The dependent variable, firm innovation (measured by patents), is a nonnegative-integer count variable and has a problem of over dispersion, which tends to bias downward the estimated standard errors (Huang, 2011). Prior research suggests that negative binomial regression model is a remedy to overcome the over-dispersion problem
(Keil, Schildt, & Zahra, 2008). Thus, this research follows the suggestion by using the negative binomial regression to examine the hypotheses. At the first stage, this research establishes regression models to examine
The mediating effect of absorptive capacity on the casual relationship between independent variable (R&D investment) and dependent variable (firm innovation). Based on the Edwards and Lambert’s (2007) study, this research has to examine three relationships among R&D investment, absorptive capacity, and firm innovation for the mediating
effect. First, this research has to examine the relationship between R&D investment and firm innovation (see Eq. (1)). Second, the research also separately examines the relationship between R&D investment and absorptive capacity (see Eq. (2)) as well as the relationship between absorptive capacity and firm innovation (see Eq. (3)).