Dear all,

I am elaborating on my diploma thesis, which is devoted to factors that influence a success of projects on reward-based crowdfunding campaign.

The dependent variable is binary, i.e. equal to 1 if the project campaign successful and 0 if not.

Before that, I plan to use BMA framework (using bms package in R) to determine, which set of regressors should be included in the model. Afterwards, I will estimate logit as a robustness check.

N(# of projects)=2,870

My question is, which transformation of Y in order to gain normal distribution (which is vital for BMA application) would you recommend to use?

is it OK just to use this method?

y-mean(y)/ 2std.dev

Thank you very much for your help!

Tereza