Table 3.

Adjusted mean comparisons on Homophobic and Transphobic Belief Scale at the 3- and 12-Month follow-up survey timepoints

Time point	Sample	N	Beta (SE)	p-value	ICC
3-month follow-up	Full sample	1357	− 0.072 (0.026)	0.005 < 0.01	0.000
	Straight & cisgender sub-group	1144	− 0.077(0.028)	0.005 < 0.01	0.000
	LGBTQ sub-group	213	− 0.021(0.065)	0.75	0.000
12-month follow-up	Full sample	1275	− 0.076 (0.028)	0.008 <0.01	0.001
	Straight & cisgender sub-group	1078	− 0.057 (0.030)	0.05	0.000
	LGBTQ sub-group	197	− 0.174 (0.068)	0.01	0.000

Models adjusted for baseline value on belief scale, age, gender, region (Midwest or South), school size, race/ethnicity, time, religion, and report on whether students’ mother had children as a teen. The beta represents the difference in the adjusted means between treatment and control groups. Cases with missing values were excluded from the models (n = 81 at 3-month follow-up and 120 at the 12-month follow-up). Baseline characteristics of those excluded were similar to those in the starting analytic sample. There was no differential attrition between treatment conditions at 3 months. At 12 months, excluded cases from the FLASH condition were more likely to report English as a primary language than excluded cases from the comparison group; no other differences were present