1
$\begingroup$

Looked but could not find so I apologize if this is a repeat of a question already asked -

Can I conduct an EFA with an N of 180 and 70 variables? I say NO but my advisor still wants me to do it.

Thank you in advance!

$\endgroup$
3
  • $\begingroup$ I know of no reason why not. EFA is based on the correlation matrix, so, while larger samples may yield more precisely estimated correlations, you don't run into the overfitting problems that you would with regression. $\endgroup$
    – Peter Flom
    Sep 30, 2012 at 0:49
  • $\begingroup$ So the sample size rules or variable to N ratio found in texts are just archaic guidelines? Just wondering, after reading the de Winter article from 2009 - seems like the rules are old! $\endgroup$
    – Glenlivet
    Sep 30, 2012 at 2:30
  • $\begingroup$ Not srue which ratios you are talking about or for what procedures. What's in the de Winter article? $\endgroup$
    – Peter Flom
    Sep 30, 2012 at 11:44

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.