School-based modifications of children’s gender-related beliefs
Gash H. & Morgan M. (1993) School-based modifications of children’s gender-related beliefs. Journal of Applied Developmental Psychology 14: 277–287. Available at http://cepa.info/2190
Table of Contents
Modification of Gender Stereotypes
Teachers and Preparation
Schools and Children
Format of Experimental Lessons
Reliability of Measures
Student teachers taught a program to reduce stereotyped beliefs among children from first to sixth grade. The program was developed on the basis of constructivist theory and involved teaching strategies designed to lead children to question existing beliefs. Compared to children in a control group, the children in the two experimental conditions were shown to have significantly less stereotyped beliefs on completion of the program. The effects of the intervention were relatively greater among boys and younger children. The effects of the program were noticeable 1 year afterwards in a subsample of girls who participated in the study. It is suggested that this approach had considerable potential for classroom application.
It would not be an exaggeration to state that the issue of equality is one that has vital implications for world peace. Particularly if one was to insist that the issue of equality is one that needs to be considered broadly, for example, in domains including class, ethnicity, gender, persons with disability, political belief, race, and religion. This article stems from research on the modifiability of gender stereotypes in elementary schools in Ireland. The context of this article was focused in Brussels by the Council of the European Commission and the Ministers for Education in the European Communities in June 1985. They resolved to initiate an action program on equal opportunities for girls and boys in education. As a result, the Association for Teacher Education in Europe responded with a course framework for teacher education (Arnesen & NíChárthaigh, 1987). Proposals were invited and projects began across Europe in the fall of 1988.
In Ireland within the educational community there was an awareness of the inequalities in the distribution of positions of responsibility within (Kellaghan, 1983) and outside the educational sector, and there was also an awareness of differences between boys and girls in their mathematical and scientific abilities (viz., Lapointe, Mead, & Phillips, 1989). The cultural climate in Ireland was right for the then Minister for Education and her department to facilitate research in this area. There were initially five projects at both the elementary and high school levels in the Irish education system. Details on all these projects can be found in Drudy et al. (1991). The research described herein was based on the idea that schools might play a role in the developmental processes by which gender stereotypes were formed, and if they could be shown to have an effective school role, appropriate programs could be initiated for interested teachers and administrators.
Modification of Gender Stereotypes
Efforts to reduce gender stereotyping have been a focus of many studies in the recent literature in psychology and education. Several experimental procedures have been found to be effective in reducing stereotyped beliefs and behaviors including countertraditional television programming (Williams, LaRose, & Frost, 1981), peer modeling (Katz & Walsh, 1991), and adult modeling (Raskin & Israel, 1981).
However, the educational relevance of much of this work is yet to be established. Attempts to modify stereotypes in educational settings have ranged from those that have a narrow, restricted focus in artificial settings to those that occur in naturalistic settings with classroom teachers. An example of a highly focused intervention is the study by Bern and Bem (1973), who found that when job advertisements were written in either unbiased or gender-reversed formats, high school seniors were more likely to express an interest in the “opposite” sex jobs than when the advertisements were written in traditional formats.
At the opposite extreme from the highly focused work is the broadly based type of intervention exemplified by the classroom experiment of Guttentag and Bray (1976). This large-scale study was carried out over a long period, involved a wide array of activities and materials, and used evaluation techniques that are distant in time and substance from the original intervention techniques.
Therefore, it would seem that many interventions suffer from problems deriving from an overconcern with either internal or external validity and to the detriment of one or the other. The narrowly focused studies have maintained internal validity but suffer from difficulties arising from the restricted range of situations to which the finding may apply. Conversely, the broadly based studies may suffer from a lack of clarity regarding the precise factor that was responsible for the observed outcomes.
Related to the matter of internal validity is the need to specify the precise elements of interventions responsible for change (Schau & Scott, 1984). In many studies the theoretical underpinnings of the intervention are not clear, resulting in difficulties in attempts to replicate findings.
The present intervention could be described as “midlevel” in terms of scope. While being carried out in a classroom context, it attempts to maintain precision in the delivery of the experimental manipulation. Furthermore, the learning factors involved are stated explicitly.
The approach taken is explicitly constructivist. In a constructivist theory, stereotypes are ways in which children come to organize their experiences and ways in which they come to an understanding of the world in which they find themselves (e.g., G. Bateson, 1972; G. Bateson & M.C. Bateson, 1987). In G. Bateson’s accounts, stereotypes are context-dependent, second-level learnings, which play an important role in an individual’s identity. That is, because stereotypes are ways in which children expect men and women to act and to be, they become ways in which their friends expect them to behave and think and so, mirrorlike, the stereotypes become part of their identity. There are a number of ways in which children can be invited to reconsider how they think about gender stereotypes. First, they can be questioned about their ideas; second, they can be presented with counterexamples. Statements that children make about males and females (their personal and social characteristics, their interests or careers) can be challenged by the teacher or by other children. Other tactics include a technique called distancing (Sigel, Stinson, & Flaugher, 1991). In constructivist theory, realization of the child’s abilities is said to be facilitated by distancing, that is, placing a cognitive demand on the child to reconsider the experience of the ongoing present. One possible outcome of distancing is of particular concern in this experiment, namely, the development of alternate perspectives on the present. Further details on the strategies used may be found in Gash (1991).
In line with the constructivist view, it is our view that children often forget or distort counterstereotypic information (Bigler & Liben, 1990), thus reducing the effectiveness of such exposure – a factor that may partially explain the limited success of intervention programs based on these procedures. A primary feature of the present study was the effort to minimize the probability of such an occurrence by increasing the salience of the stereotyped information. This was achieved by (a) the children encountering live counterstereotypic models, (b) giving the class period a special status, and (c) frequent revision of the salient features of the experiences.
The study reported here involved student teachers who learned to use such techniques in a college course. Subsequently, the teachers taught a program based on these features to classes in a field-experimental context. The effects on stereotyped beliefs and attitudes were then evaluated in comparison to control classes.
Teachers and Preparation
The 15 female student teachers were final year students at a College of Education in Dublin, Ireland. The student teachers (average age 20 years) had opted for a course on gender stereotyping offered by the senior author. This course lasted two semesters and required in part that the students teach a specified number of classes in schools over the course of the year. In the college-based component of the course students learned about the psychological and educational factors that are thought to be important in the development of gender stereotyping. They also learned about various techniques that would be appropriate in counteracting such beliefs. Particular attention was given to techniques that required the children to reconsider their ideas, such as questioning, distancing, role play, and the use of counterexamples, where, for example, children thought women could not be soldiers or men could not be nurses. The student teachers were given practice in use of these skills. Furthermore, the participants prepared detailed lesson plans, appropriate to the grade level to which they had been assigned and which featured these techniques.
There were two experimental conditions, a spaced experimental (SE) condition and an intense experimental (IE) condition. In the SE condition, the student teachers taught a program of four lessons each of average duration 45 min over a 4-month period. Students were randomly assigned to host teachers whose schools had volunteered to participate in the project. The procedures involved the student teachers taking over the class from the regular teacher at a prearranged time. The classes involved ranged from kindergarten to sixth grade. For each experimental class, a control class of the same grade, and in the same school participated in the measurement of the ou _come measure. Because the procedure of allotting children to classes within grades was random in all of these schools, the problem of getting a suitable control group was minimized.
The arrangement for the IE condition took advantage of the fact that the final student teachers are assigned to schools for about 1 month in February/March. This assignment is not fully random, but takes into account the student teachers’ previous experiences with different grade levels. Student teachers taught the same program in their assigned classes over this 1-month interval. Control classes were identified in the same way as in the SE condition.
Schools and Children
Some of the primary schools in the Dublin area are coeducational and some are single sex. Of the 997 children involved in the study, approximately 26% were in boys-only schools, 44% were in girls-only schools, and 30% were in coeducational schools. Thus, girls-only schools are overrepresented in the study and this, in turn, results in there being a higher representation of females (58%) in the sample.
Format of Experimental Lessons
As noted earlier, students prepared lesson objectives and notes that were appropriate to the ages of the children and also took into account children’s previous learning, social background, and intelligence. Information on these topics was obtained from the regular class teacher. The content of each lesson was selected with a view to maximizing the chances that pupils would encounter experiences and issues that allow for frequent use of the techniques that take seriously the constructivist perspective, namely, providing the children with opportunities to reconsider their ideas by questioning and counterexamples as described earlier.
The lessons taught to the children in this program were designed to call into question gender stereotypes, whether they referred to occupations, personal or social characteristics, or activities and interests. Where, for example, a child or group of children thought that women were only secretaries, or were warm and gentle, or were not interested in football, the teacher was to respect the child’s view but to try to create a context so that this view could be counterbalanced with another. The teachers were asked not to tell the child that he or she was incorrect but to point to examples that did not fit with the child’s expressed view. These examples could be invited from other children, or from the teacher’s own experience. Therefore, in the cases just mentioned, children were asked if they could think of women who worked at jobs other than secretarial jobs, if they could provide instances where men were warm and gentle, and if they knew any women who were interested in football. The student teachers were to seek to “wobble” the children’s certainties about stereotypes. Sometimes this was done through discussion, sometimes using worksheets, or sometimes by means of live counterexamples such as female veterinary surgeons, or male nurses, or whatever occupation the children thought of in a stereotypical way. Different lessons were used which were appropriate to the age level of the children being taught.
As part of their course requirements, the students submitted details on their lessons in the SE section of the study, together with comments on how the lessons were received by the children. Examination of these materials gives an indication of the type of activities that took place in the classrooms during these interventions. In the junior classes (6-year-olds) there were 8 lessons on careers, 13 story-based lessons, 2 on toys, and 1 lesson based on a poem. Among the 13 stories, 3 were based on media (TV, video, comics) and 2 were “reversed” fairy tales in which a prince was rescued by a princess, or in the case of Rapunzel, a girl rescued a boy from the tower. In the 7- to 9-year-old age group, the 16 lessons included 2 whose themes were explicitly careers, 3 were stories, and, in addition, there were 2 stories in which the gender of the character was deliberately hidden to stimulate guessing and discussion, and another version of Rapunzel reversed. There were 3 lessons on women in history, and 1 based on each of the following topics: singing, poetry, hobbies, fairy tales, and the media. In the 10- to 12-year-old age group, there were 6 lessons on important women in the world of history, science, and politics, there were 8 on equality issues (e.g., in housework, in the home, on the function of stereotypes), and there were 6 lessons using a story format (e.g., reversed fairy tales, poetry, stories in which the gender of key characters is hidden, and a video of stereotyping in a school context).
Two dependent variables were of particular interest in this study. The test of gender perception was a modified version of the Gender Stereotype Measure (SSM) (Best et al., 1977). The original version of this 32-item scale presents an activity or trait and asks subjects to indicate whether the activity or trait is likely to be associated with a man or a woman. One typical item is: “One of these people is adventurous” and asks subjects to select a male or a female. Our preliminary piloting of this test indicated that children resisted being forced into the stereotyped or counterstereotyped selection. Thus, we modified the response to a 5-point scale ranging from always a boy, usually a boy, either a boy or a girl, usually a girl, to always a girl. Otherwise we followed the procedures and item ordering used in the original questionnaire which was carefully counterbalanced. There were 16 male and 16 female stereotypes on this test, with the 5 points being given to responses that “correctly” identified the stereotype, for example, adventurous being categorized as always a male.
In addition, an attitude and belief scale was given to the teachers. Previous work with this scale by Lynch (1991) had indicated that the attitude scale had comprised the following subscales: (a) 6 items relating to beliefs on gender equality (e.g., “Girls can cope with figures as well as boys”); (b) 8 items focusing on awareness of gender discrimination in the educational system in Ireland (e.g., “Teachers tend to discriminate unfairly against girls in mixed classes”). This scale, which was in the traditional Likert-5 format (strongly agree – strongly disagree) was administered to the teachers (class and student teachers) at the end of the experiment. To provide a comparison for the student teachers, a group of final year students (matched on age and gender) also completed the scale.
In the spring of the following year, SSM data were again collected from a subset of the original sample. These children were all in an all-girls school.
Reliability of Measures
Two estimates of reliability were made with regard to the main dependent variable (SSM), namely, internal consistency and test – retest reliability. The Cronbach test of internal consistency yielded an alpha of .82, whereas a pilot test of the SSM in which the test was given to grade levels from kindergarten to sixth grade showed reliability coefficients ranging from .58 for junior infants to .81 for sixth graders. These outcomes are considered satisfactory as were the internal reliability coefficients for the ASD and BSE scales (a = .72 and .79, respectively).
As explained earlier, the main measure of interest in the evaluation was the
modified SSM scale. The most relevant comparison was between the SE condition, the IE condition, and the control group. Equally important was the test of the extent to which the treatment effects are similar across all subject groupings. Were the effects the same for boys and girls, for different age levels, and for different types of classes? These questions were examined using one-way analyses of variance (ANOVAs, hierarchical method).
The two experimental and control conditions gave mean scores of 95.98, 94.33, and 99.08, respectively. An ANOVA showed that the difference between the conditions was highly significant, F(2, 982) = 27.66, p .001. Furthermore, Scheffe planned comparisons of the two experimental conditions with the control condition showed that these differed significantly from the control, while the two experimental conditions were not significantly different from each other. Thus, it can be concluded that there were strong and significant treatment effects.
It also emerged that the mean scores for boys were significantly higher than for girls (i.e., were more stereotyped in their opinions). This difference was highly significant: F(1, 982) = 35.039, p .001. Second, the interaction of gender and treatment is significant, F(2, 987) = 3.33, p .05. As can be seen from Table 1, the interaction seems due to the relatively greater impact of the experimental conditions in the case of boys.[Note 1]
Table 1: Scores for Boys and Girls in Each Condition
*Spaced Experimental ConditionsIntense Experimental ConditionsControl ConditionsBoys96.51 (145)96.24 (135)101.26 (186)Girls94.38 127)93.83 (171)96.99 (229)
Another important point concerns the extent to which all grade levels were significantly influenced by the experimental conditions. Table 2 shows the results for grade-condition combinations. The overall difference between the grades was statistically significant: F(2, 982) = 8.40, p .001. There was also a significant interaction of grade with treatment, F(2, 982) = 3.91, p .025. This interaction seems largely due to the fact that the treatment had relatively stronger effects in the junior grades.
Table 2: Scores in Grade-Condition Combinations
*Spaced Experimental ConditionsIntense Experimental ConditionsControl Conditions1st/2nd Grade94.41 (61)93.89 (131)97.29 (129)3rd/4th Grade95.08 (71)92.76 (110)98.16 (140)5th/6th Grade96.78 (140)98.63 (65)99.34 (146)
Although there are important questions about the extent to which treatment effects were similar across all types of schools (single sex versus mixed) and also in relation to teachers’ gender, these questions were impossible to answer in the design of the present research. This is because these variables were strongly related to the other factors considered (i.e., pupil gender and age of pupils). Thus, although there were significant interaction effects associated with school type and also with teachers’ gender, these effects were largely due to the associated effects described earlier.
In the longitudinal sample, 130 experimental (SE) children and 80 controls were located across three grade levels (1st, 5th, and in the first year at high school). Few children were lost in this sampling; the greatest number at any grade level being 19 in the oldest group where the children had to be followed to other schools. Mean differences between the experimental (108.08) and control (111.25) groups were analyzed using ANOVA and found to be significant, F(1, 207) = 9.94, p .002. SSM scores were lower in the experimental condition at each level.
Comparisons were made between the experimental and control teachers/student teachers on scales relating to (a) awareness of sex discrimination (ASD) and (b) belief in sexual equality (BSE). The experimental group included the students who taught the program, and their host teachers in the SE part of the study who had participated in a similar program during the previous academic year. Data were obtained from 24 experimental students and teachers and from 27 control students and teachers who did not participate in the program. On the BSE scale, the experimental and control groups had mean scores of 26.13 and 25.15, respectively, F(1, 46) = 1.61 n.s. However, on the ASD scale, a somewhat different pattern emerged, with the experimental and control groups having mean scores of 17.88 and 15.69, respectively, F(1, 46) = 14.81, p .001. This indicates that the experimental experiences had the effect of increasing teachers’ awareness of sex discrimination.
The main purpose of this experiment was to modify the perception of gender stereotypes in a field setting, making use of specific teaching strategies based on the constructivist perspective. In comparison to a control group, two experimental groups showed less stereotyped responses on a modified version of the SSM scale. The difference between the spaced experimental condition and the intense experimental condition was not significant. Furthermore, the effect of the experimental treatment was relatively greater in the case of boys, which was related to the fact that all-boys schools tended to show greater experimental effects. Another feature of the results was that the experimental treatment tended to be somewhat greater in junior classes than in senior classes. The experimental treatment was also seen to have an impact on the beliefs of the participating teachers, especially in the area relating to awareness of gender discrimination.
The finding that boys showed more change in gender-stereotyped beliefs might seem at first to run counter to much of the research that suggests that girls are more susceptible to change (e.g., Ashton, 1983). However, on closer examination, it is found that girls’ greater susceptibility to change found in earlier work is more likely to emerge in behavioral measures relating to involvement in gender-typed activities (e.g., Bussey & Perry, 1982). In contrast here, it is beliefs and judgments about social characteristics that are central to the modified SSM.
The finding that younger children were relatively more susceptible is a particularly interesting finding. This outcome is consistent with the view that learning of stereotypes increases in the course of middle childhood (Carter & Patterson, 1982; Martin & Little, 1990). It is also consistent with the findings that peer influences characteristics of early adolescence tend to exert a conforming influence during those years, resulting in less flexibility in relation to gender role conceptions (e.g., Hartup, 1976).
An important consideration with the present type of study has to do with the extent to which the changes indicated by the dependent variables are reflections of deep-seated changes of opinions and attitudes rather than mere verbal learning based on the lessons on stereotyping. At issue here is the crucial issue of transfer, and although the significance of a given response is difficult to establish with certainty, we have considerable confidence in the substantive significance of the outcomes. A number of factors support this line of argument. First, the lessons were selected on the basis that they would illustrate broad aspects of gender equality. Second, care was taken to avoid as far as possible the specific descriptors that provide the basis of the SSM test. Third, the effects persisted in a sample of girls examined 1 year later.
One of the implications of the constructivist viewpoint is that children may take away from an experience something other than what the experimenter intended (Liben & Bigler, 1987). Thus, Hurwitz & White (1977) showed that when subjects were given an article to read about new occupational opportunities for women, a substantial minority did not mention women at all in their descriptions, thus avoiding the experimental manipulation. In our view, the teaching techniques used in the present study were quite effective in ensuring that subjects were confronted with the salient aspects of gender equality.
There are ways in which the present work could be improved upon. One important question, for example, pertains to the extent that the treatment effects are due to the specific components of the experimental manipulation, rather than to nonspecific expectation effects. This latter issue is especially difficult to rule out in a field experiment.
However, the results are particularly hopeful in an area where results have frequently been less than promising. The fact that these positive outcomes were achieved without any undue disruption to the normal classroom schedule is especially significant. Furthermore, the experiment did not require any resources or training in addition to those normally obtained in classrooms. These factors suggest an optimistic prospect for the extension of this kind of program to classrooms. A study is currently under way along similar lines to investigate the possibility of reducing prejudice toward persons with special needs.
This research was supported by the European Commission and the Department of Education (Ireland) under the TENET Program at St. Patrick’s College. Thanks are due to the teachers and students who cooperated in this project.
Correspondence and requests for reprints should be sent to Hugh Gash or Mark Morgan, Education Department, St. Patrick’s College, Dublin 9, Ireland.
Arnesen, A-L., & NíChárthaigh, D. (Eds.). (1987). Equal opportunities for girls and boys: A curriculum framework for teacher education with guidelines for action. Brussels: Association for Teacher Education in Europe.
Ashton, E. (1983). Measures of play behavior: The influence of sex-role stereotyped children’s books. Sex Roles, 9, 43-47.
Bateson, G. (1972). Steps to an ecology of mind. New York: Ballantine.
Bateson, G., & Bateson, M.C. (1987). Angels fear. New York: Macmillan.
Bern, S.L., & Bem, D.J. (1973). Does sex-biased job advertising “aid and abet” sex discrimination? Journal of Applied Social Psychology, 3, 6-18.
Best, D.L., Williams, J.E., Cloud, J.M., David, S.W., Robertson, L.S., Edwards, J.R., Giles, G., & Fowles, J. (1977). Development of sex-trait stereotypes among young children in the United States, England and Ireland. Child Development, 48, 1375-1384.
Bigler, R.S., & Liben, L.S. (1990). The role of attitudes and interventions in gender-schematic processing. Child Development, 61, 1440-1452.
Bussey, K., & Perry, D.G. (1982). Same-sex imitation: The avoidance of cross-sex models or the acceptance of same-sex models? Sex Roles, 8, 773-784.
Carter. B., & Patterson, C.J. (1982). Sex roles as social conventions: The development of children’s conceptions of sex-role stereotypes. Developmental Psychology, 18, 812-824.
Drudy, S., Gash, H., Lynch, K., O. Láimhín, P., Moles, R., Lane, C., Ganly, M., Fogarty, C., O’Flynn, G., & Dunne, A. (1991). Integrating equal opportunities in the curriculum of teacher education 1988-1991: TENET program dissemination phase. Irish Educational Studies, 10, 271-289.
Gash, H. (1991). Gender perceptions: A constructivist approach to stereotyping. Oideas, 37, 57-74.
Guttentag, M., & Bray, H. (1976). Undoing sex stereotypes. New York: McGraw-Hill.
Hartup, W.W. (1976). Peer interaction and the behavioral development of the individual child. In E. Schopler & R.J. Reichler (Eds.), Psychopathology and child development: Research and treatment (pp. 208-218). New York: Plenum.
Hurwitz, R.E., & White, M.A. (1977). Effects of sex-linked vocational information on reported occupational choices of high school juniors. Psychology of Women Quarterly, 2, 149-156.
Katz, P.A., & Walsh, P.V. (1991). Modification of children’s gender-stereotyped beliefs. Child Development, 62, 338-351.
Kellaghan, T. (1983). Equality of opportunity in primary-school teaching. Paper read at Irish National Teachers Organization Seminar on Equality in Primary Education, Dublin.
Lapointe, A.E., Mead, N.A., & Phillips, G.W. (1989). A world of differences; An international assessment of mathematics and science. Rep. 19-CAEP-01. Princeton, NJ: ETS.
Liben, L.S., & Bigler, R.S. (1987). Reformulating children’s gender schemata. In L.S. Liben & M.L. Signorella (Eds.), Children’s gender schemata (pp. 89-105). San Francisco, JosseyBass.
Lynch, K. (1991). Worlds apart: Gender inequalities in education and ways of eliminating them. Irish Educational Studies, 10, 275-276.
Martin, C.L., & Little, J.L. (1990). The relation of gender understanding to children’s sex-typed preferences and gender stereotypes. Child Development, 61, 1427-1439.
Raskin, P.A., & Israel, A.C. (1981). Sex-role imitation in children: Effects of sex of child, sex of model, and sex-role appropriateness of modeled behavior. Sex Roles, 7, 1067-1077.
Schau, C.G., & Scott, K.P. (1984). Impact of gender characteristics of instructional materials: An integration of the research literature. Journal of Educational Psychology, 76, 183-193.
Sigel, I.E., Stinson, E.T., & Flaugher, J. (1991). Socialization of representational competence in the family. In L. Okagaki R.J. Sternberg (Eds.), Directors of development: Influences on the development of children’s thinking. Hillsdale, NJ: Erlbaum.
Williams, F., LaRose, R., & Frost, F. (1981). Children, television and sex-role stereotyping. New York: Praeger.
Effect sizes were calculated on the basis of Effect size = ((X spaced + X intensive) / 2) – X control ) / X control. The effect sizes thus calculated were .048 for boys and .030 for girls.
Found a mistake? Contact corrections/at/cepa.infoDownloaded from http://cepa.info/2190 on 2016-05-13 · Publication curated by Hugh Gash