0
$\begingroup$

Trying to determine if a logarithmic regression analysis is valid. With good R2 values, does it matter that there are ony 7 data points?

$\endgroup$
0
$\begingroup$

In my opinion 7 observations is too small no matter what kind of regression you are running. I would recommend increasing the sample size to get more precise estimates.

It heavily depends on the specification that you are running though, as well as the model. With a logarithmic regression you mean OLS but with the log of the variables I assume. In either case, if you have many regressors and are dealing with a low sample size you'll have problems with low degrees of freedom, translating in a lack of significance. Also if the effect sizes that you are dealing with are small, you'll need more observations.

If you are working with experimental data try to do a power analysis beforehand to get an idea of the sample size needed.

Hope this helps.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.