1995 Program Evaluation of the Women
in Science Project at Dartmouth College
Return to Table
of Contents
Previous Page
This evaluation analyzed both quantitative and qualitative data.
This section first presents the results of the quantitative analysis
of data from Dartmouth College Records to probe whether WISP has
had any impact on the number of women science majors at Dartmouth.
The second part reports data from students' responses to qualitative
data--questionnaires, interview, and journal questions. The qualitative
part is organized around five foci of the data collection: the
WISP internship, WISP programming, women who major in science,
women who are considering a major in science, and women who "leave"
science. The "WISP Internship" section distills and
develops six major thematic strands that run throughout this evaluation.
The next four sections present data from specific questions, call
attention to particularly noteworthy factors, and offer some explanatory
remarks.
Return to Table
of Contents
One of the underlying goals of the WISP program is to increase
the number of women continuing and majoring in science. The program
began in 1990; therefore, the first cohort of women who held internships
their freshman year were seniors in 1994. To begin to assess whether
the WISP program is having an impact on the number of women science
majors, and to try to distill factors that contribute to a student's
decision to major in science, a statistical analysis was run on
data from all students (male and female) in three Dartmouth classes.
Data from Dartmouth College's records were used to look for significant
relationships to major choice for the following variables: whether
the student had a dual major, SAT math scores, whether the student
postponed graduating a year, graduation date, gender, GPA, and
workstudy eligibility2.
This analysis used a dataset comprised of all students graduating
from Dartmouth College in 1992, 1993, and 1994. The goal of the
analysis was to determine whether:
(a) students' major choices are related to gender
(b) if they are, whether that relationship is changing over time
(c) whether changes over time are attributable to changes in other
student demographics.
For example, an increase in the number of women majoring in science
over time might reflect an increase in the number of all students
majoring in science (perhaps because science jobs are perceived
to be more abundant). Or an increase might reflect changes in
students' precollege preparation (if Dartmouth admissions were
becoming more selective, and math SAT scores were consequently
increasing, one might expect more students to have the option
of majoring in a science). Hence, to separate out the most likely
contributors to major choice, and to see whether a gender effect
persists even when those factors are controlled for an analysis
of variance was used.
First Analysis of Variance
The first analysis of variance was run to identify demographic
covariates with students' decisions to major in science. (For
a more detailed account of the analyses in this section see
Appendix
B.) This analysis of variance showed that factors associated with
the decision to major in science are:
(1) high Math SAT scores (strongest effect),
(2) being male (next strongest effect)
(3) being work-study eligible (a measure of financial need--weakest
significant effect tested here).
It also indicated that:
(4) students with dual majors are more likely to major in science
(5) there is no relationship between GPA at graduation and the
SCIENCEMAJOR variable
(6) there is no relationship between postponing graduation (1
or more years) and the SCIENCEMAJOR variable.
However, it is important to note that this model explains only
5.6% of the variance. In other words, many other factors contribute
to a students' major decision.
Second Analysis of Variance
Next the YEAR*GENDER interaction was explored with a second analysis
of variance to determine whether the effect of gender was stable
over the three years of data analyzed (92-94). This analysis showed
a significant effect of GENDER and YEAR*GENDER, but not YEAR alone
on whether a student majors in science. In other words, the effect
of gender changes over the three years of data.
Post hoc multiple comparisons for YEAR*GENDER interactions on
SCIENCEMAJOR were run for women only to investigate the difference
between individual years. These showed that there is a significant
difference between the number of women majoring in science between
1992 and 1993; and 1992 and 1994. The following table presents
data about the number and percent of women and men majoring in
the sciences in all three years. The jump from 12.7% of the women
to 20.1% is significant. The role of WISP in encouraging this
jump is nebulous since the increase occurred a year before the
first cohort of interns were seniors (the increase is seen in
1993; the first class with women interns is 1994). That WISP had
some impact is not untenable however--for example, upperclasswomen
could have participated in the non-WISP components of the
internship.
Table 1: Table of Majors 1992-94
| |
Frequencies |
Row Percent |
| NonSci |
Science |
Total |
NonSci |
Science |
| 1992 |
Men
Women
Total
|
413
423
836
|
172
62
234
|
585
485
1070
|
70.6
87.2
78.1
|
29.4
12.7
21.9
|
| 1993 |
Men
Women
Total
|
417
354
771
|
166
89
255
|
583
443
1206
|
71.5
79.9
75.2
|
28.5
20.1
24.9
|
| 1994 |
Men
Women
Total
|
405
360
765
|
1666
96
262
|
571
456
1027
|
70.9
79.0
74.5
|
29.1
21.1
25.5
|
Third Analysis of Variance
The third analysis of variance strove to determine whether the
gender effect changes from 1992 to 1994 can be attributed to changes
in other demographics. The significant main effects of this analysis,
in decreasing order of importance are SATMATH, GENDER, WORKSTUDY,
and YEAR*GENDER.
A statistical change in the number of women science majors is
one important way to document a possible impact of the WISP program.
However, judging the "success" of the program based
exclusively on this type of data would be short-sighted. First,
the number of majors may not accurately reflect the number of
women who are actually planning to continue in science. As described
in the next section, many students who plan to attend medical
school choose not to major in a science because they know that
they will be studying only science in the future and want to explore
other non-science fields as an undergraduate. Other students who
do not major in science may still consider a career related to
science, but they selected Dartmouth because it is a liberal arts
college and feel that the requirements of a science major limit
their ability to explore other interests.
Second, the definition of "success" deserves some attention.
Women who do not major or continue in science still consider WISP
a valuable experience; they appreciate their increased comfort
with and understanding of scientific processes. As our society
becomes increasingly science and technology-based, it is important
that people in all fields have an accurate understanding of the
practice of science to make informed decisions, not only scientists.
To develop a more comprehensive understanding of impact and influence
of WISP and of the variables affecting majoring in a science,
students' attitudes and insights were probed using qualitative
methods.
Return to Table
of Contents
2
Work study
eligibility was used as an indicator of financial background and
need. A degree in the science is often a more marketable degree
upon graduation, and salaries for science jobs are often higher
than those in the humanities--factors that could be important
and influential for students with financial need.
Return to Table
of Contents
Next Page
|
