Table S6.

Summary of mixed models analyzed using multilevel modeling showing effects of mentor condition on all dependent variables in year 2 (reference group = female mentor)

Predictor	Belonging	Threat	Intentions to pursue advanced engineering degrees
β00 Reference intercept	5.24 (0.23)***	4.86 (0.18)***	4.65 (0.33)***
β01 Male mentor	−0.5 (0.32)†	0.11 (0.26)	−0.27 (0.47)
β02 No mentor/control	−0.2 (0.32)	0.35 (0.26)†	−0.59 (0.46)
β10 Reference slope	0.005 (0.01)	0.02 (0.01)*	−0.02 (0.01)
β11 Male mentor	−0.03 (0.01)*	0.03 (0.01)*	−0.02 (0.02)
β12 No mentor/control	−0.02 (0.01)	0.007 (0.01)	−0.07 (0.02)*

Note: ^†P < 0.20, ^P < 0.10, *P < 0.05, **P < 0.01, and ***P < 0.001. SEs are in parentheses. β₀₀ is the intercept for the female-mentor group at time 4 (19 mo). β₀₁ and β₀₂ represent the relative effects of having a male mentor and no mentor, respectively. The absolute intercepts for each groups can be derived by computing the difference from the female-mentor group (β₀₀). Negative values for β₀₁ and β₀₂ indicate that women in the male-mentor condition or no-mentor condition have a lower mean at time 4 than women with female mentors; positive values indicate a higher mean at time 4. β₁₀ represents change in the dependent variable over time for women with female mentors. β₁₁ and β₁₂ are the relative differences in change over time for women with male mentors and no mentors, respectively. The absolute change over time (slope) for each group can be derived by computing the difference from the female-mentor group (β₁₀). Negative values for β₁₀, β₁₁, and β₁₂ indicate decreases in the dependent variable over time; positive values indicate increases in the dependent variable over time. Significance tests for the reference indicate whether the intercept and slope for the female mentor group (β₀₀ and β₁₀) differ from zero. Significance tests for the male- and no-mentor conditions indicate differences from the female-mentor condition.