Dear all,

Im am doing an OLS regression. My dependent variable has large negative values, many zeros and large positive values. How can I transform my data in such a way that I can take the log of this data?
Log cannot be applied to negative values and zeros.

To solve for zeros, I can add a constant (1) to each variable. But this does solve for the negative values.

So my question is, how do I have to transform my data, so that I can take the log of the whole data set, including negative, positive and zero values.

I hope I formulated my question more clear.


TS Contributor
If the negative values are real, i.e. not the result of some error, then taking logs makes no sense. One option is the cube root is a transformation; it is somewhat similar to a log but can handle negative values.