STATS MAKE ME CRY

Hi Jeremy,

As per your advice, posting this on your fabulous site -

I am testing for the success drivers (independent variables such as star power, advertising spend, critic reviews, etc) to impact positively on box office revenues (the dependent variables). I am testing against the (i) Opening box office revenue, and (ii) Overall box office revenues, in two separate regression models, to elicit the IV's to positively impact box office revenue (for each model, and to compare both).

So the help required? -

My DV's (opening box office, and overall box office) are NOT 'normally distributed' at all (but this is quite usual for box office revenues in the UK and worldwide, as proven in my literature). However, it therefore does not fall under the 'normality assumption' for running reg analysis.

Prev scholarly articles that have tested for the phenomena I am researching (i.e. box office success drivers), have mainly transformed their data or some have even used 'bootstrapping'. I have tried a log transformation but data still NOT at all normally dist.

Not sure you can advise without looking at my data, but if you can, what do you advice i do Jeremy?

Thanks for your help, Best, J

Also, once bootstrapping, the bootstrapping table displayed in the output, doesn not give u a standardised beta coefficient column (just the unstardardised coeff's), so what figures do you use to interpret the results?