Hello,

I am interested in measuring how previous purchases from a given platform and time spent on the platform impact the probability of the purchase of a given item. I am using the following model

Code:
probit purchase c.lntimespent##c.lnpreviouspurchases i.categoryFE i.timeFe, cluster(customer)
However given that many customers are first time customers, my interaction term gets zero many times. In other words, although customer might spent a lot of time on the platform, if it is first time buyer, the interaction term always gets to value of zero.

What should I do in this? Center the variables? Any good references about what to do?