im using SPSS to run regression analysis for my Phd, but I keep getting low values and no significance (ie. > 1)..


My dataset is 103, does anyone know what I can do to get better significance levels ?:-(

Check for multicollinearity.

You should have 10-15 participants per variable, so maybe you need to reduce the number of variables.

Also - maybe your model just isn't predicting.

thanks, multicollinearity is ok,

i have 50 variables, 103 firms, but I dont run all of them at the same time......

usually i run 1 on 1 or 3 on 1 maximum...

is it just too small a sample?

======= Date Modified 18 Jul 2011 13:51:51 =======
yes, ifyou have 50 variables (and you need 10-15 participants/organsiations per variable) then you need 750 participants/organisations.

But even then, have 50 variables in a regression isn't going to result in any decent findings - each variable takes up variance, so by adding in so many, you're going to find very little effect of any single variable.

Most regressions I run have about 5-7 variables in.

if you're not running them all at the same time, then that's ok. but you may still need a higher sample size. OR you're results aren't significant ;-)