Note(s): (1) In line with the proposed procedure of Tabachnick and Fidell (2014) and to comply with the conditions of the logistic regression, the linearity of the logit was examined for metric variables by evaluating the interaction between the determinants and the ln transformation of the respective determinant. Due to a violation of the linearity assumption by Intern, the variable was transformed into Intern_transformed $=1(Max(Intern)+1−Intern$ ; to normalize the strong left-skewed distribution, a reciprocal transformation was used, thereby adding “Max_(Intern)+1” to ensure that the transformation was defined over all values of Intern (Field, 2018); (2) The logistic regression was repeated without the Intern_transformed variable. Both the values for Nagelkerke’s R² and AUC as well as the results concerning the significances and relations of the determinants with CyberIns remained unchanged; (3) The reference category for the categorial variable of Size was large enterprises, whereby Size (1) indicates micro-enterprises, Size (2) small enterprises and Size (3) medium-sized enterprises; ^*** indicate the 1% statistical significance level

Source(s): Authors own creation