Impact Of Accessibility Signals On Knowledge Diffusion

1926 Words8 Pages
Figures 6a, 6b, 7a, and 7b provide visual insight about the effect of utilizing accessibility signals in 8 blocks of our experiment. Figure 6a, compares the effect of utilizing accessibility signals on knowledge diffusion in blocks with heterogeneous distribution of knowledge and 1) local, and 2) global reputation mechanisms, whereas Figure 6b does this comparison in blocks with homogeneous distribution of knowledge and 3) local and 4) global reputation mechanisms. Similarly, Figure 7a compares accessibility signal’s effect in blocks with homogeneous distribution of knowledge and 5) low network density, and 6) high network density, whereas Figure 7b does this comparison in blocks with heterogeneous distribution of knowledge and 7) low…show more content…
This measure abstracts extent of diffusion in the network (Abrahamson and Rosenkopf 1997). A dummy variable called “heterogeneous knowledge distribution” is created with a value of 1 when the distribution of knowledge is heterogeneous, and 0 when the distribution is homogeneous. Further, we model accessibility signals as 0/1 dummy variables where 0 indicate no use of a signal and 1 indicates the use of a signal. Likewise, a dummy variable called “global reputation mechanism” is created which is 1 when the reputation mechanism is global, and is 0 when the reputation mechanism is local. Also, we input the inverse number of questions that each agent can answer per round of simulation as the measure of its time constraint (low- 1/10, high- 1/5). Finally, we use the parameter value of network density (0.3, 0.7) as the network density variable’s value in our OLS model. Table 1 reports the beta coefficients of accessibility signal dummy in each of the eight blocks mentioned above . Results of this split-sample analysis, conforming to patterns seen in the plots, show that accessibility signals have positive effect in blocks 4 and 6, whereas they have negative effect in blocks 1 and 8, and insignificant effect in blocks 2, 3, 5, and 7. As an alternative to the split sample test, we run regression models
Open Document