Table 2. Standardized odds ratios and regression coefficients for dispositional compassion and empathy in relation to health behaviors and prosocial attitudes, in models with and without political ideology.

Exposure Variable	Outcome Variable	Model 1		Model 2
	Health Behavior Engagement	Odds ratio (95% CI)	P	Odds ratio (95% CI)	P
Compassion	Whether one stayed home for others	1.58 (1.47, 1.70)	< .001	1.55 (1.44, 1.66)	< .001
	Health Behavior Frequency	β (95% CI)	P	β (95% CI)	P
Compassion	Time spent staying home	0.12 (0.09, 0.15)	< .001	0.11 (0.09, 0.14)	< .001
Compassion	Time keeping distance from others	0.14 (0.12, 0.17)	< .001	0.14 (0.11, 0.16)	< .001
Compassion	Frequency of mask wearing in public	0.13 (0.09, 0.18)	< .001	0.11 (0.06, 0.15)	< .001
	Prosocial Attitudes
Empathy	Can understand others’ fear	-0.26 (0.23, 0.29)	< .001	0.23 (0.20, 0.26)	< .001
Empathy	Viewing others as overreacting	-0.25 (-0.28, -0.23)	< .001	-0.19 (-0.22, -0.17)	< .001
Empathy	Belief that staying home prevents spread	-0.23 (0.20, 0.26)	< .001	-0.23 (0.20, 0.25)	< .001

Note. Standardized odds ratios and beta coefficients (β) are presented for both models. Model 1 included the following covariates: age, sex, race, ethnicity, gender, income, education, employment, social desirability, personal diagnosis of COVID-19, increased childcare responsibilities, cohort number, and sample source (MTurk or Qualtrics). Model 2 included all Model 1 covariates along with political ideology. Outcomes were assessed at different points in the study, and the resulting available data for each outcome were: n = 5,533 for the outcomes of whether one stayed home, time spent home and keeping a distance from others; n = 1,867 for the outcome of mask wearing; n = 5,533 for the outcomes “Can understand others’ fear” and “Viewing others as overreacting”; and n = 4,635 for the outcome “Belief that staying home prevents spread”.