjects of the experiment are reinforcedfor good test-taking behavior.

The present findings also have clinical andeducational significance, however. There aremany cases in which test results are part ofdiagnosis and treatment and where the test-taking behaviors of the subject prevent accu-rate appraisal. It is suggested that more validtest results could be obtained through the useof reinforcement to strengthen the "good testbehavior. " The first author plans additionalresearch in this area of study.

REFERENCES

Staats, A. W. A case in and a strategy forthe extension of learning principles to theproblems of human behavior. In L. Kras-ner & L. P. Ullman (Eds.), Research inBehavior Modification. New York: Holt,Rinehart, and Winston, 1965. (a)

Staats, A. W. An integrated-functional learn-ing approach to complex human behavior.Technical Rep. No. 28 between the Officeof Naval Research and Arizona State Uni-versity, 1965. (b) Reprinted in large partin B. Kleinmuntz (Ed. ), Problem Solving:Research, Method, and Theory. NewYork: Wiley, 1966.

Staats, A. W. Language Learning and Cognitive D e v e 1 o pm e n t. New York: Holt,Rinehart, and Winston, in press.

Staats, A. W. Integrated-functional learningtheory and clinical psychology: I. Humanmotivation and conditioning therapies. Inpreparation.

Staats, A. W. , & Butterfield, W. H. Treat-ment of nonreading in a culturally-deprivedjuvenile delinquent: An application of re-inforcement principles. Child Develop-ment, 1965, 36, 925-942.

Staats, A. W., Finley, J. R. , Minke, K. A. ,& Wolf, M. Reinforcement variables inthe control of unit reading responses. J.exp. anal. Behay. , 1964, 7, 139-149.

Staats, A. W., Minke, K. A. , Finley, J. R. ,Wolf, M. M. , & Brooks, L. 0. A rein-forcer system and experimental procedurefor the laboratory study of reading acqui-sition. Child Development, 1964, 35, 209-231.

Staats, A. W. , & Staats, C. K. Complexhuman behavior. New York: Holt, Rine-hart, and Winston, 1963.

Staats, A. W., Staats, C. K., Schutz, R. E. ,& Wolf, M. The conditioning of readingresponses utilizing "extrinsic" rein-forcers. J. exp. anal. Behay. , 1962, 5,33-40.

15