Table 2.

Multi-group QAP results on log-transformed interaction durations of dyads.

	Estimate	p	E(Est.)	Percentiles
				2.5th	97.5th
Intercept	2.504**	0.005	1.820	1.290	2.346
Sample two	0.806	0.344	0.835	0.693	0.974
At least one female	−0.095	0.129	−0.003	−0.160	0.155
Both female	−0.148*	0.036	0.004	−0.158	0.160
Age mean (centered)	0.065*	0.013	0.000	−0.059	0.057
Age similarity	0.042**	0.009	0.000	−0.035	0.035
One student organization	−0.028	0.450	−0.001	−0.452	0.451
Same student status	0.269	0.115	−0.003	−0.456	0.435
Being friends	2.128***	<0.001	0.007	−0.453	0.477
Depression mean	−0.059***	<0.001	0.000	−0.023	0.024
Depression similarity	0.047**	0.004	0.000	−0.035	0.034
Depression mean * depression similarity	−0.004***	0.001	0.000	−0.002	0.002
Depression mean * being friends	−0.012	0.333	0.000	−0.053	0.052
R2	0.123
Adj. R2	0.119

Note. Multigroup MRQAPs with 5,000 Y-permuted samples. *p < 0.05, **p < 0.01, ***p < 0.001 (two-sided p-values). We report p-values because confidence intervals cannot be computed for MRQAPs. The percentiles describe the distribution under the null hypothesis and can be interpreted similarly to confidence intervals. Various robustness analyses (the two samples separately, a standard linear regression, with a non-log-transformed dependent matrix, with non-merged RFID data, and including Big Five personality traits) are reported in Table S1 and Table S2 of the Supplementary Materials.