Table 3.

Robustness check: heterogeneous treatment effects

	(1)	(2)
	OLS	OLS
C1: Dig. increases the time pressure at work	$-$ 0.0776***
	(0.0128)
C1 * How strongly Dig. affect the work over the last year?		$-$ 0.107***
		(0.0175)
C2: Dig. puts my job at risk	$-$ 0.0343***
	(0.0133)
C2 * How strongly Dig. affect the work over the last year?		$-$ 0.0251
		(0.0154)
C3: Dig. worsens the work–life balance	$-$ 0.102***
	(0.0137)
C3 * How strongly Dig. affect the work over the last year?		$-$ 0.120***
		(0.0171)
C4: Dig. leads to a smooth transition between working hours and leisure time	0.00174
	(0.0131)
C4 * How strongly Dig. affect the work over the last year?		0.00758
		(0.0171)
C5: Dig. makes my work more interesting	0.155***
	(0.0154)
C5 * How strongly Dig. affect the work over the last year?		0.233***
		(0.0231)
C6: Dig. reduces the proportion of repetitive tasks	0.0193
	(0.0128)
C6 * How strongly Dig. affect the work over the last year?		0.0267
		(0.0178)
C7: Dig. increases my productivity	0.255***
	(0.0161)
C7 * How strongly Dig. affect the work over the last year?		0.416***
		(0.0259)
C8: Dig. increases my autonomy at work	0.0380***
	(0.0147)
C8 * How strongly Dig. affect the work over the last year?		0.0516**
		(0.0207)
C9: Dig. enables more flexible forms of working time	0.0269**
	(0.0116)
C9 * How strongly Dig. affect the work over the last year?		0.0365**
		(0.0153)
C10: Dig. simplifies interaction with colleagues and superiors	0.0548***
	(0.0136)
C10 * How strongly Dig. affect the work over the last year?		0.0772***
		(0.0196)
How strongly Dig. affect the work over the last year?	0.112***	$-$ 0.260***
	(0.0121)	(0.0281)
Worker characteristics	Yes	Yes
(N)	3089	3089
($R^2$)	0.379	0.381

This table reports the results of the OLS regression having as dependent variable the effect of digitalization on job satisfaction, which is measured on a five point Likert scale (1 = “less satisfied”, 3 = “no change”, 5 = “more satisfied”). Robust standard errors in parentheses. *$p<0.10$, **$p<0.05$, ***$p<0.01$. Worker characteristics is a vector of control variables as described in Table 1. Column (1) reports the baseline estimation as in Table 2, while column (2) reports the estimation of the model in which we allow the channel to interact with the overall effect of digitalization on work