Table 6 presents results from the sensitivity analyses. In column 1, estimations from logistic regression (odds-ratio, i.e. exponentiated parameter estimates) and in column 2, estimates from OLS using six-levels of CS (i.e. Pre-CS, Compliance-driven, Profit-driven, Caring, Synergistic and Holistic). In column 3, estimates from OLS using five-levels of CS, where the levels Synergistic and Holistic are merged into one category and in column 4, estimates from OLS using the average score from three raters. The following independent variables are used in the models: %WomenBoard − a percentage of female board members on; FemCEO − a dummy variable taking a value of one if the CEO is a woman, and zero otherwise; FemChair − a dummy variable taking a value of one if the Chair of the board is a women, and zero otherwise; AvgageBoard − the average age of the board members; ExtCEO − is a dummy variable taking a value of one if the CEO is external and zero otherwise; and VolReporter − a dummy variable taking a value of 1 if the firm reported on CS voluntarily before the NFRD, and zero otherwise. Furthermore, we control for the size of the board with the variable SizeBoard, and firms’ industry with the variable Industry based on Swedish Standard Industrial Classification (which are equivalent to EU Standard Industrial Classification), and other firm-level characteristics (i.e. profitability calculated as return on assets [ROA], leverage calculated as long-term debt by total assets [Leverage] and firm size calculated as log of total assets [FirmSize]). Standard robust errors in parenthesis; significance level *p < 0.1, **p < 0.05, ***p < 0.01