Table 2.

Linear regression models of the overall effect of DAI on SI and its regression per class (low, medium, high).

	β	Std. error of β	t	p-value
Overall effect of DAI on SInorm
Constant	− 0.40	1.00	− 0.398	0.691
DAI	4.86	1.75	2.780	0.006**
Effect of low, medium, and high DAI on the trend in SI
Constant	38.61	12.10	3.191	0.003**
Low [0;0.36]	54.18	42.01	1.290	0.205
Constant	71.18	9.98	7.133	0.001**
Medium [0.37;0.68]	− 10.61	18.23	− 0.582	0.562
Constant	102.36	18.75	5.458	0.001**
High [0.69;1]	− 55.26	24.32	− 2.272	0.028*

* p ≤ 0.05; ** p ≤ 0.001.

The regression coefficient (β) is the degree of change in the outcome variable for every one-unit change in the predictor variable. The t-statistic is the regression coefficient divided by its standard error. Digital adoption is expressed in arbitrary units of DAI. The numbers in squared brackets indicate the intervals of the classes.