Hello everyone, I am analyzing the influence of some stores attributes (e.g., quality, price, instore service, communication ...) on customer loyalty (more precisely share of wallet = how much a customer allocates to a specific store in %: 0100).
I transformed the dependent variables (share of wallet) using the log feature in order to overcome nonnormality of data.
However, I cannot figure out how to interpret the final number in an economic way.
Basic interpretation shows that increasing the IV by oneunit would increase the DV by 100*<Standardized Beta coeff>%.
Since the original unit of measurement was already in percentage (share of wallet), what would be the final interpretation?
Do I have to multiply 100*<Standardized Beta coeff>% by the average percentage allocated? or ...?
Also, my descriptive data are on an average level for each store, not per individual customer ...
Thanks a million for your help !!!
I transformed the dependent variables (share of wallet) using the log feature in order to overcome nonnormality of data.
However, I cannot figure out how to interpret the final number in an economic way.
Basic interpretation shows that increasing the IV by oneunit would increase the DV by 100*<Standardized Beta coeff>%.
Since the original unit of measurement was already in percentage (share of wallet), what would be the final interpretation?
Do I have to multiply 100*<Standardized Beta coeff>% by the average percentage allocated? or ...?
Also, my descriptive data are on an average level for each store, not per individual customer ...
Thanks a million for your help !!!
Attachments

41.1 KB Views: 1