The Gender Similarities Hypothesis Notes

The Gender Similarities Hypothesis

Introduction

• Janet Shibley Hyde proposed the gender similarities hypothesis, contrasting the popular "differences model".
• The differences model posits large psychological differences between males and females, while the gender similarities hypothesis suggests males and females are similar on most psychological variables.
• A review of 46 meta-analyses supports the gender similarities hypothesis.
• Gender differences vary in magnitude across ages and contexts.
• Exaggerated claims of gender differences have negative consequences in workplaces and relationships.

The Hypothesis

• The gender similarities hypothesis: males and females are similar on most, but not all, psychological variables.
• Effect sizes:
  • Close-to-zero: d \le 0.10
  • Small: 0.11 \le d \le 0.35
  • Moderate: 0.36 \le d \le 0.65
  • Large: 0.66 \le d \le 1.00
  • Very Large: d \ge 1.00
• Early researchers like Thorndike (1914) and Hollingworth (1918) noted gender similarities.
• Woolley (1914) observed a gap between data and scientists’ views on gender differences.

Meta-Analysis in Assessing Gender Differences

• Early reviews: Woolley (1914) and Hollingworth (1918).

• Maccoby and Jacklin's (1974) review of over 2,000 studies challenged beliefs about gender differences in areas like social behavior, suggestibility, self-esteem, and achievement motivation.

• Maccoby and Jacklin found established gender differences in verbal, visual-spatial, mathematical abilities, and aggression.

• Secondary reports disproportionately focused on gender differences.

• Meta-analysis: a statistical method for aggregating research findings across many studies.

• Effect size (d): measures the magnitude of gender difference.

  d = \frac{MM - MF}{s_w}, where:

  • M_M = mean score for males
  • M_F = mean score for females
  • s_w = average within-sex standard deviation

• Negative d: females scored higher; positive d: males scored higher.

• Steps in gender meta-analyses:

  • Locate all relevant studies.
  • Extract statistics and compute effect sizes.
  • Compute a weighted average of effect sizes.
  • Conduct homogeneity analyses.

Evidence for Gender Similarities

• Meta-analyses of psychological gender differences were collected and grouped into categories: cognitive, communication, social/personality, well-being, motor behaviors, and miscellaneous constructs.
• Analyses were updated with recent studies and larger samples when possible.
• Data from large probability samples (e.g., National Longitudinal Study of Youth) were included.
• Table 1 shows effect sizes, with asterisks indicating data from major national samples.
• Table 2 summarizes effect sizes: 30% are close-to-zero, and 48% are small (78% total).
• These results support the gender similarities hypothesis, even in traditionally believed areas of gender difference.
• Gender differences in communication and moral reasoning are generally small.

Exceptions to the Rule

• The gender similarities hypothesis acknowledges exceptions where gender differences are moderate or large.
• Motor performance: throwing velocity (d = 2.18) and distance (d = 1.98).
• Sexuality: masturbation and attitudes about casual sex show striking gender differences; sexual satisfaction differences are close to zero.
• Aggression: moderate gender differences; physical aggression differences are larger than verbal aggression.
• Relational aggression: ambiguous evidence; effect size varies by measurement method.

Interpretation of Effect Sizes

• Cohen (1969, 1988): small effect = 0.20, medium = 0.50, large = 0.80.
• Kling et al. (1999): graphed distributions differing by d = 0.21, showing significant overlap.
• For d = 0.20, U = 15% (85% overlap).
• For d = 0.20, 54% of individuals in Group A exceed the 50th percentile for Group B.
• d = 0.20 is equivalent to r = 0.10 (small correlation).
• Rosenthal (1991): even small effect sizes can be important.
• Binomial effect size display (BESD): r = 0.32 (10% variance) translates to survival rates of 34% (placebo) vs. 66% (treatment).
• A d of 0.20 is equivalent to an r of 0.10, translating to a small increase in cancer survival (45% to 55%).
• Cohen’s guidelines offer a reasonable standard for interpreting gender differences effect sizes.
• The greater male variability hypothesis suggests that males might have a larger standard deviation than females.
• Variance ratio (VR): the ratio of the male variance to the female variance.
• Empirical VR investigations: vocabulary (1.00–1.08), mathematics performance (1.05–1.25), self-esteem (0.87–1.04).

Developmental Trends

• Meta-analysis can identify age trends in the magnitude of gender differences.

• Hyde, Fennema, & Lamon (1990): girls showed a small advantage in computation in elementary school and middle school and a small difference favoring males in problem-solving during high school

• Kling et al. (1999): gender differences in self-esteem increase during adolescence but decrease in older samples.

  • Childhood (ages 7–10): d = 0.16
  • Early adolescence (ages 11–14): d = 0.23
  • High school years (ages 15–18): d = 0.33
  • Ages 23–59: d = 0.10

• Whitley (1997): computer self-efficacy shows a dramatic trend: grammar school (d = 0.09) vs. high school (d = 0.66).

• These examples show that gender differences fluctuate with age, arguing against the idea of large and stable gender differences.

Importance of Context

• Context influences the creation, erasure, or reversal of psychological gender differences.
• Lightdale and Prentice (1994): deindividuation removes the influence of gender roles, eliminating gender differences in aggression.
  • Individuated condition: men dropped more bombs than women.
  • Deindividuated condition: no significant gender differences. Women dropped somewhat more bombs than men.
• Steele (1997): stereotype threat affects cognitive performance.
• Spencer et al. (1999): gender differences in math performance depend on whether participants are told the test shows gender differences.
  • Gender-fair test: no gender differences.
  • Expected gender differences: women underperformed compared with men.
• Eagly and Crowley (1986): gender differences in helping behavior depend on social context (onlookers, danger).
  • Onlookers present: d = 0.74.
  • No onlookers present: d = -0.02.
• Anderson and Leaper (1998): gender differences in conversational interruption vary depending on the social context.
  • Dyads: d = 0.06.
  • Larger groups: d = 0.26.
  • Strangers: d = 0.17.
  • Friends: d = -0.14.
• LaFrance, Hecht, and Paluck (2003): gender differences in smiling depend on awareness of being observed and culture and age.
  • Aware of being observed: d = -0.46.
  • Not aware of being observed: d = -0.19.
• The conclusion is clear: the magnitude and even the direction of gender differences depends on the context.

Costs of Inflated Claims of Gender Differences

• Overinflated claims of gender differences have consequences in work, parenting, and relationships.
• Gilligan’s (1982) argument about different moral voices reifies gender stereotypes, affecting how men and women are perceived.
• Rudman and Glick (1999): female job applicants displaying agentic qualities receive lower hireability ratings.
• Eagly, Makhijani, and Klonsky (1992): women leaders portrayed as uncaring autocrats are evaluated more negatively.
• Stereotypes about boys and girls in math affect expectations, self-confidence, and performance.
• Inflated claims about communication differences in relationships can hinder conflict resolution.
• Overlooking self-esteem problems in boys due to media focus on girls.

Conclusion

• The gender similarities hypothesis contrasts the differences model.
• Meta-analyses support the gender similarities hypothesis, with exceptions in motor behaviors and sexuality.
• Overinflated claims of gender differences cause harm in the workplace, relationships, and self-esteem analyses.
• These claims are not consistent with the data. Therapists should base their practice on the best scientific evidence. The evidence does not support the belief that men and women have inherent difficulties in communicating across gender. Neither does the evidence support the belief that adolescent girls are the only ones with self-esteem problems.