Gender differences in attributions and behavior in a technology classroom.
The attributions that students make to explain their successes and failures have been implicated as being important in affecting their future expectations, outcomes, and decisions and could be part of the reason for the under-representation of women in the fields of computer science and engineering. This study examines the perception of accomplishment, attributions and behavior of fourth- through sixth-grade girls and boys in a technology course where students learned to build simple Lego[R] robots and program them using RoboLab[TM] software. There were no significant differences in the girls' and boys' assessments of their daily accomplishment or in their attributions for their successes or failures, but the girls' behavior during the course was significantly different from that of the boys in that they asked more questions of teachers and made fewer self-assured statements. Teachers need to be aware of these gender differences and take them into consideration in their instruction.
The National Science Foundation's most recent statistics report that in the year 2000, women earned only 28% of the bachelor's degrees in computer science and only 21% of the bachelor's degrees in engineering, demonstrating that women are still under-represented in technological areas where growing need and opportunity are likely. In their review of the literature on possible explanations for the marked gender differences in math, science and technology (MST) careers, Meece, Parsons, Kaczala, Goff and Futterman (1982) concluded that students' choices are influenced by multiple interrelated factors. They proposed a theoretical model for understanding the complex interaction of those factors, linking academic effort, achievement and enrollment choices to students' expectations for success. Those expectations are in turn influenced by students' past achievements, self-attributions about ability, and perceptions of task difficulty and value. They identified men's and women's attributions for their successes and failures as being critical and called for additional research on attributions. In a more recent review of the literature on gender differences in MST fields, Sanders, Koch and Urso (1997) identified a similar set of factors influencing women's enrollment in MST courses and career decisions.
Researchers interested in the mechanisms by which girls are socialized to have lower self-expectations and lower motivation, especially in typically "male" fields, have employed Bernard Weiner's Attribution Theory (1974, 1994), which explains motivation by examining the reasons individuals give for their successes and failures. Weiner suggested that students attribute their successes and failures to four categories of causes: ability, effort, luck and task difficulty. These four categories can be characterized as follows: 1) being internal or external to the individual; 2) stable or unstable over time; and 3) under the individual's control or not. Table 1 summarizes the categorizations.
Some researchers suggest that females exhibit a more external locus of control in their attributions for success and failure and, therefore, tend to be less motivated than males who view their own internal ability and effort as responsible for their successes and failures. Huber and Podsakoff (1985) derived a measure of externality for participants in their study by subtracting ratings of the influence of internal factors from those of the external factors. They found that both men and women were more internal than external, but that men were more so than were women, providing limited support for the internality-externality hypothesis. Simon and Feather (1973) also found that women tend to make more external attributions but only when their achievement was counter to their original prediction. Sweeney, Moreland, and Gruber (1982) and Huber and Podsakoff (1985) suggested that women don't necessarily have an externality bias; rather, men have an internality bias.
More commonly, researchers distinguish between the four kinds of attributions and examine them separately under conditions of success and failure. They suggest that girls exhibit an internality bias in failure, thus blaming themselves, and an externality bias in success, thus not taking credit for their successes. Males do the opposite, exhibiting an externality bias in failure and an internality bias in success. While a number of studies discuss their results in terms of externality or internality, a close examination of them indicates that none find the complete pattern for men or women in both success and failure. Rather they report partial confirmation of the pattern, including women's greater use of luck (Bar-Tal and Frieze, 1976; Stipek and Gralinski, 1991) and of task difficulty in failure (Huber and Podsakoff, 1985) and men's greater use of skill in success and luck in failure (Levine, Reis, Turner and Turner, 1976). Thus, these findings do not rule out the possibility that, really, stability or controllability is the determining factor rather than internality or externality. In fact, Ryckman and Peckham (1987) suggest that stability is what matters and that ascribing success to the stable factors of either low ability or task difficulty leads to a helpless attitude, while attributing success to high ability promotes motivation. They found that boys tended to take the more mastery-oriented approach.
Covington and Omelich (1979a, 1979b) focus on differences in the internal attributions of ability and effort and suggest a pattern they term Self-Worth Theory. They claim that because of the value society places on having high ability, ability is the attribution that carries the most weight in an individual's assessment of self-worth, particularly in academic situations. They suggest that females more often use the explanation of low ability as the cause of their failure, while males more often use the explanation of high ability as the cause for their success. They also suggest that rather than viewing effort as independent of ability, students--especially as they grow older--see effort as directly related to ability because the amount of effort that a student needs to expend is viewed as an indication of the student's ability (Covington, 1992). Thus, while lack of effort is not viewed as a virtue, a negative attribution to low effort can be self-enhancing in that it enables students to prevent negative assessment of their ability and thus protect their sense of self-worth. Many studies report a male-female difference in terms of ability attributions that is precisely what Covington and Omelich (1979a) predicted. The findings with respect to effort, however, have been variable. For example, Parsons, Meece, Adler and Kaczala (1982) found that the boys in their study attributed their successes to effort as well as to ability, and D'Amico, Baron and Sissons (1995) found that their effort factor was the least important predictor of performance.
An approach that would stress the unifying theme in all these studies would be to characterize students' perspectives as adaptive or mastery-oriented as opposed to maladaptive with attitudes of helplessness. Using such an approach, the research can be summarized as reporting that boys are frequently found to take a more positive approach that leads to future motivation and success and furthers their sense of self-worth, while girls are more likely to take a self-deprecating attitude that lessens motivation.
Ideally, both women and men would make realistic, mentally healthy and motivating attributions. Therefore, it is of particular interest to understand the circumstances under which researchers have not found differences or have found women to give more positive attributions than men. For example, Parsons, Meece, Adler and Kaczala (1982) asked fifth and eleventh-grade students to recall the last time they succeeded and failed on a mathematics test and asked them to attribute their success or failure on that test. They found that girls were more likely to attribute both their success and failure to their ability, and boys were more likely to attribute both their success and failure to their effort. Likewise, Sweeney, Moreland, and Gruber (1982) measured success by a comparison of each student's test score with the mean class score. They found that both men and women internalized their successful outcomes, and that women were more likely than the men to externalize their poor performance.
Factors Affecting Attributions
Looking closely at the research designs used in studies of attributions helps to explain some of the seemingly contradictory findings reviewed above. Huber and Podsakoff (1985) addressed this issue by conducting a study in which they tried to determine the effect of various dispositional and situational variables on attributions and called for others to do the same. Among the factors that seem to contribute to the variation of results are the following: 1) masculinity of the task; 2) age of the participants; 3) authentic versus contrived task; 4) operational definition of success and failure; 5) operational definition of causal factors included; and 6) relationship of attributions to beliefs, expectations and achievement behaviors. What follows is a summary of some of the patterns that have emerged to date.
Masculinity of the task. Two studies in particular support the idea that sex differences in attributions are stronger when the task is commonly considered to be one in which men will naturally be better than women. Stipek (1984) surveyed 5th and 6th grade students before and after they took either a math or spelling exam and found sex differences in attributions with respect to the math test but none with respect to the spelling test. Rosenfield and Stephan (1978) came to a similar conclusion in an interesting experiment with college undergraduates where males and females were asked to perform the same task, but for some subjects the task was characterized as masculine and for others it was characterized as feminine. For the masculine-typed task, they found the expected sex difference of men making stronger attributions to internal causes in success and to external ones in failure. For the feminine-typed task, they found a reversal of the pattern such that women made more internal attributions for success and more external ones for failure, rather than the lack of differences reported by Stipek (1984) for spelling tests.
Age of participants. Research on attributions has been done with students from grade school to college and shows changes as children grow older. Stipek and Gralinski (1991) gave pre- and post-test questionnaires to third-grade students and junior high students, asking them to predict their outcome on a test and make attributions to ability or effort in either success or failure. They found that the younger children were more likely than older children to believe that effort leads to success. Nicholls (1979) found similar results when asking children aged six, eight, ten, and twelve years old to compare their reading ability to the rest of their class. He concluded, as did Meece, Parsons, Kaczala, Goff and Futterman in their 1982 review, that by age twelve, most low achievers have sufficient experience to determine their low ability and begin to make attributions for failures to their low ability. They become less optimistic about the potential of effort to lead to success or to make them smart.
Authentic versus contrived task. Understandably, much of the research on attributions has been done in laboratory settings with contrived tasks such as solving anagrams or mazes. The advantage of such studies is the control of extraneous variables and the ability to manipulate success and failure randomly for the participants in the study. The drawback is that researchers cannot be certain that the resulting attributions would apply equally to a task such as a school exam or a task encountered at a job.
Operational definition of success and failure. In attribution studies success is often operationally defined as some minimum score or letter grade, but sometimes success is defined by the participants' own self-assessment. One of the most disturbing findings that is reported by a number of authors (Fennema, 1974, Stipek, 1984, Stipek and Gralinski, 1991) is that girls often view a particular level of accomplishment or score less favorably than boys with an identical accomplishment. That finding suggests that it is the students' perception of success that is more important than their objective score or grade and suggests further that there is a need for more studies using participants' judgment as the definition of success and failure.
Operational definition of ability, effort, task difficulty and luck. Studies employing different measures and categorizations of attributions have the potential to contribute to the robustness of research knowledge, but it is important to pay particular attention to the definitions of variables especially when comparing studies with differing results. For example, Parsons, Meece, Adler, and Kaczala (1982) used open-ended questions and also asked participants to rank in order a set of eight possible attributions. The two measures yielded different patterns. Stipek (1984) included two different measures of task difficulty. One asked about the difficulty of a particular test and was considered unstable because future tests might or might not be difficult. The other asked about the difficulty of math in general and was considered to be relatively stable over time. In addition some researchers have used difference scores. Some have summed several items to develop a scale and, in a recent study, D'Amico, Baron and Sissons (1995) used factor analysis of a 23-item questionnaire to develop their scales.
Relationship of attributions to beliefs, expectations and achievement behaviors. Many studies have examined the specifics of students' attributions about their successes and failures, but it is also important to establish that students' attributions do influence their beliefs, expectations and subsequent achievement as indicated in the model proposed by Meece, Parsons, Kaczala, Goff and Futterman (1982). If such linkages cannot be established, then students' attributions are of no practical importance in explaining sex differences in the achievement behaviors of choice, persistence and performance. A number of studies provide at least correlational evidence that these relationships may operate in the way that Meece, Parsons, Kaczala, Goff and Futterman (1982) propose. Basow and Medcalf (1988), Bar-Tal and Frieze (1976), Stipek (1984), and Stipek and Gralinski (1991) all found that girls' self-derogatory attributions resulted in lower expectations for future success. Stipek and Gralinski (1991) also found that the girls' self-derogatory nature lowered their feelings of self-worth. Huber and Podsakoff (1985) showed that, because of confidence in their ability, males have higher expectations for future success; and because of females' attributions of their success to unstable factors, they feel less pride in their success. While there is good evidence that past successes and failures and the attributions for those events do lead to emotions of happiness, shame or guilt, and to expectations, more research is needed that shows a direct relationship to future achievement behavior and decisions.
Reasons for the Present Study
The present study compares the attributions and behavior of fourth-, fifth-, and sixth-grade boys and girls in a robotics course where the students built Lego[R] robots and learned to write computer programs for them. The study contributes to the extant literature in several ways.
First, few attribution studies have used computers or construction tasks. A notable exception is the study by D'Amico, Baron and Sissons (1995), but in their study the computer was a vehicle to provide tutorial or drill and practice in the language arts skills of increasing vocabulary and identifying types of analogies, typically a more feminine-typed task. Thus, the present study extends the research to a strongly masculine-typed task.
Second, most attribution studies have been done on college students. However, studies of students from roughly ten to twelve years of age are of particular interest because it is during these years that girls develop more negative assessments of their accomplishments and maladaptive attribution patterns. The participants in the robotics course span the ages where girls begin to develop negative assessments and could provide teachers with information to use in crafting effective intervention strategies.
Third, the present study uses an authentic task in a natural setting rather than a more contrived task, such as solving anagrams. While the many laboratory studies that have been done allow for greater control over the pertinent variables and have the advantage that the researcher can manipulate the success and failure of the participants, it is also important to know if attributions operate in the same way in the daily lives of students. In fact, if Meece, Parsons, Kaczala, Goff and Futterman (1982) are correct in proposing that the value of the task is relevant, then one might expect tasks in natural settings to produce stronger differences in attributions than laboratory tasks.
Fourth, the operational definition of success or failure in this study is the participants' own subjective opinion based on their judgment of how well they were able to program their robot to perform various tasks. Studies where achievement is determined by a test score are relevant since course-work is a necessary precursor to a technical career, but it would also be useful to know if attributions follow the same pattern when the definition of success is based on the participants' self-assessment and that assessment is informed by their success in designing and programming their robot - a task that is more like one that would be encountered in the work place. Perhaps boys' more positive attributions are due to the fact that they are more likely to have informal and authentic experiences in the areas of home construction and repair projects and computer use.
Fifth, while there is substantial evidence that successes and failures and students' attributions for those outcomes are related to emotions of shame and pride, to self-concept, and to expectations, the link of actual behavior to attributions is more tenuous. The present study not only assesses students' attributions but also their spontaneous comments and behavior by directly observing and coding videotapes made of the students while they worked. Thus, the relationship between cognitions and behavior can be examined.
The primary research questions of this study are: 1) What are girls' and boys' perceptions of their relative success or failure in a robotics class, and what reasons do they give for their performance? 2) What is the relationship among the girls' and boys' self-assessment of accomplishments, attributions for those accomplishments and their behavior during the course? While there is substantial variation in the details of previous work on attributions, there are also notable trends. Those trends led the researchers to predict that for this task, which is particularly masculine, boys would report greater success than girls even though their achievement might be comparable to that of the girls. It was also expected that boys would claim high ability as being responsible for their successes and that they would make self-protective and mastery-oriented attributions about their failures. In addition, it was expected that girls might attribute their successes more to effort or to external factors and blame their lack of ability for their failures. The researchers hypothesized that girls' behavior would reflect their negative self-assessment and attributions.
The participants were students entering fourth- through sixth-grade and enrolled in a one-week robotics course during a summer Science Academy, a division of a larger program called "College for Kids" sponsored by a mid-western university. The study took place over three summers, and different students enrolled in the class each summer. The first two summers featured single-sex classes, and the third year's classes were coeducational. Each summer the number of boys who wanted to take the course exceeded the number that could be accepted due to the limited number of available computers and robots, and each year girls were recruited in order to have sufficient numbers of girls for the purpose of the study. Thus, the present study compares volunteer boys with primarily recruited girls, which could have been a serious threat to the study's validity if the girls had not fared well in the course. Table 2 shows how many boys and girls volunteered or were recruited, as well as the number of non-white participants each summer. Girls were recruited by calling area science teachers, contacting enrolled program girls who did not have a full schedule, and passing out additional flyers late in the school year. The tuition was waived for recruited girls.
The resulting classes had 18 female participants and 21 male participants, respectively, for the single-sex sessions the first two years and 6 girls and 12 boys in both coeducational sessions for the third year. The students were primarily from middle-class families willing to spend $90 on a summer program, although scholarships were available to qualifying families. While there were no academic requirements for participation, many area schools distributed materials primarily to stronger students.
Students worked in groups of three, composed as much as possible of one student from each grade. The students in each group were not acquainted with each other prior to the study. In the first two years, two groups of three students in each session were randomly selected and videotaped daily as they worked. In the coeducational sessions only one group was videotaped each session. Each of the videotaped groups had one student from each grade.
The class featured groups of three students working together to build one of four robots with Legos[R] and using the RoboLab[TM] software to write programs. The programs were downloaded through an infrared (IR) transmitter to a computerized block known as a Robotics Command System (RCX) that in turn could run the program to make the robots move, make sounds, and perform simple tasks. Students worked through a packet of worksheets with directions designed as a tutorial to teach the basic programming commands used by the RoboLab[TM] software and what effect they would have on the robot. In addition, the worksheets included challenges in which the students needed to design and construct additional features for their robot for a given purpose. For example, students with a bug robot had to design pinchers for their bug and then write a program that would use a light sensor to detect something in the bug's way and capture it by moving the pinchers. A team of teachers circulated, helping students with questions and in working together cooperatively.
Each day at the end of the class, students were asked to fill out a brief questionnaire. In all three years the researchers measured students perception of success by asking them to rate their accomplishment on a scale of 1 to 5, 1 being the highest and 5 being the lowest. In year three, items were added that asked the students to attribute their accomplishment to ability (I am good at Legos[R] or I am not so good at Legos[R], and I am good at computers or I am not so good at computers), effort (I worked hard today or I didn't work so hard today), task difficulty (The work today was challenging or The work today was easy), or luck (I have a good group or I don t have such a good group). Students chose between the sets of opposing statements and could check as many attributions as they felt applied.
In interviews students were asked to recall a time they had done well and a time they had done poorly on a math and a reading test and to give a likely reason for their success and failure. Students were also asked about any previous experience they had with computers or Legos[R].
Each videotape was observed and coded using codes developed for the purpose of this study. Each student group was observed while working on the first through fifth pages of their ten-page packet, as well as during the engineering challenge on the ninth page, resulting in roughly 300 minutes of videotape footage per group over a representative sample of the tasks in the course. Colleagues of the researchers practiced coding, and differences were discussed until the coders were able to code half an hour of videotape with an average inter-coder reliability of 88.4%. They had reliabilities of 94.4%, 83.3% and 84.6% for the comments, questions and off-task categories, respectively. The observation frequencies were examined for significant gender differences using chi-square analyses. Survey responses about accomplishment for boys and girls were compared using t-tests.
Codes for Videotaped Behavior
The researchers developed codes for students behaviors that they believed were indications of students' self-confidence, perceptions of and attributions for successes and failures, and expectations and persistence. Table 3 lists the codes in each of the three categories-- comments, questions and off-task-- and provides definitions and examples for each.
Comments. The comments category included a total of six codes in three opposing pairs: a) success and failure statements and any attributions made regarding the success or failure; b) comments indicating self-assurance or lack of assurance about a suggestion, idea or likely cause, particularly about an unexpected result; c) independent or dependent comments indicating a desire not to seek or to seek help from a teacher. Only comments that applied to and included the speaker were coded. For example, if a student said, "We did it!" the comment was counted. However, a comment, such as, "You did that wrong," is an accusation that does not apply to the speaker and was not counted. An example of a negative comment attributed to luck would be "This didn't download. I think the computer is broken." Self-assurance was often indicated by students repeating their ideas and insisting that the group act on them. A comment indicating a lack of assurance would be, "I don't get this." An independent comment was one that a student made indicating a desire to figure something out without teacher assistance, such as, "I think we can get this," and a dependent comment was one in which a student suggested a need for help from a teacher, such as, "Let's ask for help."
Questions. The second coding category of questions examined to which individual students addressed their questions. The possible codes were: 1) to another student; 2) to a teacher who was currently working with the group; 3) to a teacher who was near and available; and 4) to a teacher who was far enough away that the students needed to make an effort to get the teacher's attention, often by getting up out of their seats and walking over to the teacher. Only task-related questions were coded. Procedural questions, such as, "What page are we on?" or "Whose turn is it at the computer?" were not included, nor were questions that functioned as suggestions, such as, "Why don't we change the motor icon to go that way?" Examples of questions that were coded are: "Why didn't this download?" or "Which icon do I need to use?" If a student repeated a question to the same person, it was counted only once, but if the student repeated it to another person in a different coding category, it was coded each time.
Off-task. Two types of off-task behavior were coded: individual and social. The off-task category was measured by the number of off-task episodes and the total time spent off task. Individual off-task behavior involved only one group member in such activities as twirling around in a chair, staring off into space or leaving the group. Social off-task behavior occurred when a student was engaged in an off-task activity with another group member. Social conversation sharing personal information was a typical social off-task behavior. The total number and length of social off-task episodes were coded for each group rather than for each individual. T-tests were used to compare boys' and girls' mean number of off-task episodes and mean off-task time.
To answer the research questions of whether boys and girls viewed their performances differently, gave different explanations for their outcomes, and behaved in ways that were consistent with their self-assessments, the researchers examined course registration, student interview responses, student responses to a daily survey and videotapes of student behavior during instruction.
Girls showed a marked lack of initial interest in the course. In all three years the program organizers had to recruit girls, but they turned away boys for whom they had no room. It was not especially difficult to recruit girls, but girls seemed to need an individual invitation and the incentive of a tuition waiver. Overall, girls outnumbered boys in the College for Kids program, but boys outnumbered girls in science courses, with the exception of pre-med and rain forest.
In all three years students responded daily to a brief survey rating their accomplishment and interest. Unlike other studies that have given students grades or scores, the evidence that students had on which to base their accomplishment in this study was more ambiguous. Students could compare their progress with other groups by comparing the number of completed pages in the instruction packet. They also received feedback from running their programs and seeing if they had managed to program the robot to do what they expected. In this context, nearly all students saw themselves as successful and were pleased with their accomplishment despite the fact that all groups also experienced many small failures along the way. For the first two single-sex years there was no significant difference in the girls' and boys' ratings of their accomplishment during the first four days of the course, although the girls consistently rated their accomplishment higher than the boys all four days. On Friday the difference between the girls' mean ratings ([bar.x] = 4.9) and boys' mean ratings ([bar.x] = 4.4) of their accomplishment, on a scale of 1-5, with 1 being the worst and 5 being the best, did reach significance (t = -2.35, p = .022, df = 69). That trend continued in the third year with the mixed gender groups. None of the daily differences was significant, but a pattern emerged: on average, girls felt they accomplished slightly more ([bar.x] = 4.7) than the boys thought they accomplished ([bar.x] = 4.4). This result was especially interesting because in the third year the boys and girls were in mixed-sex groups, so their accomplishments were actually the same.
In the third year the daily survey also asked students to make attributions for their accomplishment. The ratings of accomplishment were so uniformly positive for boys and girls that virtually all the attributions were for success. Of a total of 34 students, 21 attributed their excellent accomplishment every day to high ability with Legos[R] and computers, to the luck of having a good group, and to working hard. About half the students said that the tasks were challenging with a few more students saying Thursday, more than any other day, was challenging, but the difference was not related to student sex (chi-square = 3.09, p = .079, df = 1). In addition, even students who characterized the tasks as challenging did not consider that much of an obstacle to their accomplishment. Of the 13 students who made some negative attributions, six made such attributions on only one day, and for both boys and girls the most common negative attributions were the self-protective ones of bad luck or lack of effort.
In their interviews, girls and boys uniformly attributed their remembered success on a math or reading test to ability. Likewise, they almost all had difficulty answering the question about failure because it was hard for them to even imagine that happening, but they almost always said that if it did happen, it would be due to a lack of effort on their parts.
Behavior Observed on Videotapes
Chi-square tests were used to compare the frequencies of observed verbal behaviors for girls and boys in two categories: 1) questions (asked of peers and teachers); 2) comments (assured/unsure, success/failure and independent/dependent). Table 4 shows the results of the chi-square analyses.
Questions. The difference in the behavior of girls and boys by number of questions asked of teachers and peers was highly significant (chi-square = 15.65, p = .000, df = 1) with girls and boys addressing similar numbers of questions to peers, but with girls asking many more questions than expected of teachers. Because girls did not seem to have discernibly more trouble with the tasks than boys, teacher proximity was examined as a possible cause of girls asking more questions. Perhaps girls were simply more articulate or more interested in developing a relationship with teachers and thus asked questions of nearby teachers for that reason rather than because they genuinely had more questions. All of the teacher questions were categorized according to the proximity of the teacher when the question was asked. The three categories of teacher proximity were: 1) helping teacher (already working with the group); 2) close teacher (addressed to a teacher in close proximity) 3) far teacher (addressed to a teacher that one of the group members had to go ask to come over and help the group). Again, the chi-square test was highly significant (chi-square = 31.548, p = 0.000, df = 2). Figure 1 shows the frequencies of boys' and girls' questions by the person addressed. The pattern was also interesting. Girls did, in fact, address more questions to teachers during the course of the teacher's interaction with the group, but girls were also much more likely than expected to seek teacher assistance even if they had to get up and find a teacher. In contrast, boys almost never went in search of a teacher. Girls also asked a few more questions of close teachers, but that difference made a negligible contribution of the chi-square value.
Comments. The proportion of unsure and assured comments also showed a statistically significant difference between girls and boys. Figure 2 shows the frequencies for assured and unsure comments for girls and boys. Boys were more likely (chi-square = 4.843, p = .028, df = 1) than were girls to make assured comments, yet boys and girls made similar numbers of unsure comments. The difference between girls' and boys' spontaneous success and failure statements was not statistically significant (chi-square = 2.033, p = 0.15). Figure 2 also shows that girls made somewhat more failure statements than boys but a similar number of success statements.
[FIGURE 1 OMITTED]
On the video, students did not make many spontaneous attributions for their success and failure statements. Nonetheless, it was interesting to compare the pattern of the attributions they did make with what they claimed in the survey. Table 5 indicates that in their spontaneous attributions girls were at least as likely to claim ability in success as were boys, and possibly even more so. However, in failure, the pattern of attributions was similar to that reported in the literature where girls are more likely to attribute their failures to a lack of ability, and boys are more likely to attribute failure to task difficulty.
The frequencies of independent and dependent comments were too low in some cells to perform a chi-square test, but the pattern was clear in that girls made 18 dependent statements while boys made only two. Virtually no independent statements were made by either girls or boys on the taped segments chosen for the study, although one boy group asked teachers to leave them alone during some other portions of the course.
[FIGURE 2 OMITTED]
Off-task. The total off-task time, total number of off-task episodes and mean time per episode were calculated for each student as he or she worked on the course activities chosen for the study. The typical length of video-tape from which data was collected was about 300 minutes for each group, but some groups took longer to complete the work than others, sometimes because of greater off-task time. Table 6 shows the results of the t-tests adjusted for lack of homogeneity of variance that were performed to compare the off-task behavior of boys and girls. The p-values of 0.091 for gender differences in total off-task time, 0.148 for the number of off-task episodes, and 0.453 for average seconds per episode indicate that there was no significant difference between girls and boys for any of the measures, although on average the girls were off task for more time and for a greater number of episodes. However, the major finding was that neither boys nor girls were off task for very much time, although a few individual girls were off task for as long as 30 minutes.
This study also compared girls' and boys' off-task social time. Social time was defined as a time when two or three group members were off task and engaged in social conversation with other group members. Thus, off-task social time was calculated for each group rather than for individuals, making our sample too small to do a statistical test to determine whether the means were significantly different. The results in Table 7 show that girls did engage in more social behavior than boys, and it is interesting to note the pattern of the differences by the composition of the groups. The pattern suggests that the greater the number of girls in a group, the more socializing takes place. The group with one girl and two boys was the exception to the pattern.
The major question raised by the results of this study is why girls' spontaneous behavior in making enrollment decisions and in working during the course was more negative than that of boys, while their self-assessment of accomplishment and their attributions were as positive or even more positive than that of the boys. Examining the results from the perspective of the model proposed by Meece, Parsons, Kaczala, Goff and Futterman (1982) suggests that choices, in this case the girls' choice not to take the robotics course, are based on socializers, success and failure in related past events and the child's interpretation and attributions for those events, as well as task-specific beliefs and expectancies. Since the fourth- through sixth-grade students had little prior experience with computer programming or engineering tasks, socializers must have played the dominant role in influencing the girls' enrollment decisions. It is possible that the girls saw the course as related to some prior math or science experiences that had a negative influence, but the girls' attributions for the question about a failed math test during the interview and their expressed confidence in working with computers make the researchers think that it was much more a matter of socializers than of related negative experiences. Such an interpretation is further supported by the girls' willingness during the course to claim ability with Legos[R] and computers as responsible for their success rather than effort or luck alone.
If socialization was the primary influence responsible for the low girl enrollments, the next question is why that socialization did not have a similar influence on the girls' assessments of their accomplishment in the course and their attributions for those accomplishments. Since nothing on Meece, Parsons, Kaczala, Goff and Futterman's (1982) list of influences is likely to have had a big impact in the girls' lives between their registration and actually taking the course, then perhaps it was the course experience itself that caused the girls' positive assessment of their accomplishment and attributions for success. It is also possible that girls in 2001 have received the message that they can do whatever they choose and be as successful at it as boys, and that once girls had registered for the course, they expressed that view on the daily survey.
Given the girls' positive reactions to the robotics course, a further surprise was that their behavior during the course seemed not to be consistent with those reactions and was not as positive and confident as that of the boys. Perhaps their behavior reflected a different set of socialization messages that they had received about what boys and girls are interested in and where they are naturally talented. While the Meece et al. model suggests and the researchers expected that there would be consistency among all the measures of performance, attributions, and behavior, the finding that socialization gives mixed messages to girls is a very real possibility.
Another likely explanation for the seeming inconsistency of the girls' behaviors, such as asking more questions of teachers and making fewer self-assured comments, is that those behaviors do not reflect frustration with failures, a lack of self-confidence, over-reliance on teacher expertise and unwillingness to persevere with a masculine-typed task, but rather reflect a style of interaction. Perhaps girls have been socialized to be more modest and self-effacing and to develop relationships with their peers and teachers, or perhaps they have a different understanding of the task at hand as one where getting it right is more important than figuring it out. If that interpretation is correct, then teachers need to take that into consideration rather than viewing the girls as deficient and in need of fixing. In contrast, boys' seeming self-assurance might actually be a false self-confidence or a way of trying to get one's ideas accepted by the group as opposed to actual knowledge and self-confidence. In fact, some of the teachers' informal impressions were consistent with that interpretation. Rather than feeling that the girls were too dependent, some of them felt that the boys were independent to the point of being hampered in making progress because they were working under some basic misunderstandings.
Possible Reasons for Apparent Success of Course
Because the technology course and the research design had a number of features that made it different from the circumstances in which accomplishment and attributions are often studied, it is not possible to know whether it was one of those differences or some combination of them that made the course successful, or whether any technology course that managed to recruit girls would be successful. Thus, it is important to identify the ways in which these particular circumstances might be responsible for both girls and boys having positive experiences.
Operational definition of success via feedback from the robots. Students in the robotics course did not receive test scores or grades. Rather, they could consider themselves successful when they had met a challenge. For example, students who engineered a sunroof for their car and successfully programmed the car to open its roof when bright light was encountered could count themselves as successful and move on to the next challenge. Students could see what other groups were doing and could compare progress based on pages completed in the packet, but each group's solution to a particular challenge was different, and the teachers did not publicly label any solutions as more sophisticated than others. It might not be surprising to find no differences in sense of accomplishment under such circumstances, but that recruited girls (even those in mixed-gender classes) considered themselves to have accomplished more than the volunteer boys was unexpected. Perhaps as a result of socialization, the girls did not expect to be successful, and then when they worked in a situation in which it was relatively easy to interpret one's accomplishment in a positive way, they felt even more accomplishment than the boys who expected to succeed.
Operational definition of attributions. Although the results from the interviews and the surveys were consistent with one another, it is possible that using the procedure that some studies have adopted and forcing students to select the one most important attribution or to rate the relative importance of attributions would have revealed differences that we did not detect. However, in a later study with older students in which participants were asked to rate the importance of various attributions, the results were similar.
Single-sex. It is possible that the single-sex class could have influenced girls' sense of accomplishment or their explanation of that accomplishment, and there is some support in the literature for the effectiveness of single-sex classes. However, the fact that the girls' self-assessments were equally positive in the third year with mixed gender groups suggests that this factor did not substantively affect the girls' self-assessments.
Cooperative learning. A number of researchers, most notably, Peterson and Fennema (1985) and Oakes (1990), suggest that cooperative learning is a style of instruction that promotes the learning of girls. In addition, Peterson and Fennema (1985) and D'Amico, Baron and Sissons (1995) even suggest that cooperative learning may interfere with boys' learning in some cases. Clearly, the cooperative structure of the class is a possible explanation for the results and deserves further study.
Higher achieving students. Nearly all of the participants were above-average achievers in their schools and in their interviews had a very difficult time making likely attributions for a failed math or reading test because they simply could not imagine that scenario happening. Perhaps because of their successes in all school subjects in the lower grades, high-achieving girls are slower to adopt some of the typical "girl" attitudes and attributions, even though they have clearly learned "girl" behavior and make decisions consistent with their gender role.
Ample access to teacher help. While the tasks in the curriculum were quite challenging, the students did have the benefit of several teachers being available to provide assistance, instead of the more typical teacher-student ratio of 1:30. It is possible that it was this access to teacher assistance that promoted girls' positive sense of accomplishment.
The findings of this study underscore the influential role of socialization in girls' decisions and make it clear that even young girls are much less likely than boys to take advantage of opportunities to engage in technology experiences. Furthermore, even when girls consider themselves successful in using technology and make positive attributions about their performance, they display behavior that might well be interpreted as self-deprecating. Apparently, such behavior is learned and internalized prior to actual experience and prior to making self-deprecating attributions. The implications for teachers are that they should not only try to provide positive technology experiences for girls, but they should also address cultural messages directly. They also need to be aware of girls' learning style and accommodate it rather than view it as deficient. Finally, both male and female students might profit from a discussion of when to ask questions and when to figure things out on their own.
Table 1 Four attributions according to internality, stability, and controllability Internal External Stable Unstable Ability X X Effort X X Task X X difficulty Luck X X Controllable Uncontrollable Ability X Effort X Task X difficulty Luck X Table 2 Number of volunteer, recruited, and non-white participants each year Girls Girls Boys Non-white Year Volunteered Recruited Volunteered Participants First Year 3 15 28 3 (single-sex) Second Year 12 6 21 8 (single-sex) Third Year 6 6 24 4 (coeducational) Table 3 Definitions and examples of codes developed for the study Code Definition Example Comments Success statements Positive comment about "We did it!" and attributions progress or outcome Failure statements Negative comment about "This didn't and attributions progress or outcome download. I think the computer is broken." Assured Self-confident and insistent "No. Change the about knowledge arrow. Change the arrow. That's what we need." Unsure Not confident about knowledge "I don't get of the task at hand this." Independent Does not wish to have help "Let's keep trying from a teacher this before we ask a teacher for help." Dependent Desires help from a teacher "We're not getting anywhere. Let's ask for help." Questions To a student Task-related question asked of "Which icon do I a group member need to use?" "Why didn't this download?" To a helping Task-related question asked of "So this number teacher a teacher already working with always means the group speed?" To a close teacher Task-related question asked of "We need help." a teacher who is close to the group To a far teacher Task-related question to a "Can you come help teacher far enough away that us get our door to the student needs to get up to open? It worked attract the teacher's once, but now it attention isn't working." Off-task Social Two or more group members "What grade are you engaged together in the same in?" off-task behavior Individual A group member is off-task Watching or without distracting another conversing with member of own group member of another group, twirling in chair Table 4 Comparisons of behavior frequencies for girls and boys Test Chi-Square P-value Degrees of freedom Questions addressed to 15.658 0.000 ** 1 peers and teachers Questions to teachers by 31.548 0.000 ** 2 teacher proximity Assured/Unsure Comments 4.843 0.028 * 1 Success/Failure Comments 2.033 0.154 1 Independent/Dependent Insufficient cell Comments counts * p < 0.05 ** p < 0.001 Table 5 Boys' and girls' spontaneous attributions for success and failure Ability Effort Task Difficulty Luck Boys in Success 4 0 0 1 Girls in Success 9 0 1 1 Boys in Failure 5 0 11 4 Girls in Failure 11 0 3 3 Table 6 Comparison of boys' and girls' off-task time Group Composition Average time off task per episode All-girl groups 37.9 seconds/episode Two girls & one boy 20.2 seconds/episode All-boy groups 14.89 seconds/episode One girl & two boys 10.7 seconds/episode Table 7 Comparing social off-task time by gender makeup of each group Off-Task Girls' Boys' t-statistic df p-value Measure mean mean Total time 583 sec 275 sec -1.82 14 .09 Total 24.9 19.1 -1.50 21 .148 episodes Seconds per 18.7 24.9 .75 21 .453 episode
Bar-Tal, D. & Frieze, I. H. (1976). Attributions of success and failure for actors and observers. Journal of Research in Personality, 10, 256-265.
Basow, S. A. & Medcalf, K. L. (1988). Academic achievement and attributions among college students: Effects of gender and sex typing. Sex Roles 19(9/10), 555-567.
Covington, M. V. (1992). Making the grade: A self-worth perspective on motivation and school reform. Melbourne, Australia: Cambridge University Press.
Covington, M. V. & Omelich, C. L. (1979a). Effort: The double-edged sword in school achievement. Journal of Educational Psychology, 71(2), 169-182.
Covington, M. V. & Omelich, C. L. (1979b). It's best to be able and virtuous too: Student and teacher evaluative responses to successful effort. Journal of Educational Psychology, 71(5), 688-700.
D'Amico, M., Baron, L. J., & Sissons, M. E. (1995). Gender differences in attributions about microcomputer learning in elementary school. Sex Roles, 33(5/6), 353-385.
Fennema, E. (1974). Mathematics learning and the sexes: A review. Journal for Research in Mathematics Education, 5, 126-139.
Huber, V. L. & Podsakoff, P. M. (1985). Dispositional and situational moderators of female and male causal attributions. Sex Roles, 13(7/8), 441-461.
Levine, R., Reis, H. T., Turner, E. S., & Turner, G. (1976). Fear of failure in males: A more salient factor than fear of success in females? Sex roles, 2(4), 389-398.
Meece, J. L., Parsons, J. E., Kaczala, C. M., Goff, S. B. & Futterman, R. (1982). Sex differences in math achievement: Toward a model of academic choice. Psychological Bulletin, 91(2), 324-348.
Nicholls, J. G. (1979). Development of perception of own attainment and causal attributions for success and failure in reading. Journal of Educational Psychology, 71(1), 94-99.
Oakes, J. (1990). Opportunities, achievement, and choice: Women and minority students in science and mathematics. Review of Research in Education, 16, 153-222.
Parsons, J. E., Meece, J. L., Adler, T. F., & Kaczala, C. M. (1982). Sex differences in attributions and learned helplessness. Sex Roles, 8(4), 421-432.
Peterson, P. L., & Fennema, E. (1985). Effective teaching, student engagement in classroom activities, and sex-related differences in learning mathematics. American Educational Research Journal, 22, 309-335.
Rosenfield, D. & Stephan, W. G. (1978). Sex differences in attributions for sex-typed tasks. Journal of Personality, 46, 244-259.
Ryckman, D. B. & Peckman, P. D. (1987). Gender differences in attributions for success and failure. Journal of Early Adolescence, 7(1), 47-63.
Sanders, J., Koch, J. & Urso, J. (1997). Gender equity right from the start. Mahwah, New Jersey: Lawrence Erlbaum.
Simon, J. G. & Feather, N. T. (1973). Causal attributions for success and failure at university examinations. Journal of Educational Psychology, 64(1), 46-56.
Stipek, D. J. (1984). Sex differences in children's attributions for success and failure on math and spelling tests. Sex Roles, 11(11/12), 969-981.
Stipek, D. J. & Gralinski, J. H. (1991). Gender differences in children's achievement-related beliefs and emotional responses to success and failure in mathematics. Journal of Educational Psychology, 83(3), 361-371.
Sweeney, P. D., Moreland, R. L., & Gruber, K. L. (1982). Gender differences in performance attributions: Students' explanations for personal success or failure. Sex Roles, 8(4), 359-373.
Weiner, B. (1974). Motivational psychology and educational research. Educational Psychologist, 11, 96-101.
Weiner, B. (1994). Integrating social and personal theories of achievement striving. Review of Educational Research, 64(4), 557-573.
|Printer friendly Cite/link Email Feedback|
|Publication:||Journal of Computers in Mathematics and Science Teaching|
|Date:||Sep 22, 2004|
|Previous Article:||Combing online and paper assessment in a web-based course in undergraduate mathematics.|
|Next Article:||Developing a TI-92 manual generator based on computer algebra systems.|