Effect of TA background on student laboratory achievement - Journal

Does anything in a TA's background have an effect on how well a student will perform ... Does it matter what sort of undergraduate school he or she co...
2 downloads 0 Views 3MB Size
-

~~

Effect of TA Background on Student Laboratory David L. Monts a n d Miles Pickering' Princeton University. Princeton. NJ 08544 Does anvthine . .. in a TA's backrrround have a n effect on how well a student will perform? D i e s it matter whether a T A is a n experienced teacher or not? Does i t matter what sort of undergraduate school he or she comes from? This paper is a n assessment of the impact of T A backgrounds on student performance on a qua-ntitative laboratory practical examination. T h e answers to the above questions not only are clearly of importance t o anyone faced with choosing TAs, b u t also they help to illuminate the nature of the TA-student interaction. Method The use of a permenganate-oxalate titration as a "practical exam" has heen described previously in detail ( 1 . 2, 3) and will only be summarized here. Under exam conditions (silence. no collaboration. imposed time limitation), n student standardizes a KMnOI solution and titrates an unknown containing oxslate. Results are reported on an "income tax form" requiring no chemical knowledge, and all arithmetical errors are corrected. The grade is determined' purely by the deviation of the student's value from the true value. This test offers a convenient benchmark of pure lab technique without contamination by any student difficulties with arithmetic or comprehension. The prartical exam prading method, thus, offers an objective appraisal of a student's laboratory proficiency, without depending on TA opinion in any way. For thisstudy, student results have been judged ss"suecessful" if their deviation from the true analvsis is less than 1% relative. (For example, if the true percentage of oxalate is 30%. the student must report a number between 29.7% and 30.3%; otherwise he is "unsuccessful.") While the exact same exam was run each year for four years, a numher of changes in the teaching environment took place which may well have a hearing on the results. In the second and third years the Princeton-Yale titration contest (3)was held, in which our group was pitted against our arch rival in New Haven. Although theeontest also twk place in the fourth year, there wasalmost nnemphasison it. The effect of the contest seems to have been important. Data on TA hackp(roundswas drawn from many sources. The TA's date of birth, marital status,and year in graduatesehwl weredetermined from personnel records. Information about previous teaching experience wasohtained from a questionnaire returned by the TAs. Data on undergraduate school enrollment and graduate degrees in chemistry offered hy the undergraduate schwl were taken from reference books ( 4 , s ) .Undercraduate school eataloa were examined to see if quantitative and qualitative analysis %.ereoffered by the chemistry departments. We did not determine on an individual basis whether each TA had taken these courses, but feel that schools offering such courses probably have a strong commitment to analytical ehemistryand so will insist that chemistrymajors have moreexpaswe to it. We initially attempted to useguestionnaires, but this approach relies on TA memory of the content of courses taken many years ago, and it is also very difficult to write questions that assess the degree of exposure to analytical methods.

Discussion T h e x2 values for t h e correlations between the T A variables a n d the student success rates are summarized in Table 2. W e

' Author to whom correspondence should be addressed.

768

Journal of Chemical Education

Table 1. Source of Data Year

CourseLeve14 Tihation Contest %

1976

1977

1978

1979

Lower

Lower Yes

Upper

Upper

Yes

Yes lbUf

49%

56%

44%

No

SUCCBJS~Y~44%

Table 2.

not emphasized)

Student Petformance Correlated with TA Background

x2 Value lor Correlation with Student Pedormance .~

Charaaeristlc of TA A.

.

Contest Emphasized

Experience 1. Previous Teading Experience of

Omer

Years

0.001'

1.63

0 13 0 023

1.63 1.06

0.73 0.14 0.036

0.82 0.016 6.12'

0.60

2.47

AOV Kind 2. P ~ ~ W OLab U I Teachlog Experience 3 RG+incetan Teaching Experience

Undergraduate School 1. Offered Ouant 2. OfferedOuai 3. Was Less Than 7000 Undergraduates 4. Offered no PhD Rogram C. Personal Characteristics 1. SingleC 2. Female B.

3. 22 Years Old (versus Older) 4.

First Year in Grad School

Signlllc8"Ily C O R B b W * p = 0.9 OT tat)' DIMnnelalion lp = 0.05 or benal = ~ a r r ~ TAS e d ~d Dener pertmiw n ~ j a n win y-

@asized.

-1.14 0.11 0.042 0.060

3.32. 0.96 2.93. 0.O0Zb

when u ~ ~ t lconlest m war e m

have broken down t h e d a t a into years when t h e titration contest was emphasized a n d t h e other years. T h e latter are probably much more typical of conditions a t most institutions. I n t h e "normal" vears we find that. statisticallv s ~ e a k i n e . t h e best T A s are s k g l e , 22 years old; a n d come from smaii, undereraduate-oriented schools. T h e w 2 for teachina experience variables fall short of statistical si&ficance in th% ssudy. Surprisingly, while being young is strongly correlated with having successful students, t h e year in graduate school does n o t seem t o matter. (Not all t h e first year graduate students are 22 years old, since some come t o Princeton with Master's degrees, or after working in industry.)

Table 3.

Magnitude of TA Effects on Student Pertormance Probability

of Student

Success TA

Characteristic

1%).

Incremental

Success Rate

(Studentl Se~tion)~

Experience 1. Previous teaching experience of any kind. No previous teaching experience ol any kind. 2. Wwious lab teaching experience. No previous lab teaching experience. 3. Pre-Princeton teaching experience. No TA experience before Princeton. B. Undergraduate School 1. Offeredquant Did not offerquant 2. offeredqw1 A.

Oid not offer qua1 3. Less than 7000 undergraduates Greater than 7000 un6sryaduater 4. Offereda PhD prqlram Offeredonly BS a MS degree

C. Personal Characterlstlcs 1. Single Married 2.

Female Male

3. 22 years old

Older 4. First Year in Grad Schaal at PU

In years when the contest was emphasized, the correlations fell short of statistical significance. The desire to win led to much more intervention in the lab by the instructor, (himself a Yalie), who had originally challenged Yale to the contest. Also, since the job of "beatingYaleV was seen as agroup effort, there was more student-to-student teaching in the laboratory periods before the exam. These efforts bore fruit, as can be seen by the increased percentages of success listed in Table I. Under these conditions the TAs were, therefore, relativelv less important as sources of information and motivation, and so the correlation disappeared. We conclude from this study that a TA's effect on students must be largely motivational. It is hard to believe that variables like aee or marital status could affect his or her direct competencLor the task of information transfer. In addition, it is reasonable to assume that the TA's skill a t information transfer would he increased by teaching experience, and that it would be affected by subject matter exposure. In our sample neither of these variables show correlation with student success a t statisticallv significant levels. Hence, we believe that the effects shownaredue to the TA's ability to get students excited (or perhaps worried) about the practical exam. In a "Provocative Opinion" in this Journal (6)one of the authors argued, on the basis of anecdotal information, that one of the maior thines that motivated TAs was the need to "root" for thestudent:. This is probably easier if thestudents and TA are similar in age and background. Nishibayashi (7) documents a tutoring program using undergraduates a t Occidental College in which seniors were used as tutors for freshmen. Columbia University has also restructured their well-known freshman supplementary course (8,9 ) to use section leaders who are undergraduates. In such programs enthusiasm, and obvious caring about student success, seem to be more important than na;rowly defined subject matter and pedagogical competence.

The lack of correlation with experience variahles leads us to the conclusion that direct skill a t information transfer is either of much less importance than is generally helieved, or that it reaches a limiting value early in the first teaching year. It may he that teaching success, a t least for a TA, depends more on intangibles like leadership abilities than on teaching experience. Our results are what one would expect, if a major part of producing proficient students is motivating them to learn, and if such motivation is most easily accomplished by a TA who is similar to the students being taught. We are concerned also ahout the magnitude of such effects. Does a student who has a TA who is single, for example, have a important edge over students in another section with a married TA? T o answer this question, Table 3 presents the data in terms of ~rohahilitvof student success as a function cdTA hackgnrund fartwi, in years when the titratinn amtest did not ohwure thc correlation. \Ve have also cvmrmted the "incremental success rate," the number of additional students succeeding when the TA variable is favorable, assuming a section size of twenty. In almost all cases fewer than two out of a section are affected. By way of comparison, the effect of the titration contest was an incremental success rate of 1.8 students per section, and so its motivational effect is of comparable magnitude. We feel that the TA background effects are small enough not to pose administrative problems, or to warrant selection of TAs on this hasis. This study is statistical in nature and does not offer useful information about individual cases. We have known a number of successful TAs who do not fit our categorization. Clearly, this study has major limitations. First, we have defined "successful TA" in a very narrow way; other results (such as students' enjoyment o f the course) may he as important as student skill a t analytical chemistry. We do not argue that other results are not important but will leave them to others to quantify. Second, there may he other extraneous variables; if, for example, our lab manual were completely inadequate, the class would depend more on the TA's knowledge of technique, and so pronounced subject matter correlations might appear. And, of course, our results apply strictly only a t Princeton; they would have to be replicated elsewhere before a general conclusion should be drawn. T h e value of this studv is that it is one of the verv few in which it is shown that a TA has an objective effect onitudent proficiencv. It also raises a number of worthwhile auestions for furthe; research. If the work were replicated a t several places, a general picture of TA background effects might emerge. w e also need to know how student success on practical exams relates to more conventional measures of teaching assistant prowess such as student and faculty evaluations (10).

Acknowledgment Professors David Brooks and Robert Crabtree contributed useful ideas to an early draft of this paper. Literature Cited

Volume 58 Number 10 October 1981

769