Table 2. Differences in Stages of Action for Radon Testing and Mitigation, Participants in Treatment (n = 257) and Control (n = 258) Groups by Participant Characteristics, Calculated by Linear Mixed Models, the FRESH Intervention,^a Central Kentucky, January 2013–August 2017 .

Variable	Stage of Actionb
	Radon Testing (n = 499)		Radon Mitigation (n = 499)
	F	P Valuec	F	P Valuec
Age	2.73	.098	1.43	.23
Male	0.35	.55	<0.01	.96
Non-Hispanic white	4.00	.046	0.07	.80
College education	1.64	.20	1.09	.30
Employed for wages	1.20	.27	1.78	.18
Years living in home	0.50	.48	2.05	.15
Family history of lung cancer	0.32	.57	0.03	.86
Smokers in the home	2.89	.089	1.13	.29
Self-efficacy	47.29	<.001	36.18	<.001
Lung cancer worry	0.04	.83	0.23	.63
Lung cancer risk	<0.01	.97	0.88	.35
Synergistic risk	3.21	.074	15.42	<.001
Health-related self-concept	3.82	.051	0.13	.71
Risk group	—	—	85.73	<.001
Time	426.88b	<.001	226.62b	<.001
Treatment	49.90b	<.001	9.57b	.002
Treatment × time	23.21	<.001	34.12	<.001

Abbreviation: —, not applicable.

^a FRESH was a randomized controlled trial to promote stage of action to reduce radon and air nicotine levels in the home.

^b Regressions modeling stage of action to test for radon and radon mitigation ranging from 1 (unaware) to 5 (maintenance over time). Main effects were not interpretable in either model given the presence of a significant interaction effect; means for the interaction effect are shown in Figure 2. Although the full sample size was 515, 16 participants missing 1 or more variables in the models could not be included in the multivariable analysis.

^c P values calculated by type 3 tests of fixed effects in the mixed models.