Abstract
The Girls Creating Games program is a demonstration project designed to increase the number of girls who become producers, rather than just users, of technology. The activities and instructional approach are aimed to increase girls’ expectations for success with computers and the extent to which they value computer work and the support they receive. In over 23 sessions after school and in the summer, middle school girls worked in pairs to design and program a computer game that was supposed to help other students. Two hundred fourteen girls participated in a study of the program. Data from surveys and interviews suggest that compared to girls not in the program, participants increased their computer skills, knowledge about computers, and perceived social support. They decreased their negative stereotypes about girls and information technology workers. Implications for applying aspects of the program to middle school classrooms are discussed.
Introduction
Although men and women use computers at equal rates, there are vast gender differences in who designs and produces new computer hardware and software technologies (U.S. Department of Education, 2000). Women, Latinos, and other minorities are the least likely to be interested in or aware of high-tech careers ( Kearney, 2002), and women make up only 27% of workers in the areas of computer and mathematical operations (U.S. Department of Labor, 2005). Barriers include a lack of confidence, negative attitudes toward computers, lack of social support, and the belief that computers require solitary work with little social relevance (American Association of University Women, 2000; Goode, Estrella, & Margolis, 2006; Zarrett, Malanchuk, Davis-Kean, & Eccles, 2006).
Middle school may be a key time for intervention. In these years, students make critical choices regarding their identity and perceived ability, which shape their educational and career paths (Brickhouse, Lowery, & Schultz, 2000; Tang & Cook, 2001). However, most middle schools and high schools still focus on building computer literacy rather than promoting higher order thinking (Goode, Estrella, & Margolis., 2006), even though hands-on experiences that include programming can engage students with information technology (IT; Tucker et al., 2004). In the last decade, hundreds of after school programs have tried to increase gender equity in science, technology, engineering, and math (National Science Foundation, 2003), but few collect the kind of data needed to determine whether and why (or why not) the program had the desired impact (American Association of University Women, 2004; Dryburgh, 2000). This paper presents findings from a study of an out-of-school program where girls work in pairs to create computer games.
Only two other published studies (with elementary and high school students) have used game creation as a strategy to increase girls’ interest in IT as part of a research-driven agenda (Kafai, 1995; Miller, Chaika, & Groppe, 1996). Although controversial, computer games have an early influence on the skills and attitudes that are the best predictors of later technology-related behavior (Greenfield & Cocking, 1996; Levine & Donitsa-Schmidt, 1998; Subrahmanyam & Greenfield, 1998). Computer game design is an exciting way to put girls in the role of producers of technology, rather than simply as consumers. Game design is also innovative in that it involves identity exploration as a way to transform, not simply respond to, existing gender roles (Cassell & Jenkins, 1999).
This study is designed to examine the effects of the Girls Creating Games program on the participants and to identify ways to strengthen the program in future implementations. Two research questions are addressed:
1) Did participants increase their capacity to pursue and persist with computer technology?
2) What aspects of the program can be improved?
The evaluation of this program is grounded in the expectancy-value model developed by Eccles and colleagues (1983). This model suggests that achievement-related behaviors can be explained by expectations for success and subjective task values. In other words, girls are more likely to pursue and persist in IT careers if they believe they will be successful and enjoy and see value in doing well in the field of IT. Previous studies suggest that girls’ interest in computers, including their perceived value and relevance to other parts of their lives, plays a role in decisions to pursue computer courses and careers (Dickhauser & Stiensmeier-Pelster, 2003; Goode et al., 2006; Zarrett et al., 2006). Also, expectations for success, including confidence and self-efficacy, play a critical role in whether a girl chooses a non-traditional career path (Nauta & Epperson, 2003; Zarrett et al., 2006). The current study builds on findings that show the importance of support from peers in whether or not girls participate and persist in computer science (Goode et al., 2006; Margolis & Fisher, 2002). Below, we describe some of the activities in the Girls Creating Games program and present data that suggests this approach can help to prepare girls to pursue and persist in computer science courses and careers by increasing their expectations for success, subjective task value, and perceived support.
The Girls Creating Games Program
The Girls Creating Games (GCG) program was implemented in its entirety six separate times over two years. The curriculum contains 23 sessions, each lasting two hours. The program was held four semesters after school for 12 weeks (two days a week), and over two summers for six weeks (four days a week). The program design employed some of the techniques used in Intervention Mapping (Bartholomew, Parcel, Kok, & Gottlieb, 2001) which starts by identifying specific program objectives, linking them to theoretical models and evidence-based interventions, and developing practical strategies that lead to specific program activities. This process increases the likelihood that a program is research-based, will appeal to participants, and will have the desired impact. Also, a clear theoretical model allows for a closer examination of why a program succeeded or failed. The program in this paper builds on previous research-based approaches that involve learning by design, collaboration with peers, female role models, and a focus on the practical applications of what is being learned (Campbell et al., 2002; Clewell & Campbell, 2002; Cohen et al., 1996; Lee, 1997).
Program activities were organized within four strands that build on the four critical design features identified by the Cognition and Technology Group at Vanderbilt University (2003). The strands are designed to link the activities and expected outcomes, based on previous research. In the first strand, Learning by Design, instruction is organized around a meaningful problem: how to design and create a computer game using Macromedia’s Flash MX software. Girls were encouraged to create a game that would help incoming students adjust to middle school. In the second strand, Scaffolding and Modeling, instructors support the development of conceptual understanding by providing students with the resources to create their games and to solve problems independently. Collaborative Learning is the third strand and involves activities that build a community of learners, such as having students work in pairs to both design and then program their game. Instructors support the development of effective relationships within and across pairs by role modeling behaviors and leading fun activities designed to strengthen communication and mutual decision-making. In the final strand, Identity Formation, girls explore careers in IT, interact with female role models who challenge gender stereotypes, and get public recognition to promote their “tech savvy” identity.
The strands run concurrently over the course of the program. All activities are designed to make working with computers fun and to link technology to real-world applications.
The evaluation reported here can inform efforts to integrate technology into middle school classrooms. To that end, we report 1) whether participants in the GCG program increased their expectations for success, subjective task value, and perceived social support, compared to non-participants, and 2) what the students liked and disliked about the program.
For more information about the instructional approach and specific activities in each strand, see Denner, Werner, Bean, and Campe (2005) and the program guides and lesson plans . The 45 games created by the girls can be played at this website, and an analysis of the games can be found in Denner, Werner, Bean, and Tyner (2005). |
Methods
Participants
 |
A total of 126 girls enrolled in the GCG program. They were recruited from two middle schools in a small city and a nearby recreational club by offering pizza parties, posting flyers, asking teachers to nominate students, and mailing information to parents. The first school serves 650 students and has 200 computers. The student population is 67% White (non Hispanic), and 26% are eligible for free or reduced fee lunch. The second school serves 590 students and has 113 computers. |
The student population is 57% White, and 26% are eligible for free or reduced fee lunch. Participants were in the sixth through eighth grades, and at pre-test, the average age was 11.70 years (SD = 1.03). Most (94%) were born in the U.S., and slightly more than one third reported speaking a language other than English at home at least some of the time. Most (87%) of the students had access to a computer at home, and they used it on average one to two times a week for games or accessing the Internet. Girls of the same age from different schools who were either in, or wanted to be in, an after school program that involved computers made up the comparison group (n = 88).
Procedures
A quasi-experimental, pre-post test design was used to evaluate the program. Most surveys were administered on the computer. However, due to logistical challenges, 17 students in the comparison group completed paper surveys, and these students reported significantly higher skills at post-test than those who completed the survey on the computer (p < .05). Since that was the only difference between the groups, the data were combined for subsequent analyses.
Two sources of qualitative data were collected. Interviews were done with 31 girls who were selected to represent the range of grade levels, computer expertise, and race/ethnicity. Interviews lasted between 20-40 minutes and were conducted with individual girls at the school or club after the program was completed. Electronic notebooks were used to collect data on satisfaction with the program. During one of the last program meetings, students were asked to open their password- protected notebook on the computer and type a response to various questions, including what they liked about the program.
Measures
Race/ethnicity was collected in a variety of ways. For students in the GCG group who participated after school, race/ethnicity was obtained from parent report in school records Students in the GCG group who participated during the summer, as well as the entire comparison group, reported their own race/ethnicity by checking one of five categories on the survey.
Survey items measured subjective task value and expectations for success as well as social support. Subjective task value was measured with three scales, which were adapted from previous studies or developed by our research team. The Stereotypes about Computer Workers scale was adapted from Lawhead, Wilkins, and Rheningas (n.d.). This scale includes seven items that describe people who work with computers. Examples of items include: “Works with other people,” “Is creative,” and “Does not have time for their family.” Responses ranged from strongly disagree (1) to strongly agree (5), with items coded so that a high number reflects endorsement of negative stereotypes. Cronbach’s alphas were .70 at pre-test (n=207) and .73 at post-test (n=170). Our team created four items to measure Intentions to Study Computers. Students were asked about their plans to take courses such as computer graphics and computer programming. Responses ranged from definitely not (1) to definitely yes (5). The Cronbach’s alpha at pre-test was .72 (n = 207) and at post-test was .79 (n = 171). The Attitudes Toward Computers scale was based on items from Todman and Dick (1993). The seven items included choices like, “I like using computers in my free time” and “Computers are a waste of time.” Responses ranged from strongly disagree (1) to strongly agree (5) with high numbers reflecting a more positive attitude. Cronbach’s alphas were .79 at pre-test (n = 210) and .78 at post-test (n = 172).
Five scales measured expectations for success. Confidence with Computers items were derived from Levine and Donitsa-Schmidt (1998) and from Todman and Dick (1993). The scale includes nine such as, “I find using the computer easy,” and “I get frustrated with all the different keys and computer commands.” Responses ranged from strongly disagree (1) to strongly agree (5). Negative items were reverse-coded so a high score reflects high confidence. Cronbach’s alphas for the confidence scale were.82 at pre-test (n = 206) and .84 at post-test (n = 166). The Computer Skills scale had eight items, including “Make a copy of a file” and “Burn a CD.” Responses ranged from “I don’t know what this means” (1) to “I can do this so well that I can teach someone how to do it” (5). Alphas were .87 at pre-test (n = 204), and .85 at post-test (n = 171). Knowledge about Computers was measured by four items that assessed what their teachers, other kids, people in their family, and they think they know about the computer. Responses ranged from nothing (1) to a lot (4). Many of the skills and knowledge items were from Rockman et al. (n.d). Problem Solving was measured by one item created by the research team. Students were asked to complete the sentence: “If I don’t know how to do something on the computer at school, the first thing I do is….” Options included “try to figure it out myself,” “ask for help,” and “give up.” Gender Stereotypes were measured by one item, which built on a survey by Levine and Donitsa-Schmidt (1998): “Boys my age usually do better than girls when using computers.” Responses ranged from strongly disagree (1) to strongly agree (5).
