We first consider results for gende

We first consider results for gender as the moderatorvariable. Coefficient estimates in Table 1 show that gendermoderated the path from feedback to satisfaction (aXZ50.14, p .01), the path from feedback to commitment(bXZ200.13, p .01), but not the path from satisfactionto commitment (bMZ200.01, p .05). Equation 25 wasapplied to coefficients in Table 1 to compute simple effects,as reported in Table 2 and portrayed in Figures 2A and 2B.For men, Z 0, and the first stage, second stage, and directeffect reduce to aX5, bM20, and bX20, respectively, whichequal 0.81, 0.31, and 0.28. The indirect effect for menequals the product of the first and second stages, or 0.810.31 0.25, and the total effect equals the sum of the directand indirect effects, or 0.28 0.25 0.53. For women,Z 1, such that the first stage of the indirect effect becomesaX5aXZ50.81 0.14 0.67, the second stage becomesbM20bMZ200.31 0.01 .030, and the direct effectbecomes bX20bXZ200.28 0.13 0.15. As for men,the indirect effect for women equals the product of the firstand second stages, or 0.67 0.30 0.20, and the totaleffect is the sum of the direct and indirect effects, or 0.150.20 0.35. Comparing these effects for men and womenshows that the first stage of the indirect effect was strongerfor men (0.81 0.67 0.14, p .01), whereas the secondstage did not differ for men and women (0.31 0.300.01, p .05). When multiplied, the first and second stagesdid not produce a significant difference in the indirect effectfor men and women (0.25 0.20 0.05, p .05).However, the direct effect was stronger for men than forwomen (0.28 0.15 0.13, p .01) and, when combinedwith the indirect effect, produced a larger total effect formen (0.53 0.35 0.18, p .01).Differences in these effects are depicted as simple slopes inFigures 3A through 3E. As seen by comparing Figures 3A and3C, the moderating effect of gender on the first stage was notsufficient to produce a meaningful difference in slopes for theindirect effect because of the absence of a moderating effect ofgender on the second stage indicated by Figure 3B. ComparingFigures 3C and 3D further shows that the difference in slopesfor the direct effect was the primary reason for the differencein slopes for the total effect in Figure 3E. Thus, gender mod-erated the direct effect of feedback on commitment and the firststage of the indirect effect of feedback on commitment medi-ated by satisfaction, and these differences were sufficient to produce a larger total effect for men.