Following Baron Kennys 1986 three step regression approach to test for

# Following baron kennys 1986 three step regression

This preview shows page 9 - 10 out of 18 pages.

Following Baron & Kenny’s (1986) three-step regression approach to test for mediation, we estimated the following three equations: MEDIATOR IV e i i i = + + α α 0 1 (1) DV IV e i i i = + + β β 0 1 (2) DV IV MEDIATOR e i i i i = + + + δ δ δ 0 1 2 (3) For mediation to exist, the independent variable ( IV ) must first affect the mediator variable in equation (1) (i.e., a 1 must be statistically significant). Second, the independent variable must affect the dependent variable ( DV ) in equation (2) (i.e., b 1 must be statistically significant). Third, mediator must then affect the dependent variable in equation (3) (i.e., d 2 must be statistically significant). In addition, the effect of the independent variable on the dependent variable must be less in equation (3) than in equation (2). 15 The results in Table 2, Panel A and Figure 1, Panel A, provide support for the Hypothesis 3 prediction that perceived oversight effectiveness will mediate the relation between IAR disclosure and confidence in financial reporting reliability. The equation (1) results indicate that IAR disclosure is positively related to investor perceptions of oversight effectiveness ( p -value = 0.03, one-tailed). In equation (2), we find that IAR disclosure is positively related to confidence in financial reporting reliability ( p -value = 0.002, one-tailed). Finally, the equation (3) results show that perceived oversight effectiveness is positively related to confidence in financial reporting reliability, even in the presence of IAR disclosure ( p -value < 0.001, one-tailed). In addition, the effect of the IAR disclosure on confidence in financial reporting reliability is less in equation (3) ( d 1 = 7.05) than in equation (2) ( b 1 = 10.91). The Sobel test statistic (Baron & Kenny, 1986) provides evidence of mediation ( p -value = 0.07, two-tailed). 16 The results in Panels B and C of Table 2 and Figure 1 provide support for the Hypothesis 4a and 4b predictions that perceived oversight effectiveness and confidence in financial reporting reliability, respectively, will mediate the relation between IAR disclosure and investment recommendation. The equation (1) results indicate that IAR disclosure is positively related to perceived oversight effectiveness and confidence in financial reporting reliability (one-tailed p -values = 0.03 and 0.002, respectively). Next, the equation (2) results provide evidence that IAR disclosure is positively related to the likelihood of investment recommendation ( p -value = 0.06, one-tailed). The equation (3) results indicate that, even in the presence of IAR disclosure, perceived oversight effectiveness and confidence in financial reporting reliability are positively related to the likelihood of investment recommendation ( p -value < 0.001, one-tailed) and that the effect of the IAR disclosure on the likelihood of investment recommendation is less in equation (3) ( d 1 = 2.98 and - 0.25, respectively) than in equation (2) ( b 1 = 7.67). The Sobel test statistics suggest mediation for both oversight effectiveness and financial reporting reliability (two-tailed p -values = 0.08 and 0.007, respectively).  #### You've reached the end of your free preview.

Want to read all 18 pages?

• • • 