SPSS for newbies: log transformation

SPSS/PASW tutorial for beginners
www.statisticsmentor.com
How to do and undo a log transformation in SPSS/PASW.

[@ThereseKearney]

Excellent video, Very easy to follow, precise and to the point. thank you.

[@georgetaylor100]

Great video but when would you use log10 instead of Ln? Someone answer ASAP

[@abhijithvasista4952]

Every video shows only one variable. I have 60 variables. Should I do it manyally one by one or is there a one click solution for multiple variables?

[@loaybal]

Thanks a million mate, that was very helpful!

[@ThereseKearney]

Thank you soo much. Very helpful tutorial.

[@00Saiyan]

Thanks, you really helped me out here :)

[@TRIPLEJUMP1975]

Thanks for this, very helpful.

[@ASleeplessWolf]

What about doing this conversion for coefficients for the sake of writing out predicted values equations? Then you need to use Dian's Smearing Factor, I believe.

[@PhilChanstats]

Interesting. You may use the Duan Smearing Factor is the error term of the regression is not normally distribution, but is homoskedastic. The Factor is used to go from lnY back to Y, so it's not about the coefficients.

[@brac5933]

Is there a journal article that I can reference on adding the +1 because my measure included zero?

[@PhilChanstats]

Google for "add start for log transform"

[@TheAg605]

spot on guvna!

[@ThereseKearney]

Thank you sooo much

[@mvgaming1703]

Hi im from Indonesia

[@rassten]

I get the how, but not why.

[@PhilChanstats]

Sten. Your question is very natural. The answer to your question why one may wish to take log in regression is in my video on the box cox transformation. Outside of regression, you may take a log transform to help make a variable that is positively skewed more symmetric for a procedure that requires the variable to be or approx normally distributed.