Psychological Bulletin Copyright by the American PsychoLogical Association Inc - PDF document

by the Vol. 100, Aggressive Behavior: A Meta-Analytic Psychological Literature Eagly and sex differences in the social psycholog- were somewhat average, sex differences were inconsistent across studies. sex differences was more than women was injury than ALICE H. J. STEFFEN (e.g., Eagly, 1983; & Crowley, 1986). underlie sex aggressive behavior. a single, ing sex that may Maccoby and Frodi et ences in The malegender role. aggression. Psychologists about men men to be aggressive (see 1978). Similarly, stereotype research women. Survey this role aggressiveness (see & Crowley, 1986). that men and support traditional female this role & Crowley, 1986) avoiding physical intended in less aggressive assertiveness, a has often & Baer, 1975). gender role, may be posed to only within may transmit gressive behavior. AGGRESSIVE BEHAVIOR women to implications for an element women in behavior in in the (Eagly, 1983). higher status status enjoins Overall sex-of-subject difference. all suggest direction. Indeed, ways social contextual variables differences in in the hypothesis stems the prediction emphasize primarily (and therefore jects faced whether the & Taylor, 1979). predict the in the for each Buss, 1963) for each multiple observa- distinction between physical psychological aggression has one variant aggression literature. Some investigators (e.g., Buss, 1961) have distinguished between these two overtly physical versus vocal nature aggressor's act. Following Frodi, Macauley, and Thome (1977), we the target. Yet Frodi collaborators labeled aggression causing psychological and, unlike it encompasses behaviors as- the studies we the research experimental situations typically allowed either psychological cal aggression--not both. VALERIE J. STEFFEN and the this meta-analysis involved strenuous people behave legitimize aggression people behave (e.g., Feather, the male role's emphasis role's emphasis and concern are usually primarily when gender role reported mixed only some in women should be with the gender role in various with this are the for the in the & Averill, 1982). inherent in in the research reports. in those in the are often ined in involved some toward women. to the extent that they if women tent that social status toward men, toward other toward men. aggression studies and perhaps this meta-analysis, expected that on social-role analy- As we 1974 (Vol. 1983 (Vol. 7); (b) differences; (c) aggression studies; aggression studies. the dependent results were a sex-of-subject correlate with process. 2 studies were assessed aggressive drugs on 1977) was (a) as- sessed a provided), these a single From Each source), (c) male, male, variables were the information field); (c) aggression (physical, i.e., accessible aggress (aggression free choice); provocation (minimal, than minimal, opportunity to subject). 3 variables were following studies (1974, 1975), aggression was instigated anger to be aggression (e.g., presumed to be aggression was as its value in teaching vary with aggression (e.g., magnitude and which constituted the dependent variable in by the to generate in the included to Respondents. 4 female experimenter. haviors investigated in the in the you are to teach in front to three (a) How be to toward? (b) How or guilt enact this enact this likely is enact this to one peared in one subtracted from the male subtracted from and Analysis Effect Sizes effect size means of men and performed independently by each effect size the population effect sizes may be effect sizes to publication preted with caution (Hedges, 1982a, 1978, 1984). in aggression shall In this who participated added to Sample with Variables sizes 1974.67 1974.38 84.50 90.20 58.50 58.34 aggregated 10.00 7.75 0.09 0.09 (0.04/0.15) (0.05/0.13) (0.28/0.40) (0.28/0.38) (0.17/0.24) (0.21/0.29) 0.31 0.31 (0.26/0.37) (0.27/0.35) (0.77/0.90) (0.75/0.86) f 13/37 aggression g 20/30 42/35 aggress i n = n = 77 sex difference were excluded, b Values c Values aggressing), a Values are numbers are greater-than-minimal/ naire respondents' likelihood that than male or by additional to choose included only female sex-of-subject differences. in Table one to = 50) size 0.43 (d+) a All reports M d Significant differences (.34) .001 represent the naire respondents' by the value indicating sex difference). that they in Table differences. Effect sizes a Effect sizes Frequencies are (one-tailed) is p = .05, p = .025 significant difference ALICE H. VALERIE J. STEFFEN effect size 1985), a procedure that effect size W = p effect size the nonsignificant 0.00 (indicating this step in the in the male in the male effect sizes. can be questioned in has an k - I (Hedges, 1982a; Q = p only for the nonsignificant effect size could be the study effect sizes are ordered the end effect sizes for the a signifi- an approximate p - 1 distribution with m - 1 effect sizes in the effect size for each the reciprocal & Becker, in pub- x2(2) = p x2(2) = p required rather For each effect sizes in the male models for the were also models are with each effect size each predictor in the this test k - p - 1 effect sizes 6 Alternative calculations representing each study only one effect size yielded results similar those we report. for the variables, which relate significantly the sex-of-subject effect size: date publication, source publication, percentage male authors, first author, GENDER AND AGGRESSIVE BEHAVIOR 317 ALICE H. GENDER AND AGGRESSIVE BEHAVIOR 319 Sex-of-Subject Effect Weighted effect class effect (QB) n size class (Qw,) ~ 37 0.35 aggression 19.27"** Free choice male direction a Significance sex differences: male) respondents estimated they or anxiety also larger Sex-of-Subject Effect Variable b b* b b* b b* sex difference sex difference aggression e aggress r 0.40** .18 0.68*** .38 0.92*** .47 0.40*** .27 1.08"** .38 weighted least squares size. b = coefficient, b* = female direction, n = 50. a b = 0.0037, = .0019. b Values female estimates aggressing). r Values a 0 = field, 1 = e 0 = psychological, 1 = physical, 1 = aggression AGGRESSIVE BEHAVIOR categorical variables were significant various multivariate questionnaire respondents' study attributes, dition, the p p p effect sizes, be regarded p categorical variables from dan- the inclusion with this when subjects that male proved moder- variability in effect sizes, although it p not added Effect size analyses effect sizes = 20) M effect size CI for M d 0.06/0.57 effect size 0.29 M weighted effect size (d+) a 0.13 = 26) subjects 2,349 Counting methods Frequencies Exact p b 20/23 (.87) Significant differences the male direction c 10/26 (.38) a value 0.00 (exactly ence) was assigned sex differences nonsignificant. Effect sizes were calculated for all nificant differences. Effect sizes a Effect sizes were weighted p (one-tailed) was based on p = c Frequencies are the significant differ- male direction divided parentheses. Exact p (one-tailed) was tion with p = (Robertson, 1960). female direction. Means smaller 0.00 values, W = p ) = 7 Also, effect sizes effect sizes. Given the sion models, we Effect Sizes Between-class Weighted effect class effect n size (lower/upper) class 9 0.37 0.29/0.45 9.92 Physical 7 0.40 0.24/0.56 8.69 Surveillance b 0.55 0.43/0.67 21.66" 7 0.37 0.20/0.53 9.12 0.37 0.25/0.50 59.61"* Effect sizes in the = confidence a Significance homogeneity, b private category effect size. .001. supported by the results the two sex-of-target difference significant sex-of-target Q = p provocation. Although were also article. Because ables on further clouded precluded, test- could be calculated. As did not nor did presented for following variables, which sex-of-target effect sizes: publication, source sex-of-target effect sizes were significantly different in male-au- versus female-authored articles, effect sizes male-authored articles = 5) us hesitant present this AGGRESSIVE BEHAVIOR be regarded upper bound the aggregated in the available sex-of-subject than that sex differences behaviors such which there clear-cut evidence Eagly, 1986). as in research on in the in the affect publishability. sizes suggests The most external validity sex differences in research studied in sex differences aggression. Also to be in child in psy- the typical to confront to avoid these di- require further the absence narrow band in the dence in the target person (Baron sex differences J. STEFFEN along with As in the sex differences in the in semi- target) versus freely chosen. are reluctant the cat- differences in beliefs about in the the variability would cause and Crowley's perceived danger sex differences the guilt/ interactions involving act has strongly negative this regression significantly larger As we subjects aggressed same four to be tory than semiprivate rather freely chosen, sex differences been contested the existence psychological analysis for the AGGRESSIVE BEHAVIOR are for the most part alternative theories, research methods, effect size One may behavior and female gender by women they encour- in an situation provides overall success success was than that the questionnaire respondents' probably unreason- for in a less for the review (Eagly in the social psychologists for studying Validity considerations). published meta-analyses Comparisons with Frodi al.'s conclusions. up to their focus in the sex differ- often quite in the within the in research. popular methods ated with sizes: a ALICE H. VALERIE J. STEFFEN semiprivate context, gender role their preference researchers, especially they are male authors should aggression experi- other laboratory-experimental should be out. Practitioners experiments involving conditions, features this meta-analysis. in the context relatively deficient are somewhat in the example, analyzing own and reflects scientific and ob- note that in the challenging, especially additional predictors & Dobrofsky, L. R. (1978). Military socialization Social Issues, 34(1), Bakan, D. (1966). on psychol- Baron, R, A. (1977). New York: Plenum Baron, R. A., & R. J. (1972). Performance the "aggression machine": Motivation Psychonomic Science, L. (1964). ity catharsis. L., & Donnerstein, I. (1982). External validity skin deep: Some answers laboratory experi- American Psychologist, (1983). Methodology the study R. G. Geen & I. Donnerstein Aggression: Theoretical pp. 213-245). York: Academic R. (1984). A critique and its ries on women. Bloom, L. Z., Coburn, K., & Pearlman, J. (1975). The new New York: Delacourt. Lexington, MA: Lexington Books. & Gordon, L. (1983). new scholarship family vio- Signs: Journal Broverman, I. K., R., Broverman, D. M., Clarkson, (1972). Sex-role stereotypes: A Social Issues, 28(2), 59-78. 1961 ). A. H. (1963). Physical aggression different frustra- Carlson, R. (1972). Understanding women: Implications Social Issues, 28(2), 17-32. (1974). Facilitation physical aggression (1975). Physical aggression as a function attitudes toward punishment. Cicone, M. & Ruble, D. N. (1978). 34(1), 5-16. Cohen, J., & Cohen, H. ( behavioral sciences (2nd ed.). Hillsdale, N Doob, A. N., & Gross, A. E. (1968). Status frustrator as horn-honking responses. Social Psychology, (1983). Gender social influence: A social psychological American Psychologist, A. H. (1986). Some meta-analytic approaches Hyde & M. gender: Advances through meta-analysis (pp. 159-177). Baltimore, MD: Hopkins University social behavior: A social-role (1981). Sex sex differences Psychological Bulle- & Crowley, social psychological ical Bulletin, (1983). Sex differences Psychological Bulletin, E (1974). New York: T. (Ed.). (1982). actions: Expectancy-value models in psychology. J. (1975). How assertiveness training can change New York: N. D. (1982). Sex differences development ofprosocial be- S. (1955). I. (1975). Belief, attitude, intention, Macauley, J., & always less A review logical Bulletin, & AveriU, J. (1982). Differences on emotion York: Springer- & E. Perspectives on aggression (pp. 1-9). York: Aca- E. (1964). psychological measurement. McGaw, B., & L. (1981). CA: Sage. Psychological Bulletin, 82, J. A. differences: Communication accuracy expressive style. gender: Advances through meta-analysis (pp. 136-158). B. (Eds.). (1981). Biobehavioral as- York: Liss. B. (1974). a field Social Psychology, Laboratory. (1955). Computation Laboratory Harvard University: Vol. 35. Tables cumulative binomial probability distribution. ( 1981). Educational Statistics, Hedges, L. effect sizes Educational Statistics, effect size Educational Statistics, & Becker, B. J. (1986). J. S. gender: Advances through recta-analysis Hedges, L. Statistical methods M. L. (1977). Sex differences related behav- Psychological Bulletin, 84, C. M. aggression? A Developmental Psychology, 20, M. C. Advances through recta- Keller, E. Reflections on gender A. (1970). Sex-role Psychological Reports, Krebs, D. L., & Miller, & E. social psychology ed., Vol. 2, (1977). Sex differences effect size: British Jour- Leventhal, D. B., & Shemberg, (1969). Sex & Pillemer, B. (1984). (1979). A social psychological social psychology. Buss (Ed.), social context Macaulay, J. (1985). aggression research: Unger, & B. social psychology Hillsdale, N J: Maccoby, E. E., & C. N. Maccoby, E. E., & (1980). Sex differences C. (1961). (1983). A psychological research. Berkowitz (Ed.), Advances in experimental social psychology pp. 1-47). R. D., & Slaby, child psychology York: Wiley. aggressive adolescents. L. A. (1983). Roles Harvey, T. VALERIE J. L. A. D. R. Close relationships New York: Rajecki, D. E. I. (Eds.),Aggression: Theo- empirical reviews 189-211). New York: Aca- S., & Taylor, S. P. retaliative behavior. binomial distribution No. SCR- Washington, DC: (1976). Sex differences aggression: Phylogenetic chological Bullet 85, 185-193. Meta-analytic procedures social research. CA: Sage. L. (1981). cial inquiry New York: P., & Worchel, (1964). Ego L. (1983). Sex stereotypes: Issues & Pisor, S. S., & B. (1965). test for K. M., B., & L. (1968). Aggression E. C., E., & Waiters, (Eds.). (1983). neural approaches. Hillsdale, N Public Opinion Quarterly Spence, J. T., & L. (1978). Masculinity & psychological dimensions, correlates, & antecedents. Texas Press. Issues, 34( 1 ), can Behavioral Scientist, 23, doors: Violence in City, NY: T. (1983). Social influence theory & E. Aggression: Theoretical 135-162). New York: Tedeschi, J. T., C., Jr. (1974). A Psychological Bulletin, 81, W., & Viney, (1969). Expression Perceptual & Skills, 29, Justice. Office Assistance, Research, Statistics. ( 1979 Washington, DC: Statistics. (1980). tives on working women: A databook Wallston, B. are the feminist approach to Women Quarterly, (1983). Sex aggression research. E. I. Aggression: Theoretical empirical reviews (Vol. 2, Jr. (1956). organization man. New York: Social Behavior & Polivy, (1976). Effects Psychology, 85, & Ball, R. L. Abnormal and Psychology, 60, S. (1959). & Rawlings, (1963). Effects violence on Psychology, 66, Bond, M. aggressive behavior. in relation Abnormal and Research in Personality, Dissertation Abstracts Interna- 890) (25) L. J., N. H. aggression: A Psychology, 84, K. K. Social Psychology, 218. (20) (1968). Aggressive Dissertation Abstracts International, 922) (10, 35) S. (1955). Psychology, 50, G., Kelm, H., J. L., B. (1976). in women Research in S., & M. A. 977). Verbal aggression control and Dissertation Abstracts International, B. (1973). - 139. (12) B. (1974a). Aggressive B. (1974b). B. (1976). B., & (1974). Aggression B., & Klingbeil, D. confederate on (1973). Aggression, B., & Social Psy- Abstracts International, T. (1982). Aggression T., & Esselman, J. A. Psychonomic Society, and winning Psychonomic Society, (23, 37) S. (1974). arousal and & Baer, and the Koerner, E Abstracts International, I. (1978). aggressive behavior. 43, 44) G. K. (1974). Effects hostility arousal Mueller, C., & Donnerstein, E. arousal upon aggressive behavior. Research in D. (1974). Dissertation Abstracts International, 494) (3) (1980). Expressions aggression: Aggression-inhibiting Social Psychology Bulletin, 315-320. (11) & Schlottmann, R. (28, 29) Scheier, M. & Buss, D. M. (1978). Self-consciousness, Research in Z., Schuck, HaUam, E., & Wells, Sex differences Psychological Reports, 931-936. (2) & Dintzer, 399-410. (18) & Epstein, (1967). Aggression & Viney, (1969). Expression Perceptual & Turner, C. Layton, J. & Simons, aggressive behavior: Aggressive stimuli, Social Psychology, Unger, R. J., & Levine, M. (1974). women a group? Sometimes! International Journal Group Ten- M., Jaffe, & Feshbach, I. (1968). race, sex, hostility, verbal stimulus Psychonomic Science, Revision received 16, 1986, Select Press 16, 1986, could order existing systems, publisher could to be Select Press Select Press also Further information Select Press GENDER AND AGGRESSIVE BEHAVIOR 319 GENDER AND AGGRESSIVE BEHAVIOR 319 GENDER AND AGGRESSIVE BEHAVIOR 319 GENDER AND AGGRESSIVE BEHAVIOR 319 GENDER AND AGGRESSIVE BEHAVIOR 319 GENDER AND AGGRESSIVE BEHAVIOR 319 GENDER AND AGGRESSIVE BEHAVIOR 319 GENDER AND AGGRESSIVE BEHAVIOR 319 GENDER AND AGGRESSIVE BEHAVIOR 319 GENDER AND AGGRESSIVE BEHAVIOR 319 GENDER AND AGGRESSIVE BEHAVIOR 319 ALICE H. J. STEFFEN (e.g., Eagly, 1983; & Crowley, 1986). underlie sex aggressive behavior. a single, ing sex that may Maccoby and Frodi et ences in The malegender role. aggression. Psychologists about men men to be aggressive (see 1978). Similarly, stereotype research women. Survey this role aggressiveness (see & Crowley, 1986). that men and support traditional female this role & Crowley, 1986) avoiding physical intended in less aggressive assertiveness, a has often & Baer, 1975). gender role, may be posed to only within may transmit gressive behavior. AGGRESSIVE BEHAVIOR women to implications for an element women in behavior in in the (Eagly, 1983). higher status status enjoins Overall sex-of-subject difference. all suggest direction. Indeed, ways social contextual variables differences in in the hypothesis stems the prediction emphasize primarily (and therefore jects faced whether the & Taylor, 1979). predict the in the for each Buss, 1963) for each multiple observa- distinction between physical psychological aggression has one variant aggression literature. Some investigators (e.g., Buss, 1961) have distinguished between these two overtly physical versus vocal nature aggressor's act. Following Frodi, Macauley, and Thome (1977), we the target. Yet Frodi collaborators labeled aggression causing psychological and, unlike it encompasses behaviors as- the studies we the research experimental situations typically allowed either psychological cal aggression--not both. Sample with Variables sizes 1974.67 1974.38 84.50 90.20 58.50 58.34 aggregated 10.00 7.75 0.09 0.09 (0.04/0.15) (0.05/0.13) (0.28/0.40) (0.28/0.38) (0.17/0.24) (0.21/0.29) 0.31 0.31 (0.26/0.37) (0.27/0.35) (0.77/0.90) (0.75/0.86) f 13/37 aggression g 20/30 42/35 aggress i n = n = 77 sex difference were excluded, b Values c Values aggressing), a Values are numbers are greater-than-minimal/ naire respondents' likelihood that than male or by additional to choose included only female sex-of-subject differences. in Table one to = 50) size 0.43 (d+) a All reports M d Significant differences (.34) .001 represent the naire respondents' by the value indicating sex difference). that they in Table differences. Effect sizes a Effect sizes Frequencies are (one-tailed) is p = .05, p = .025 significant difference Sample with Variables sizes 1974.67 1974.38 84.50 90.20 58.50 58.34 aggregated 10.00 7.75 0.09 0.09 (0.04/0.15) (0.05/0.13) (0.28/0.40) (0.28/0.38) (0.17/0.24) (0.21/0.29) 0.31 0.31 (0.26/0.37) (0.27/0.35) (0.77/0.90) (0.75/0.86) f 13/37 aggression g 20/30 42/35 aggress i n = n = 77 sex difference were excluded, b Values c Values aggressing), a Values are numbers are greater-than-minimal/ naire respondents' likelihood that than male or by additional to choose included only female sex-of-subject differences. in Table one to = 50) size 0.43 (d+) a All reports M d Significant differences (.34) .001 represent the naire respondents' by the value indicating sex difference). that they in Table differences. Effect sizes a Effect sizes Frequencies are (one-tailed) is p = .05, p = .025 significant difference Sample with Variables sizes 1974.67 1974.38 84.50 90.20 58.50 58.34 aggregated 10.00 7.75 0.09 0.09 (0.04/0.15) (0.05/0.13) (0.28/0.40) (0.28/0.38) (0.17/0.24) (0.21/0.29) 0.31 0.31 (0.26/0.37) (0.27/0.35) (0.77/0.90) (0.75/0.86) f 13/37 aggression g 20/30 42/35 aggress i n = n = 77 sex difference were excluded, b Values c Values aggressing), a Values are numbers are greater-than-minimal/ naire respondents' likelihood that than male or by additional to choose included only female sex-of-subject differences. in Table one to = 50) size 0.43 (d+) a All reports M d Significant differences (.34) .001 represent the naire respondents' by the value indicating sex difference). that they in Table differences. Effect sizes a Effect sizes Frequencies are (one-tailed) is p = .05, p = .025 significant difference Sample with Variables sizes 1974.67 1974.38 84.50 90.20 58.50 58.34 aggregated 10.00 7.75 0.09 0.09 (0.04/0.15) (0.05/0.13) (0.28/0.40) (0.28/0.38) (0.17/0.24) (0.21/0.29) 0.31 0.31 (0.26/0.37) (0.27/0.35) (0.77/0.90) (0.75/0.86) f 13/37 aggression g 20/30 42/35 aggress i n = n = 77 sex difference were excluded, b Values c Values aggressing), a Values are numbers are greater-than-minimal/ naire respondents' likelihood that than male or by additional to choose included only female sex-of-subject differences. in Table one to = 50) size 0.43 (d+) a All reports M d Significant differences (.34) .001 represent the naire respondents' by the value indicating sex difference). that they in Table differences. Effect sizes a Effect sizes Frequencies are (one-tailed) is p = .05, p = .025 significant difference ALICE H. VALERIE J. STEFFEN effect size 1985), a procedure that effect size W = p effect size the nonsignificant 0.00 (indicating this step in the in the male in the male effect sizes. can be questioned in has an k - I (Hedges, 1982a; Q = p only for the nonsignificant effect size could be the study effect sizes are ordered the end effect sizes for the a signifi- an approximate p - 1 distribution with m - 1 effect sizes in the effect size for each the reciprocal & Becker, in pub- x2(2) = p x2(2) = p required rather For each effect sizes in the male models for the were also models are with each effect size each predictor in the this test k - p - 1 effect sizes 6 Alternative calculations representing each study only one effect size yielded results similar those we report. for the variables, which relate significantly the sex-of-subject effect size: date publication, source publication, percentage male authors, first author, ALICE H. VALERIE J. STEFFEN effect size 1985), a procedure that effect size W = p effect size the nonsignificant 0.00 (indicating this step in the in the male in the male effect sizes. can be questioned in has an k - I (Hedges, 1982a; Q = p only for the nonsignificant effect size could be the study effect sizes are ordered the end effect sizes for the a signifi- an approximate p - 1 distribution with m - 1 effect sizes in the effect size for each the reciprocal & Becker, in pub- x2(2) = p x2(2) = p required rather For each effect sizes in the male models for the were also models are with each effect size each predictor in the this test k - p - 1 effect sizes 6 Alternative calculations representing each study only one effect size yielded results similar those we report. for the variables, which relate significantly the sex-of-subject effect size: date publication, source publication, percentage male authors, first author, GENDER AND AGGRESSIVE BEHAVIOR 317 ALICE H. AGGRESSIVE BEHAVIOR categorical variables were significant various multivariate questionnaire respondents' study attributes, dition, the p p p effect sizes, be regarded p categorical variables from dan- the inclusion with this when subjects that male proved moder- variability in effect sizes, although it p not added Effect size analyses effect sizes = 20) M effect size CI for M d 0.06/0.57 effect size 0.29 M weighted effect size (d+) a 0.13 = 26) subjects 2,349 Counting methods Frequencies Exact p b 20/23 (.87) Significant differences the male direction c 10/26 (.38) a value 0.00 (exactly ence) was assigned sex differences nonsignificant. Effect sizes were calculated for all nificant differences. Effect sizes a Effect sizes were weighted p (one-tailed) was based on p = c Frequencies are the significant differ- male direction divided parentheses. Exact p (one-tailed) was tion with p = (Robertson, 1960). female direction. Means smaller 0.00 values, W = p ) = 7 Also, effect sizes effect sizes. Given the sion models, we AGGRESSIVE BEHAVIOR be regarded upper bound the aggregated in the available sex-of-subject than that sex differences behaviors such which there clear-cut evidence Eagly, 1986). as in research on in the in the affect publishability. sizes suggests The most external validity sex differences in research studied in sex differences aggression. Also to be in child in psy- the typical to confront to avoid these di- require further the absence narrow band in the dence in the target person (Baron sex differences AGGRESSIVE BEHAVIOR are for the most part alternative theories, research methods, effect size One may behavior and female gender by women they encour- in an situation provides overall success success was than that the questionnaire respondents' probably unreason- for in a less for the review (Eagly in the social psychologists for studying Validity considerations). published meta-analyses Comparisons with Frodi al.'s conclusions. up to their focus in the sex differ- often quite in the within the in research. popular methods ated with sizes: a AGGRESSIVE BEHAVIOR are for the most part alternative theories, research methods, effect size One may behavior and female gender by women they encour- in an situation provides overall success success was than that the questionnaire respondents' probably unreason- for in a less for the review (Eagly in the social psychologists for studying Validity considerations). published meta-analyses Comparisons with Frodi al.'s conclusions. up to their focus in the sex differ- often quite in the within the in research. popular methods ated with sizes: a AGGRESSIVE BEHAVIOR are for the most part alternative theories, research methods, effect size One may behavior and female gender by women they encour- in an situation provides overall success success was than that the questionnaire respondents' probably unreason- for in a less for the review (Eagly in the social psychologists for studying Validity considerations). published meta-analyses Comparisons with Frodi al.'s conclusions. up to their focus in the sex differ- often quite in the within the in research. popular methods ated with sizes: a AGGRESSIVE BEHAVIOR are for the most part alternative theories, research methods, effect size One may behavior and female gender by women they encour- in an situation provides overall success success was than that the questionnaire respondents' probably unreason- for in a less for the review (Eagly in the social psychologists for studying Validity considerations). published meta-analyses Comparisons with Frodi al.'s conclusions. up to their focus in the sex differ- often quite in the within the in research. popular methods ated with sizes: a AGGRESSIVE BEHAVIOR are for the most part alternative theories, research methods, effect size One may behavior and female gender by women they encour- in an situation provides overall success success was than that the questionnaire respondents' probably unreason- for in a less for the review (Eagly in the social psychologists for studying Validity considerations). published meta-analyses Comparisons with Frodi al.'s conclusions. up to their focus in the sex differ- often quite in the within the in research. popular methods ated with sizes: a Social Behavior & Polivy, (1976). Effects Psychology, 85, & Ball, R. L. Abnormal and Psychology, 60, S. (1959). & Rawlings, (1963). Effects violence on Psychology, 66, Bond, M. aggressive behavior. in relation Abnormal and Research in Personality, Dissertation Abstracts Interna- 890) (25) L. J., N. H. aggression: A Psychology, 84, K. K. Social Psychology, 218. (20) (1968). Aggressive Dissertation Abstracts International, 922) (10, 35) S. (1955). Psychology, 50, G., Kelm, H., J. L., B. (1976). in women Research in S., & M. A. 977). Verbal aggression control and Dissertation Abstracts International, B. (1973). - 139. (12) B. (1974a). Aggressive B. (1974b). B. (1976). B., & (1974). Aggression B., & Klingbeil, D. confederate on (1973). Aggression, B., & Social Psy- Abstracts International, T. (1982). Aggression T., & Esselman, J. A. Psychonomic Society, and winning Psychonomic Society, (23, 37) S. (1974). arousal and & Baer, and the Koerner, E Abstracts International, I. (1978). aggressive behavior. 43, 44) Social Behavior & Polivy, (1976). Effects Psychology, 85, & Ball, R. L. Abnormal and Psychology, 60, S. (1959). & Rawlings, (1963). Effects violence on Psychology, 66, Bond, M. aggressive behavior. in relation Abnormal and Research in Personality, Dissertation Abstracts Interna- 890) (25) L. J., N. H. aggression: A Psychology, 84, K. K. Social Psychology, 218. (20) (1968). Aggressive Dissertation Abstracts International, 922) (10, 35) S. (1955). Psychology, 50, G., Kelm, H., J. L., B. (1976). in women Research in S., & M. A. 977). Verbal aggression control and Dissertation Abstracts International, B. (1973). - 139. (12) B. (1974a). Aggressive B. (1974b). B. (1976). B., & (1974). Aggression B., & Klingbeil, D. confederate on (1973). Aggression, B., & Social Psy- Abstracts International, T. (1982). Aggression T., & Esselman, J. A. Psychonomic Society, and winning Psychonomic Society, (23, 37) S. (1974). arousal and & Baer, and the Koerner, E Abstracts International, I. (1978). aggressive behavior. 43, 44) G. K. (1974). Effects hostility arousal Mueller, C., & Donnerstein, E. arousal upon aggressive behavior. Research in D. (1974). Dissertation Abstracts International, 494) (3) (1980). Expressions aggression: Aggression-inhibiting Social Psychology Bulletin, 315-320. (11) & Schlottmann, R. (28, 29) Scheier, M. & Buss, D. M. (1978). Self-consciousness, Research in Z., Schuck, HaUam, E., & Wells, Sex differences Psychological Reports, 931-936. (2) & Dintzer, 399-410. (18) & Epstein, (1967). Aggression & Viney, (1969). Expression Perceptual & Turner, C. Layton, J. & Simons, aggressive behavior: Aggressive stimuli, Social Psychology, Unger, R. J., & Levine, M. (1974). women a group? Sometimes! International Journal Group Ten- M., Jaffe, & Feshbach, I. (1968). race, sex, hostility, verbal stimulus Psychonomic Science, Revision received 16, 1986, Select Press 16, 1986, could order existing systems, publisher could to be Select Press Select Press also Further information Select Press G. K. (1974). Effects hostility arousal Mueller, C., & Donnerstein, E. arousal upon aggressive behavior. Research in D. (1974). Dissertation Abstracts International, 494) (3) (1980). Expressions aggression: Aggression-inhibiting Social Psychology Bulletin, 315-320. (11) & Schlottmann, R. (28, 29) Scheier, M. & Buss, D. M. (1978). Self-consciousness, Research in Z., Schuck, HaUam, E., & Wells, Sex differences Psychological Reports, 931-936. (2) & Dintzer, 399-410. (18) & Epstein, (1967). Aggression & Viney, (1969). Expression Perceptual & Turner, C. Layton, J. & Simons, aggressive behavior: Aggressive stimuli, Social Psychology, Unger, R. J., & Levine, M. (1974). women a group? Sometimes! International Journal Group Ten- M., Jaffe, & Feshbach, I. (1968). race, sex, hostility, verbal stimulus Psychonomic Science, Revision received 16, 1986, Select Press 16, 1986, could order existing systems, publisher could to be Select Press Select Press also Further information Select Press 6 Alternative calculations representing each study by only one effect size yielded results similar to those we report. No results are presented for the following variables, which did not relate significantly to the mag- nitude of the sex-of-subject effect size: date of publication, source of publication, percentage of male authors, sex of first author, amount of provocation, and sex of target of aggression. model in the sense that the model sufficiently accounted for the systematic variation in the effect sizes), it would yield a signifi- cant between-class effect and homogeneous effect sizes within each class. The between-class effect is estimated by QB, which has an approximate chi-square distribution with p - 1 degrees of freedom, where p is the number of classes. The homogeneity of the effect sizes within each class is estimated by has an approximate chi-square distribution with m - 1 degrees of freedom, where m is the number of effect sizes in the class. Table 4 also includes (a) the mean effect size for each class, cal- culated with each effect size weighted by the reciprocal of its variance, and (b) the 95% confidence interval for each mean. Significant between-class effects showed that the tendency for men to aggress more than women was greater in laboratory than field settings and for physical versus psychological aggression. Consistent with the significant between-class surveillance effect, post-hoc comparisons among the mean effect sizes for the three classes (Hedges & Becker, 1986; Hedges & Olkin, 1985) showed that the sex difference in semiprivate contexts (target and/or experimenter present) was larger than the sex difference in pub- lic contexts (additional onlookers present), x2(2) = 9.85, p .01, and marginally larger than the sex difference in private con- texts, x2(2) = 4.86, p .10. Also, the tendency for men to ag- gress more than women was significantly larger when aggression was required rather than freely chosen. Despite these significant between-class effects, none of these categorical models can be regarded as having fit the effect sizes. For each model, the hypothesis of homogeneity of the effect sizes was rejected within each class, except for one category containing only six effect sizes (the private category of the sur- veillance variable, see Table 4). With the exception of this same category, all of the category means differed from 0.00 and thus indicated a significant sex difference in the male direction. No category mean was positive, which would indicate a difference in the female direction. of continuous models. and multivariate tests of continuous models for the sex-of-subject differences were also conducted (Hedges, 1982b; Hedges & Olkin, 1985). These models are least squares regressions, calculated with each effect size weighted by the reciprocal of its variance. Each such model yields a test of the significance of each predictor as well as a test of model specification, which evaluates whether sig- nificant systematic variation remains unexplained in the re- gression model. The error sum of squares statistic, QE, which provides this test of model specification, has an approximate chi-square distribution with k - p - 1 degrees of freedom, where k is the number of effect sizes and p is the number of predictors (not including the intercept). As Table 5 shows, univariate tests indicated that six of the continuous variables were significantly related to the sex-of- subject differences. The first of these variables, the number of behaviors aggregated in the aggression measure, was related from the 0.00 value that indicates exactly no difference suggests an overall sex difference. The mean of the known effect sizes differed from 0.00 in the direction of more aggression by men than women. Weighting each known effect size by the reciprocal of its variance (Hedges & Olkin, 1985), a procedure that gives more weight to effect sizes that are more reliably estimated, yielded a significant mean effect size in the male direction that was smaller than the unweighted mean. The distribution of the known effect sizes was normal, W = .97, with values lower than .95 indicating rejection of the hypothesis of normality at p .05 (Shapiro & Wilk, 1965). There is no completely satisfactory method to compute a mean effect size that takes into account the nonsignificant effects that could not be calculated because of a lack of suffi- cient information. Nevertheless, one possible solution is to give these nonsignificant effects the value of 0.00 (indicating exactly no sex difference). When this step was taken, the mean (un- weighted) effect size decreased but remained significant in the male direction. This mean is reported in Table 2, under "All reports" As Table 2 shows, the conclusion that men aggressed more than women was supported by counting test results (Rosenthal, 1978) demonstrating that .89, the proportion of reports indicat- ing a sex difference in the male direction (disregarding signifi- cance) departed significantly from. 50, the proportion expected under the null hypothesis. As Table 2 also shows, greater aggres- sion by men than women was also consistent with a second counting test, which demonstrated that .34, the proportion of reports indicating a significant sex difference in the male direc- tion, departed significantly from .025, the proportion expected under the null hypothesis. of effect sizes. the aggregated sex differences in Table 2 are of interest in relation to our predic- tions, their importance can be questioned in view of the incon- sistency of the findings across the studies. Calculation of a ho- mogeneity statistic Q, which has an approximate chi-square distribution with k - I degrees of freedom, where k is the num- ber of effect sizes (Hedges, 1982a; Hedges & Olkin, 1985), indi- cated that the hypothesis that the known effect sizes were homo- geneous was rejected, Q = 202.44, p .001. Therefore, study attributes were used to account for variability in the sex differ- ences. Prediction was attempted only for the 50 known effect sizes because the 0.00 values used to estimate the nonsignificant effects that could not be calculated are too inexact to warrant an attempt to fit statistical models. 6 Table 3 presents each sex-of-subject effect size that could be calculated, along with the study attributes that predicted these effect sizes and a brief description of each aggressive behavior. Effect sizes are ordered by their magnitude and direction so that the largest sex differences in the male direction appear at the beginning of the table and the largest differences in the female direction appear at the end of the table. of categorical models. 4 presents tests of the univariate categorical models that yielded significant between- class effects (analogous to main effects in an analysis of vari- ance) for sex-of-subject differences. In addition to a test of the significance of between-class effects, this approach provides a test of the homogeneity of the effect sizes within each class. If a categorical model were correctly specified (i.e., the data fit the ALICE H. EAGLY AND VALERIE J. STEFFEN ALICE H. VALERIE J. STEFFEN 1985), a procedure that effect size W = p effect size the nonsignificant this step in the in the male in the male effect sizes. has an k - I (Hedges, 1982a; was rejected, Q = p only for the nonsignificant the study effect sizes are ordered at the end for the a signifi- p - 1 distribution with m - 1 effect sizes in the for each the reciprocal & Becker, in pub- x2(2) = p x2(2) = p required rather effect sizes in the male which would models for the were also models are with each the reciprocal each predictor in the this test k - p - 1 effect sizes 6 Alternative calculations representing each study only one effect size yielded results similar those we report. for the variables, which relate significantly the sex-of-subject effect size: date publication, source publication, percentage male authors, sex first author,