The Impactof Teaching"Assistants

on StudentRetention inthe SciencesLessons for TA Training

By Christopher O'Neal, Mary Wright, ConstanceCook, Tom Perorazio, and Joel Purkiss

Attrition from the sciences remains a national problem. We present results from a survey of over 2,100 undergraduates that,contrary to previous research, suggests that teaching assistants (TAs) influence student retention in the sciences in multiple ways.Multiple linear regression (MLR) and student comments suggest that TAs influence lab climate, course grades, and students'knowledge of science careers, all of which have an effect on students' decisions to stay in or leave the sciences. We close withrecommendations for TA training, mentoring, and management to positively impact student retention.

ationally, retention in sciencehas become a matter of realimportance, as educationalinstitutions try to slow the

sizable flow of undergraduates out ofscientific fields (Campbell et al. 2002;National Science Foundation 2003;Strenta et al. 1994). Attrition in thesciences is especially problematic inthe undergraduate years because ap-proximately 40% of students who cometo college intending to major in thesciences ultimately decide to major insomething else (Astin and Astin 1993;Seymour and Hewitt 1997; Strenta etal. 1994). Many of those who leave thesciences are capable students with theaptitude to do well in science (Mont-gomery and Groat 1998; Seymour andHewitt 1997; Tobias 1990), and the at-trition problem is particularly acute forwomen and people of color (Astin andAstin 1993; Holstrom et al. 1997; Sey-mour 2002; Xie and Shauman 2003).Clearly, students have every right tochange majors and explore topics theyhad not considered before coming tocollege, but it is a problem (and in this

24 JOURNAL of COLLEGE SCIENCE TEACHING

case, a national problem) if they changetheir prospective major because it didnot satisfy their needs.

There are many hypotheses aboutthe reasons for undergraduate attritionin the sciences, and this article exploresone of them, the impact of graduatestudent instructors or TAs in gatewaycourses. The gateway course is the initialcollege course in the sciences taken by afirst- or second-year student who expectsto be a science major in college. Attritionin the sciences is especially likely in thefirst two years of college (Seymour andHewitt 1997), at the time when studentsare still taking gateway courses.

A majority of science, technology,engineering, and mathematics (STEM)bachelor's degrees are awarded atresearch and doctorate-granting uni-versities, which employ large numbersofTAs in introductory science courses

(National Science Board 2006). In fact,in the life sciences and physical scienc-es, TAs are responsible for more thantwo-thirds of the lab sections (NationalCenter for Education Statistics 2000 ascited in Seymour 2005, p. 3; Nyquist,Abbott, and Wulff 1989; Rushin et al.1997). As anyone who has talked withparents and incoming college studentscan attest, there is considerable publicconcern that TAs do not teach as wellas faculty. In this research study, weasked: What influence do TAs have onunderclass students' plans to major inor leave the sciences?

Literature on student attritionand the role of TAsIn their groundbreaking study onstudent attrition, Talking About Leav-ing: Why Undergraduates Leave theSciences, Seymour and Hewitt (1997)

Christopher O'Neal (coneal@umich.edu) is the senior consultant for institutional initiatives,Mary Wright is coordinator of GSI Initiatives, and Constance Cook is associate vice provostand director at the Center for Research on Learning and Teaching (CRLT) at the University ofMichigan in Ann Arbor (U-M). Tom Perorazio is a doctoral candidate in U-M's Center for theStudy of Higher and Postsecondary Education. Joel Purkiss is associate director of curriculumevaluation at the Medical School and lecturer in the Department of Sociology at U-M.

spoke with numerous undergraduatesabout why they had either stayed in orswitched from a science major. Over-all, only 3% of students in their studywho were making a switch mentioned"language difficulty with foreign fac-ulty or TAs," and none of the studentsmentioned "poor teaching, lab, orrecitation support by TAs" as a factorin their decisions to switch majors.

Given the significant roles that TAsplay in the sciences, we were surprisedby the Seymour and Hewitt findings thatTAs do not influence students' plans fora major. While students may quite obvi-ously see the impact of the professor ona course, they may not be fully aware ofthe division of labor in a course: whodoes the grading and plans lab assign-ments, and how labs are managed sothat everyone can have a positive expe-rience. Therefore, our research focuseson the multiple roles, both obviousand subtle, that TAs might play in thedecisions that science students makeregarding a major in the sciences. Wehypothesized that TAs indirectly influ-ence the experiences of students in thesciences through the informal mentor-ing and role-modeling they provide andthe learning environments they create intheir labs or discussion sections (oftenreferred to as classroom climate).

Research designThe surveyTo determine the role of TAs in under-graduate attrition from the sciences, inJanuary 2004, we surveyed all under-graduates at a large, Midwestern univer-sity who took at least one of the sevengateway courses for prospective majorsin chemistry, biology, and physics in fall2003 (e.g., introductory biology, generalchemistry, mechanics and sound, and soon). The survey asked students abouttheir intention to major in the sciences(or have a career in the health sciences)before the class they had just taken, aswell as after they had taken the course(i.e., when they took the survey). It is im-portant to note that this research focusedon students'plans to stay in or leave theirscience major, not actual retention or at-trition, and it was retrospective, so it alsofocused on their perceived attributions

for any change. To address our concernsthat undergraduates may not know whotakes responsibility for different parts ofthe course, our survey allowed studentsnot only to choose from instructionalroles (e.g., "faculty" or "TA") but alsoinstructional events (e.g., lab climate,course grades). The complete survey ispresented in Appendix 1.

A web-based survey tool was used tocreate and distribute the survey. Becausethe survey tool allowed us to track indi-vidual responses, individuals' responseswere paired with demographic data (sex,race/ethnicity, class, and course grade)obtained from the university registrar.Out of 3,656 undergraduate studentssurveyed, 2,669 students responded tothe survey (a 73% response rate). Ofthe 2,669 students surveyed, 2,102 werefirst- and second-year students (78.8%of respondents). We focused our analysisonly on the underclass students (i.e., first-year students and sophomores) enrolledin the seven courses because they werenot yet committed to a major. Abouthalf (46.7%) of the respondents werefirst-year students, and a small majority(53.3%) were sophomores. Nearly half(49.4%) were female, and only a smallminority (10.2%) were underrepresentednonAsian minorities. The seven gatewaycourses had a total of 113 TAs teachinga total of 263 sections or labs.

A separate survey collecteddemographic data on the 113 TAsteaching in the sampled sections (sex,race/ethnicity, experience, and soon), as well as data on TAs' opinionsof their training. Summaries of datafrom both the undergraduate and TAsurveys are presented in Table 1.

Data analysisWe used multiple linear regressionanalysis (MLR) to examine the impactof various factors on student plansfor retention (independent variablesare summarized in Table 1). The TA'ssection was the unit of analysis in theregression. We ran two models. In one,called "Attrition," the dependent vari-able was the percentage of a TA!s stu-dents whose interest in a science majordecreased. In the other model, called"Retention," we used the percentage of a

TA's students whose interest in a sciencemajor increased, combined with the per-centage of that TA's students who startedinterested in a science major and stayedinterested. We excluded from both mod-els the portion of students who starteddisinterested in science and stayeddisinterested in science (13% of the to-tal). For statistically significant factors,qualitative comments are presented toillustrate how students attributed thesefactors to their TAs.

Appendix one contains the online surveydistributed to undergraduates In December2003 and January 2004.

Survey of student decisions about sciencemajors or careersPlease answer the following questions about yourexperience as a student in [course] in the 2003 fallterm.The survey will take less than five minutes andall individuals' answers will be confidential.

1. Please indicate your agreement with the follow-ing statements (choices range from "definitelynot"to "definitely so"):

a. BEFORE the fall 2003 term, I planned to major inscience, engineering, or math,

b. As of NOW, I plan to major in science, engineer-ing, or math.

c. BEFORE the fall 2003 term, I planned topursue a career in a health field (such asmedicine or nursing).

d. As of NOW, I plan to pursue a career in ahealth field.

e. For the winter 2004 term, I plan to enroll in a sci-ence, engineering, or math course.

f. After the winter 2004 term, I plan to continue en-rolling in science, engineering,or math courses.

2. How have the following factors influenced yourchoice of major and/or career? (choices were,"decreased my interest in science,'no influence,"and "increased my interest in science"):

a. The professor for this science courseb.The graduate student instructor (TA) for

this coursec. The environment or climate of the lecture

for this coursed.The environment or climate of this lab sectione. My grade in this science coursef. My grades in mathg. Learning more about a career or major outside

of scienceh. Other (please specify)

3. If yourTA increased or decreased your interest ina science career or major, please explain why.

4. Any other comments?

MARCH/APRIL 2007 25

ResultsEncouragingly, the vast majority ofstudents reported that their plans tomajor in science actually became morelikely over the course of the term.Less than 12% of students reported adecreased interest in their decisions tobe science majors (Table 1).

In regression models using TA sec-tion as the unit of analysis, there was onlyone statistically significant factor associ-ated with both retention and attrition: labclimate (Table 2). Students' commentsrevealed that for them, lab climate

encompassed a number of elements, in-eluding enthusiasm of the TA, their ownanxiety levels, and how welcome theyfelt in the class. The comments beloware typical examples of what studentshad to say about lab climate.

The TA ' ability to make the lab at-mospherefimn...definitely made doingthe labs a lot easier Furthermore,his cool attitude made it a lot easierto listen to what needed to be done,making the learning process moreeffective. [1It increased my interest in

doing a science major (Asian malesophomore, course grade=A)

The lab environment was verystressful, and I often found myselffrustrated and confused. Thisenvironment definitely decreasedmy interest in furthering a sciencecareer (female first-year student,race unknown, course grade=B)

Students indicated that a TA's efforts tomake the lab atmosphere a positive learn-ing environment enhanced the learning

Variables used in multiple linear regression modeling for TA-level analysis.

Depeden variable

Retention score: Percentage of TA's students moving toward, or remaining

in, science

Attrition score: Percentace of TA's students movino away from science

Sex

Race/ethnicity

Primary language of undergraduate education

Is the TA a first-term grad student?

Is this the TA's first term as a TA?

"TA training prepared me well"5 pt. scale, 5=strongly agree

"I plan to pursue a faculty position",5 pt. scale, S=strongly agree

Student opinions of TAEnd-of-term course evaluation question:"Overall, this TA was an excellentinstructor." (TA's average rank, 5 pt. scale, 5=best score)Did the TA influence your interest in STEM? (TA's average rank, 3 pt. scale,3=increased interest, =decreased)

Non-TA facto rs-Averages for students in a TA's classDid the professor influence your interest in STEM?(3 pt. scale, 3=increased, 1 =decreased)Did lecture climate influence your interest in STEM?(3 pt. scale, 3=increased, 1 =decreased)Did lab climate influence your interest in STEM?(3 pt. scale, 3=increased, 1 =decreased)

Did course grade influence your interest in STEM?(3 pt. scale, 3=increased, 1 =decreased)Did your math grades influence your interest in STEM?(3 pt. scale, 3=increased, 1 =decreased)Did learning of other careers influence your interest in STEM?(3 pt. scale, 3=increased, 1 =decreased)

Sub~jecnd section characteristisHow hard is the course this TA taught? (Proxy: average student grade, 4 pt.scale with 4.0 being the highest)

Min=0.0%

Min=0.0%

Female:

White:

Non-English:

Yes:

Yes:

Min=1

Min=1

Min=2.20

Min=1.00

Min=1 .20

Min=1.20

Min=1.42

Min=1.00

Min=1.40

Min=1.33

Min=1.87

Max=100.0%

Max=100.0%

52 (46.0%)

64(56.6%)

33 (29.2%)

53 (46.9%)

76 (67.3%)

Max=5

Max=5

Max=4.91

Max=2.80

Max=3.00

Max=2.63

Max=2.75

Max=2.67

Max=2.54

Max=2.57

Max=3.54

Mean=75.1 % S.D.=20.2%

Mean=1 1.9% S.D.=13.6%

Male:

All ID'd others:

English:

No:

No:

Mean=3.46

Mean=3.22

Mean=3.99

Mean=2.06

Mean=2.12

Mean=1.99

Mean=2.00

Mean=1.91

Mean=2.07

Mean=2.17

Mean=3.15

61(54.0%)

47(41.6%)

80 (70.8%)

60(53.1%)

37 (32.7%)

S.D.=0.91

S.D.=1.15

S.D.=0.63

S.D.=0.27

S.D.=0.25

S.D.=0.20

S.D.=0.26

S.D.=0.25

S.D.=0.20

S.D.=0.16

S,D.=0.21

26 JOURNAL of COLLEGE SCIENCE TEACHING

process and drew them to science ma-jors. However, students who perceived astressful and fi-ustrating lab environmentreported a decreased interest in science.

One additional factor was statisti-cally associated with retention: mathgrades (Table 2). Students who com-mented positively on their mathematicsgrades indicated that they performedwell in the science class and were moreinterested in staying in science.

My calculus TA increased myinterest because he was very en-couraging and motivating. Overall,my math and science classes wentpretty well. (white female first-yearstudent, course grade=A)

Three other MLR factors were notsignificant at the 0.05 level, but ap-proached significance in their associa-tion with retention. They were: gradein the course, learning about othercareers, and students' rating of the TAon end-of-term evaluations.

Grade in the courseStudents' positive comments abouttheir course grades highlighted suc-cess with difficult course material,emphasized by positive feedback andcourse grades that students felt were afair reflection of their learning.

My TA " positive response to my labreports encouraged me to continue

pursuing a career in science re-search. (Asian female sophomore,course grade =A +)

Students' negative comments abouttheir course grade focused on threethemes. First, some students reportedthat they struggled with the coursematerial and found it too difficult;they often interpreted this difficultyas evidence that they did not belongin science.

Although I struggled with this classgreatly, I very much enjoyed it as Ilove [the discipline]. However, myseeming incapability to understandit has rendered it impossible for

WMactors assooateo wutn 1As retention score imocel A) ana attritron Model_A:_Retention' _ Model_B:_Attrition'score (Model B): Summary of multiple linear regression results. S

Bet chapvalS p-value

VA's sex (female)

TA's race/ethnicity (all others)

TA's primary language of undergraduate education (Non-English)

Is theTA a first-term graduate student? (yes)

Is this the TA's first term as a TA? (yes)

"This TA was an excellent instructor"rating from students

"My TA training prepared me well" rating by TA

"1 plan to pursue a faculty position" rating byTA

TA influences interest

Professor influences interest

Lecture climate influences Interest

Lab climate influences interest

Course grade influences interest

Math grades influences interest

Learnina of other careers influences interest

Sujc an seto chrateisicHow hard is the course thisTA taught? (average student grade) (avg.student grade)

Chemistry course? (yes)3

Physics course? (yes)3

R2 =.451 Adj.R

2 = .304

1 Positive beta indicates increasing retention, negative indicates decreasing retention.2 Likewise, positive beta indicates increasing attrition, negative indicates decreasing attrition.3 Biology course is the excluded category.tp<0.10 * p < 0.05 **p < 0.01

R2=,302 Adj.R2= .115

MARCH/APRIL 2007 27

.002

.018

-.068

-.077

.091

-.262

.086

-.037

.040

.048

.058

.042

,045

.042

,024

.019

.986

.879

.609

A59

.390

.051 t

.432

.740

.086

.010

-.096

.048

-.179

.076

.117

.127

.030

.036

.044

.032

.034

.032

.018

.015

.440

.940

.520

.685

.135

.613

.343

.309

.026

-.097

-.140

.368

.248

.338

.188

.124

.126

.169

.127

.106

.125

.130

.873

.534

A07

.026*

.065t

.009**

.076t

.216

-.056

-.016

-.387

-.164

-.012

.048

-.009

.351

.006

.094

.095

.129

.096

.080

.095

.099

.102

.044

.055

.246

.749

.932

.038*

.276

.933

.683

.955

.351

.006

.164

-.009

.171

.134

.058

.073

.243

.953

.302

I TABLE 2 1

What non-TA factors influence student interest in a science major?

me to major in [it], sadly enough.(female first-year student, raceunknown, course grade=B+)

Second, other students reflected onthe grading climate, commenting onthe stressful nature of an instructor'sexpectations, the course curve, or thelack of encouragement.

Although Ido not wish for the uni-versity to lower their standards forstudents, [the course], along withmost of the other science coursesoffered at the university, can bethe most discouraging and unmo-tivating courses offered. I realizethis class is a weeder class andit definitely shows! (white femalesophomore, course grade=A +)

Finally, a third group of studentspointed to dissatisfaction over thegrading process due to unclear assess-ment standards.

Ifelt that the TA should give you atleast an idea of how you're doingin the class before you receive yourgrade. ... They should at least giveyou an idea or talk with you on howthey feel you can improve your labreports rather than just returningthem with no notes or comments oranything. (Hispanic female sopho-more, course grade=B+)

Learning about careersStudents who learned of summeropportunities, internships, and theflexibility of career choices reportedthat they were more likely to pursue acareer in the sciences. We should notethat the corresponding question on thesurvey actually addressed learningabout careers outside of science, butstudent comments clearly indicatedthat learning about careers withinscience was important too.

The class increased my interestin working in a laboratory envi-ronment and finding out what it'sreally like to run your own experi-ments. (African-American femalesophomore, course grade=A)

28 JOURNAL of COLLEGE SCIENCE TEACHING

Students'end-of-term rating of their TASurprisingly, TAs who received highend-of-term course evaluations fromstudents were more likely to havelower levels of student retention thanTAs with low evaluations (thoughthe differences are not statisticallysignificant). The reasons for this trendare unclear.

None of the MLR factors associ-ated with TA demographic character-istics had a significant effect on themodel. Neither TA gender, race, Eng-lish language training, or experienceappears to significantly impact studentchoices about major. Likewise, coursedifficulty (as measured by averagestudent GPA across the course) hadno impact on student plans.

DiscussionSeymour and Hewitt (1997) report thatundergraduates do not identify theirTAs as a reason for switching majors.Our findings are similar to theirs in thatundergraduates do not attribute the TA asa major cause of their change in plans fora science major. However, undergradu-ates in our study did identify a numberof factors that educators know are af-fected by science TAs: most importantly,the lab climate, but also course grades,and learning about careers. Similarly,math grade was important for retention,although it is a factor more within thedomain of math instructors.

Training and mentoring TAs toaid science retentionAlthough science TAs were not directlyimplicated as a factor in students' aca-demic planning, they likely influencethose factors which were found to besignificant, most notably, lab climate,learning about other careers, and as-pects of course grades. If TAs couldpositively influence these factors, thenthey could conceivably have a dramaticinfluence on students' plans to stayin or leave the sciences. How can webetter train and support TAs? We makethe following recommendations forscience TA training programs:

1. Sessions on issues of retentionWe recommend that TA training

programs enlist novice TAs in discus-sions of best practices for retention.The authors of this study were able tofind only one paper directly address-ing TA training for retention (Bol-giano and Horton 1993). While notspecifically aimed at science disci-plines, and not directly assessed, theone-hour retention training programdescribed in the paper was viewedpositively by TAs. Following fromthe example presented by Bolgianoand Horton, training on retention forscience TAs might include presenta-tions on science retention data fromtheir own institution, presentations ofnational retention figures and dispari-ties, discussions of the undergraduateexperiences of the TAs in training,and case studies on why studentsleave and why they stay.

2. Sessions on the role of the TA infostering a positive lab climatePearson (1991); Galvin (1991); andCano, Jones, and Chism (1991, pgs.87-94) describe a number of bestpractices for TAs seeking to createa classroom climate that is inclu-sive and supportive for all students,regardless of gender, race/ethnicity,or background. Science TA trainingprograms should include discussionsof lab climate, the best use of labgroups, the experiences of underrep-resented students in the lab, and howto handle domineering and uncoop-erative students.

3. Sessions and readings onthe role of the TA in modelingpossible science careersVisible connections to the real worldand potential careers appear to be animportant hook for students in intro-ductory courses. TAs can be excellentrole models of possible careers inscience fields. TA training programsshould involve discussions of whatinfluenced the TAs themselves to enterscience, what resources exist to guidestudents in career choices, and how totalk about alternative careers in sciencewith their students. Another possibilityis to have TAs occasionally teach thelecture class, not just the discussion

section or lab. Guest lectures give theTA some experience in the lectureenvironment, and allow them to showstudents their own research and experi-ences as scientists (Gaither 1994).

4. Sessions and readingson making explicit gradingstandards and procedures andcommunicating to studentshow they are doing in the courseWhile TAs often do not control thegrading policies of a course, theyoften have some discretion in howthey grade individual assignments.At the very least, TAs control howstudents are made aware of and com-fortable with those policies. Studentcomments revealed that by explicitlydiscussing expectations for the courseand individual assignments, TAs caneliminate much of the uncertainty thatcauses grade anxiety.

Finally, we would like to notethat TA training should not be viewedsimply as basic training designed toget TAs through their first term ofgraduate teaching. In order for train-ing to transfer to classroom practice,it is likely that continual follow-up,practice, mentoring, and advancedtraining is necessary (Notarianni-Girard 1999; Shannon, Twale, andMoore 1998). Therefore, TA trainersand faculty developers should con-sider ongoing mentoring and trainingthroughout a TA's career. We also notethat the principles discussed should beequally useful for faculty managinglarge introductory courses with TAs.Weekly course meetings represent anideal venue to train TAs to retain sci-ence students.m

ReferencesAstin, A.W, and H.S. Astin. 1993. Un-

dergraduate science education: Theimpact of different college environ-ments on the educational pipelinein the sciences. Los Angeles, CA:Higher Education Research Institute,University of California.

Bolgiano, C.F., and G.R. Horton.1993. Effective use of trainedteaching assistants in improvingthe retention of university students.

The Journal of Graduate Teach-ing Assistant Development 1 (2):57-73.

Campbell, RB., E. Jolly, L. Hoey, andL.K. Perlman. 2002. Upping thenumbers: Using research-baseddecision making to increase diver-sity in the quantitative disciplines.Newton, MA: Educational Devel-opment Center, Inc.

Cano, J., C.N. Jones, and NY Chism.1991. TA teaching of an increasinglydiverse undergraduate population. InPreparing theprofessoriate oftomor-row to teach: Selected readings inTA training eds. J.D. Nyquist, RD.Abbott, D.H. Wulff, and J. Sprague,87-94. Boston: Kendall/Hunt.

Gaither, Jr., J.C. 1994. Teaching thescientific method: An effective wayfor graduate teaching assistants toexperience the lecture hall. The Jour-nal of Graduate Teaching AssistantDevelopment 1 (3): 115-18.

Galvin, K.M. 1991. Building aninteractive learning community:The TA challenge. In Preparingthe professorate of tomorrow toteach, eds. J.D. Nyquist, R.D. Ab-bott, D.H. Wulff, and J. Sprague,263-74. Boston: Kendall/Hunt.

Holstrom, E.I., C.D. Gaddy, VV VanHorne, and C.M. Zimmerman.1997. Best and brightest: Educa-tion and careerpaths of top scienceand engineering students. NewYork: Commission on Profession-als in Science and Technology.

Montgomery, S.M., and L.N. Groat.1998. Student learning styles andtheir implications for teaching.CRLT Occasional Paper, No. 10.Ann Arbor, MI: Center for Re-search on Learning and Teaching,University of Michigan.

National Science Board. 2006. Scienceand engineering indicators, Vol.2. (NSB 06-O1A). Arlington, VA:National Science Foundation.

National Science Foundation, Divi-sion of Science Resources Statis-tics. 2003. Women, minorities, andpersons with disabilities in scienceand engineering: 2002. (NSF03-312). Arlington, VA: NationalScience Foundation.

Notarianni-Girard, D. 1999. Transferof training in teaching assistantprograms. Journal of GraduateTeaching Assistant Development 6(3): 119-47.

Nyquist, J.D., R.D. Abbott, and D.H.Wulff, eds. 1989. The challengeof TA training in the 1990's. InTeaching assistant training in the1990's: New directions for teach-ing and learning, No. 39, 7-14.San Francisco: Jossey-Bass.

Pearson, J.C. 1991. Gender bias in thecollege classroom. In Preparingthe professorate of tomorrow toteach, eds. J.D. Nyquist, R.D. Ab-bott, D.H. Wulff, and J. Sprague,73-79. Boston: Kendall/Hunt.

Rushin, J.W, J. DeSaix, A.S. Lums-den, D.P. Steubel, G. Summers,and C. Bernson. 1997. Graduateteaching assistant training: A basisfor improvement of college biologyteaching and faculty development?The American Biology Teacher 59(2): 86-90.

Seymour, E. 2002. Tracking the pro-cesses of change in U.S. undergradu-ate education in science, mathemat-ics, engineering and technology.Science Education 85 (6): 79-105.

Seymour, E. 2005. Partners in innova-tion: Teaching assistants in collegescience teaching. Boulder, CO:Rowman and Littlefeld.

Seymour, E., and N.M. Hewitt. 1997.Talking about leaving: Why un-dergraduates leave the sciences.Boulder, CO: Westview Press.

Shannon, D.M., D.J. Twale, andM.S. Moore. 1998. TA teachingeffectiveness. Journal of HigherEducation 69 (4): 440-66.

Strenta, A., R. Elliot, R. Adair, M.Matier, and J. Scott. 1994. Choos-ing and leaving science in highlyselective institutions. Research inHigher Education 35 (5): 513-47.

Tobias, S. 1990. They're not dumb,they're different: Stalking thesecond tier Tucson, AZ: ResearchCorporation.

Xie, Y., and K.A. Shauman. 2003.Women in science: Career pro-cesses and outcomes. Cambridge,MA: Harvard University Press.

MARCH/APRIL 2007 29

COPYRIGHT INFORMATION

TITLE: The Impact of Teaching Assistants on Student Retentionin the Sciences: Lessons for TA Training

SOURCE: J Coll Sci Teach 36 no5 Mr/Ap 2007

The magazine publisher is the copyright holder of this article and itis reproduced with permission. Further reproduction of this article inviolation of the copyright is prohibited. To contact the publisher:http://www.nsta.org/college