I have results from multiple regression where t statistics are not reported. I know sample size, number of predictors, p-values of predictors etc. To estimate partial r, I need to know the t-statistic. Would you know, how to calculate it? I know the other way round, but that is useless for me now... Thanks for your considerations.
My Dear
you can depend by more than methods :one use table and search of t-statistic using two criteria(df ,p-value), second one by matlab software or Excel
Guys, thanks. Appropriate software would be Excel or SPSS in my case. I would appreciate more elaborate answer though.
It is for the purposes of a meta-analysis, therefore I do not have the raw data etc.
Great, thank you very much. I will check it and if it works it is exactly what I need.
In SPSS, the function is INV.T(p, df). You can read about it in the Universals section of the Command Syntax Reference manual (see link below).
https://www.ibm.com/support/knowledgecenter/en/SSLVMB_21.0.0/com.ibm.spss.statistics.help/invdistfunctionlist.htm
I checked in Excel and it seems T.INV.2T works little bit better for me. I guess it depends on whether the hypothesis originally tested was directed and therefore whether the reported significance is two-tailed or one-tailed. As I have results from multiple regressions, I believe they report two-tailed significance as they do not work with hypothesised direction... Jochen, thank you very much. Bruce, thank you as well.
- Badges
- Science topic