Aicha Sajid An Le David Edelman Susan Warren Gender and Software This paper reviews and compares "The Effects of Gender, Expectations of Success and Social Comparison on Children's Performance on a Computer-Based Task," by Jonier and Messer, and "An investigation in to Gender Bias in Educational Software Used in English Primary Schools," by Bradshaw and Clegg. Both are case studies dealing with gender issues within computer proficiency. The first deals with how the gender of the user affects use; the latter deals with how the user views gender within the program. In the first article, Jonier and Messer studied the expectations of success between children of different genders when faced with a computer-based task. They initially surveyed the children on the level of confidence they held in achieving a stated task of solving five math questions. The results were measured using the score of each student. In addition, they introduced another independent variable, whether the student was working alone or working in the prescence of the same gender. It is critical to note that the participants in the experiment were screened to ensure that they had similar backgrounds in computers. The results of the experiment varied with many surprises. Generally, children who self-assessed themselves to have a high level of confidence scored higher. Females had lower expectations of success than males, but not statistically significant. Surprisingly, when comparing the scores, there was no gender difference in performance. This contradicts previous studies where they found that boys performed better than girls on a computer-based mathematics task. The results of varying children's environment also produced a difference in performance between genders. Males who were surveyed to have a greater expectation of success generally scored higher when tested alone versus tested in the presence of other males. Inversely, males with lower expectations scored higher when in a group and lower when alone. Compared to females, the results were the exact opposite. Females with high expectations scored higher under a social environment, where as those with lower expectations did worse in this enviroment. Therefore, the authors proposed an effort to raise positive expectations for girls in computer-related tasks, as to increase their performance in a social environment. This approach, although promising, does not take into account some of the critical variables that affect performance. For example, other educational studies have shown that whether students are praised for their hard work versus their intelligence strongly affects later performance. It is important to ensure that the programs in use praise students for the effort perhaps even more so than for their final results. This is a variable that is more easily controlled by the teacher. The Bradshaw and Clegg article, in contrast, focused on how both boys and girls projected gender onto androgynous characters in computer program. In their preliminary research, they found that there was widespread evidence of gender conditioning that begins at an early age. In one example, they found that "when adults read to children, the adults tend to ascribe male gender to neutral characters and stereotype images". Based on this, they began with the hypothesis that children would not understand the concept of a genderless person and would revert to stereotypes of characters as male; girls would be more likely than boys to view characters as female. This was of critical importance because "if both girls and boys see computing characters as predominantly male, the implicit assumption, on the part of the manufacturers and educationists, that girls will find these packages more friendly than overtly sexist ones, will be weakened. To test this hypothesis they evaluated childrens' interactions with programs that follow best practices (i.e. programs that were considered by the researches to contain a minimal amount of gender bias). The results strongly supported their thesis. They found that "male as norm" was the dominant strategy. Regardless of the androgynous of the character, both boys and girls were more likely to view the characters as male. Women, when pressed, were often willing to view them as female; however, males strongly stuck to their original assignment. The authors proposed that teachers could proactively attempt to modify student opinions about characters regarding gender identification. Unfortunately, they do not provide strategies that could achieve that goal. We agree with their suggestion that it may also be important to strive for a diversity of characters with different genders rather than characters with no gender at all. It seems silly to ignore the fact that in the "real world", all people have genders. It may be more effective to actively put female characters into typically male roles such that the girls identify with those roles. This may ultimately be a more effective way to improve girls' performances on computer-related tasks. Each of these articles aim to improve girls' performances in computer-related tasks. The first focuses on girls' performance as related to their expectations in social situations. The latter focuses on how gender within software itself may affect their performance. Each article suggests that the teacher can have a powerful role in negating the potential negative effects studied. By being aware of the potential problems that girls may encounter, their performance can be significantly improved. It is also important to measure how changes to aid girls affect the performance of boys. An ideal solution would allow girls to perform without negatively affecting the performance of boys.