Hello,
I have quite a complex model with 10 first order LVs, a second order LV (global service benefit) that then impacts the outcome variables satisfaction and general service quality. the latter two on their behalf influence another outcome vairable (loyalty).
I can calculate bootstraps for this model just fine. But then, I also want to test interaction effects for four moderators who influence the relationship from the global service benefit to satisfaction and general service quality. When I run bootstraps for this model, my computer goes on strike. The bootstrap calculation goes to about 136 samples, but I want to run it for 500.
Do you think it is detrimental to only calculate the bootstrap with about 136 samples instead of the 500?
A friend of mine, a computer wiz, said that the problem I report has to do with the computer's RAM. Apparently, the RAM is used for storing the ongoing calculation. The required memory of the calculation probably exceeds the RAM. The computer wiz said that the program might have a flaw because in such cases, whenever the limit of the RAM's capacity is reached, the program should use the computer's additional hard drive space. Any takes on that?
Calculating bootstraps for complex models
Calculating bootstraps for complex models
Doctoral Candidate
Gottlieb-Duttweiler Chair of International Retail Management
University of St. Gallen
Gottlieb-Duttweiler Chair of International Retail Management
University of St. Gallen