Working from home is here to stay, but how does it affect workplace learning?

Swiss Journal of Economics and Statistics

Table 4 Instrumental variable results

Dependent variable	Practical knowledge				Theoretical knowledge
Estimation	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Working from home	− 0.506***	− 0.335***	− 0.322***	− 0.337***	0.123	0.0681	0.0576	0.0675
	(0.0797)	(0.108)	(0.108)	(0.110)	(0.0825)	(0.113)	(0.113)	(0.115)
Observations	11,450	11,450	11,450	11,450	11,411	11,411	11,411	11,411
Month fixed effects	No	Yes	Yes	Yes	No	Yes	Yes	Yes
Occupation fixed effects	No	Yes	Yes	Yes	No	Yes	Yes	Yes
Other working activities	Yes	No	Yes	Yes	Yes	No	Yes	Yes
Respondent controls	Yes	Yes	No	Yes	Yes	Yes	No	Yes
Sanderson-Windmeijer (2016) F-Statistic for underidentification and Weak Identification	250.01	137.19	134.49	133.36	248.01	138.22	135.67	134.42
First Stage coefficient of instrument	0.492*** (0.031)	0.534*** (0.046)	0.531*** (0.046)	0.524*** (0.045)	0.491*** (0.031)	0.540*** (0.046)	0.536*** (0.046)	0.529*** (0.046)

The table shows weighted two-stage least squares coefficients. The coefficient shown for “Working from Home” is the second stage coefficient of an instrumental variable estimation. Robust standard errors clustered at the respondent level in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. Practical knowledge and Theoretical knowledge capture the impact of COVID19 on practical knowledge and theoretical knowledge of apprentices, respectively. They have a 5-Point Likert scale from substantially worse (− 2) to substantially improved (2). Details of each variable are elicited in Sect. 4.3