##### **Please cite as:**
Reimer, N. K., Becker, J. C., Benz, A., Christ, O., Dhont, K., Klocke, U., ... Hewstone, M. (in press). Intergroup contact and social change: Implications of negative and positive contact for collective action in advantaged and disadvantaged groups. *Personality and Social Psychology Bulletin*.
##### **Figures**
*Figure 1.* Structural equation model showing direct and indirect paths from negative and positive contact to collective action for sexual-minority participants (Study 1a, *n* = 233). Negative and positive contact (*r* = -.32 [-.52, -.12]) were correlated. Standardized coefficients are reported; only significant paths are shown.
*Figure 2.* Alternative model including direct and indirect effects of positive but not negative contact for sexual-minority students (Study 1a, *n* = 233). Previously significant paths are shown as dashed lines. Standardized coefficients are reported. Model fit: χ² (154) = 194.17, χ²/df = 1.26, CFI = .98, TFI = .98, RMSEA = .03.
*Figure 3.* Structural equation model showing direct and indirect paths from negative and positive contact to collective action for heterosexual participants (Study 1b, *n* = 249). Negative and positive contact (*r* = .02 [ .16, .20]) were not significantly correlated. Standardized coefficients are reported; only significant paths are shown.
*Figure 4.* Panel model showing autoregressive (in grey) and cross-lagged (in black) paths for sexual-minority participants (Study 2a, *n* = 433). Negative and positive contact (*r* = -.19 [-.32, -.05]) were correlated at T1. Standardized coefficients are reported; only significant paths are shown.
*Figure 5.* Panel model showing autoregressive (in grey) and cross-lagged (in black) paths for heterosexual/cisgender participants (Study 2b, *n* = 1,036). Standardized coefficients are reported; only significant paths are shown.
##### **Tables**
*Table 1.* Effect sizes with 95% confidence intervals for indirect, direct, and total effects of intergroup contact on collective action among sexual-minority participants (Study 1a). Created with *ggplot2* (Wickham, 2009).
*Table 2.* Effect sizes with 95% confidence intervals for indirect (half-longitudinal), direct (cross-lagged), and total effects of intergroup contact on collective action among sexual-minority participants (Study 2a). Created with *ggplot2* (Wickham, 2009).
##### **Online Appendix**
*OA 2.2.* Structural equation model showing all paths from negative and positive contact to collective action for sexual-minority participants (Study 2a, *n* = 421) at T1. Paths that were significant in Study 1a but not 2a are shown as dashed lines; paths that were significant in Study 2a but not 1a are shown as grey lines. Negative and positive contact (*r* = .18 [ .31, .04]) were correlated. Model fit: χ² (182) = 410.02, χ²/df = 2.25, CFI = .94, TFI = .93, RMSEA = .06
*OA 2.3.* Panel model showing autoregressive (in grey) and cross-lagged (in black) paths for sexual-minority participants (Study 2a, *n* = 435). Standardized coefficients are reported; only significant paths are shown.
*OA 2.4.* Panels model showing autoregressive (in grey) and cross-lagged (in black) paths for sexual-minority participants (Study 2a, *n* = 433). *Solidarity* and *Centrality* measured participants’ identification with (a) people of the same sexual orientation (e.g., bisexual people for a bisexual participant), or (b) the LGBT+ movement. Standardized coefficients are reported; only significant paths are shown.
