Introductory Programming:
Bridging the Gender Gap?
PSY/ORF 322
Sharla Cloutier
Meera Krishnan
Yan Zhong
Advisor:
Joel Cooper
Abstract
The purpose of this project was to examine the social interactions in introductory programming classes at Princeton University in order to identify and suggest possible methods for bridging the “digital gender divide”. Based on past studies, we hypothesized that males would be more comfortable in both programming and computing than females; females would be more willing to seek help from both peers (particularly males) and preceptors; and males would be more willing to help females. With the prevalence of iPods in today’s college campuses, we wanted to examine if such personal music devices would have any effect in lessening the gender differences. In a survey conducted on ORF 201 and COS 126 classes, we found that our hypotheses were mostly supported. Significantly, females were more willing to seek the aid of preceptors (outside of precept), while the music devices seemed to have some effect in alleviating the programming-comfort gender gap. This suggests that there may be potential, through accessible technology and willing teachers, for improvement in the gender gap.
Introduction
The Presence of the Digital Divide
In our society, computers are becoming increasingly important to everyday life. In fact, as technology progresses, computers have become necessary to almost any profession, and have permeated nearly every home. It is not surprising, therefore, that computer usage is slowly being integrated into the educational system. In fact, in an effort to keep up with the technologically-driven world, many classrooms have started using computers as a learning tool (Cooper ix). Unfortunately, increased usage of computers in schools has opened up another set of problems. Although males easily assimilate into the new technological classrooms, females find computers a source of anxiety, and have an aversion towards them from a very young age (Cooper x). The gender difference in these early attitudes have led to the “digital divide,” a term used in Joel Cooper and Kimberlee D. Weaver’s book Gender and Computers to describe the gap between females’ and males’ comfort levels with computers (Cooper x).
An important question to ask is, does this aversion that starts at a young age continue through the rest of school and even college? Several studies show that females do continue feel less comfortable than males with computers in college. Temple and Lips asked over 300 students at the University of Winnipeg in Canada about their attitudes and comfort levels related to computers. They discovered that female students, in comparison to males, reported anxiety towards computers, less inclination to pursue technology as a career, and less comfort and confidence with computers (Cooper 14). Dambrot and colleagues conducted a survey on almost 1000 students at the University of Akron, finding negative and fearful attitudes towards computers by females (Cooper 14). In the University of Santiago de Compostela in Spain and the University of Leicester in Britain, similar female anxiety was exhibited (Cooper 14). As shown by Cooper and Weaver, female computer anxiety does not get “cured” over time. In fact, they believe the opposite, explaining that aversions to computer use at a young age cause females to simply avoid them later, leading to an even greater divide as they get older (Cooper 25).
Wilder, Mackie, and Cooper conducted an important study of Princeton University students, suspecting that in this selective sample, in which many students had previous experience with computers, gender differences might be decreased. At first, they came to the conclusion that this was true; men and women seemed to be equal in the number of programming languages they knew and their opinions on the appropriateness of computers for each gender. However, a closer look revealed important differences between those who had computer experiences and those who did not. Females who had computer experience reported that they were less comfortable than males with no experience. In this example, gender differences remain even in selective groups with a great deal of exposure to computers (Cooper 27-28).
Stereotypes and the Self-Fulfilling Prophecy
In a computer class, stereotypes and expectancies of teachers and parents becomes extremely important. People expect females to be worse at using computers than males, and this stereotype may become internalized in students. A study by Rosenthal and Jacobson of elementary school teachers showed the influence of teachers’ expectations towards students. Rosenthal and Jacobson randomly assigned some students as “high performers” to the teachers. They found that, as a result of the teachers’ extra encouragement, the “high performers” actually did perform better on an IQ test (Cooper 69-70). Likewise, if parents and teachers believe that males are better than females at using computers, they may influence the students to believe the stereotype as well. These false beliefs cause anxious behavior that may result in a “self-fulfilling prophecy,” in which females make the stereotype come true through their anxiety. One study on black and white interview candidates confirmed the self-fulfilling prophecy belief. The stereotypes against the black candidates caused them to act less calm and composed and perceived their interviewers as less friendly; the internalization of the stereotype actually made them worse candidates for the job (Cooper 74). Another similar analysis by Dipboye found that often interviewers were also at fault. Their impressions of the candidates before the interview caused them to show signs of either approval or disapproval that were mirrored by the interviewee, which ended in the pre-interview impressions coming true (Dipboye 584).
The self-fulfilling prophecy effect that turns stereotype into reality may be an important cause of the digital gender divide. Cooper and Weaver describe differences in social dynamics that are caused by females believing the stereotype that males are better at using computers than females. Speaking to Princeton students, they found that women in male-orientated subjects like computer-science and math often looked towards male students for help on assignments. In addition, one young woman noted that males “seemed to think it was their responsibility” to help female students on assignments (Cooper 64-65). Canada and Pringle conducted an analysis of the differences in classroom interactions between all-female and mixed-sex schools. They found that in mixed-sex schools, males tended to dominate classroom interactions, and female-initiated interactions decreased in comparison to single-sex schools (Canada and Pringle 179). Asking males for help and allowing males to dominate classroom interactions indicates the truth of the self-fulfilling prophecy effect in real-life interactions; gender stereotypes lead to behaviors that only further perpetuate them.
Bridging the Divide
A follow-up to the Wilder, Mackie, and Cooper study on gender differences for selective groups in Princeton revealed that the self-fulfilling prophecy effect can be lessened. They asked students to imagine that they were in a psychology statistics class and had to perform a statistical analysis on a computer, and they told students to indicate their expected comfort level. As expected, women reported more anxiety than men, but even more intriguing was the effect of previous computer experience. It was found that previous experience with computers was a particularly important factor in that females with more experience reported greater expected comfort for the computer analysis (Cooper 29-30). Rosemary E. Sutton, in a literature review on computer-related gender and race inequities, found that when computer experience is held constant, gender differences in attitudes towards computers are greatly reduced. She cites studies conducted by Chen (1986), Gordon and Levin (1989), Gressard and Loyd (1987), and Campbell (1989). She also found that more computer experience is correlated with better attitudes and lessened anxiety towards computers, two factors that contribute greatly towards the self-fulfilling prophecy (Sutton 490).
A study conducted last year at Duke University introduced a new potential topic of research- the effect that personal music devices such as iPods have on comfort with computers. In the study, Duke gave every freshman a free iPod in order to “stimulate creative uses of digital technology in academic and campus life” (Duke University website). The purpose of the study was to see if students’ academic work improved through using iPods in different ways, such as to record lectures, take oral notes, and make electronic flashcards (Moore). After the study, 75 percent of freshmen reported that they used the iPods for academics, and a number of professors said that iPods made the students more engaged in class (Moore). The effects of iPods and other personal music devices on computer comfort was not explored in the study.
Purpose
The study we have conducted is an extension of all those described here. We decided to further explore the idea of the self-fulfilling prophecy with respect to gender and computers, and we were particularly interested in college-level computer classes with more subtle gender differences. In accordance with the ideas Cooper and Weaver touched on, we planned to focus on the amount of help students gave and received as an indicator of the extent to which the self-fulfilling prophecy was at work. In addition, we were interested in finding out the part that personal music devices play in the digital gender divide, an idea that was fueled by the Duke University study. This goes along with Wilder, Mackie, and Cooper’s suggestion that females with more computer experience feel more comfortable with computers; we wish to find out if using personal music devices can be considered a type of previous computer experience.
Hypothesis
Before conducting our survey, we hypothesized that females in introductory computer classes would seek more help overall than males, from both preceptors and peers. Because of the stereotype that males are better than females at using computers, we hypothesized that females who asked for help from peers would be more likely to ask males than females. Males would also be more likely to ask males than females, for the same reason. In addition, males would be more likely to help females than other males on assignments, while females would be more likely to help other females as well.
We also hypothesized on the effects of iPods and other personal music devices on overall computer comfort and programming in introductory computer classes. Our hypothesis was that males and females who owned personal music devices and used them often would be more comfortable with computers and programming than those who did not. Also, males and females who owned them would also be closer to each other in their comfort levels.
These hypotheses reflect several things outlined in the studies above. First of all, we are testing to see if a gender difference actually exists in students’ confidence towards their computer assignments. In asking students about the help they receive, we are essentially trying to find out how much faith the students have in their own computer skills. We expect that females have less confidence than males because they have internalized the stereotype that females are worse than males at using computers. Next, we are interested to find out how much students play out the stereotype when asking for help, by seeing if students are more confident in the skills of males because of the stereotype. Finally, we are testing to see if the general use of personal music devices will increase the general level of confidence towards using computers for females, based on the idea that females with more technological experience are more comfortable with computers and programming.
Procedure and Sample
We chose to create a survey with various questions involving computing and programming comfort and whom students get help from and give help to. We also asked about iPod and MP3 player (collectively referred to as “personal music devices”) ownership and use. Responses were mainly given in the form of indicating yes or no, circling a number on a scale from 1 to 5, and providing the names of peers (for gender classification purposes during survey analysis). The survey we handed out is attached at the end of the paper (see appendix).
The students we surveyed came from two popular beginning programming classes, Doug Clark’s COS 126, and ORF 201, which is co-taught by Robert Vanderbei and Hugo Simao. We collected 26 surveys from the COS class and 24 from the ORF class, for a total of 50 surveys. Nineteen females and thirty-one males filled out the survey, and eighty percent of the respondents were freshmen. Intended majors represented included all of the engineering fields (civil and environmental, chemical, computer science, electrical, mechanical and aerospace, and operations research and financial), as well as architecture, astrophysics, chemistry, economics, math, molecular biology, and physics, as well as the ubiquitous “undecided.”
Results
We will focus on data that we have deemed to be statistically significant, either by exhibiting a p-value of less than 0.25 (acceptable for small sample sizes such as ours), or by following or highlighting a trend established by a related significant result. Therefore, certain pieces of data will not be mentioned in the following discussion because they are inconclusive and not entirely essential to our analysis. Statistically speaking, both the two-sample t-test and the two-proportion z-test will be employed, with a summary of the tests immediately following the discussion.
Males (n1 = 29) reported an average computing comfort of 3.73, while females (n2 = 18) indicated an average of 3.53 (Fig. 1). A two-sample t-test conducted on this data provided significant results, with a t-statistic of 0.734 and a p-value of 0.23. Males (n1 = 28) also reported greater comfort with programming, with an average of 3.22, as compared to 3.05 for females (n2 = 18) (Fig. 2). This two-sample t-test showed the same trend as the first, but it was not significant according to our established standards (t = 0.663, p = 0.26). Overall, males are more comfortable than females in the areas of computing and programming.
The most noteworthy personal music device data comes from comparing the gap between males and females who use these devices to the gap between non-using males and females, in the context of programming comfort. Male and female device users (n = 27) showed approximately the same level of programming comfort, with respective averages of 3.13 and 3.18 (Fig. 3). However, non-using males were significantly more comfortable with programming than their female counterparts. Non-using males (n1 = 12) displayed an average comfort level of 3.32 (which, unexpectedly, is even higher than all those who use devices), while females (n1 = 7) reported an average of 2.86 (Fig. 3). Even with such small sample sizes, a two-sample t-test gave significant results of t = 1.170 and p = 0.13. This data supports the idea that personal music devices help to bridge the gap between males and females in the level of programming comfort.