Table 4. Binary logit model: change in working from home during pandemic

Independent variables	Dependent variable: Increased working from home
	First survey (Y: 491, N: 358) coef. (s.e.)	Second survey (Y: 337, N: 335) coef. (s.e.)
Gender: Female	0.109	0.197
	(0.158)	(0.164)
Aged 50 years or older	0.372*	–0.522***
	(0.187)	(0.202)
White	0.270	0.182
	(0.167)	(0.174)
Children present in the household	–0.289*	–0.040
	(0.163)	(0.169)
Educational attainment (Ref: high school degree or less)
Associates or some college (no degree)	0.768***	0.591*
	(0.279)	(0.278)
Bachelors	0.968***	0.782***
	(0.268)	(0.273)
Masters or higher	0.973***	0.483*
	(0.273)	(0.291)
Attitudes towards the COVID-19 pandemic
Believers	0.038	–0.060
	(0.087)	(0.100)
Deniers	–0.477***	–0.247***
	(0.088)	(0.094)
Transportation
Vehicles per adult	0.274	0.111
	(0.170)	(0.310)
Long commute (1 + hour) before COVID	0.678***	0.567*
	(0.245)	(0.252)
Had a job could be done from home before the pandemic began	0.193	0.813***
	(0.162)	(0.181)
Constant	–0.832***	–1.002***
	(0.319)	(0.304)
Observations	849	672
Log likelihood	–534.474	–438.044
Log likelihood of intercept only model	–585.936	–466.481
Akaike Inf. Crit.	1,094.949	902.088
TJUR R2 (Tjur, 2009)	0.10	0.08

p<0.10,
p<0.05,
p<0.01.