1
$\begingroup$

I need to compare two groups of people, where the independent variable is having a college degree, and the dependent variable is the income.

The problem is that if I divide the whole population of the study into two groups, one of them has signally more persons, so the mean is affected by the size of each group and the value of the outliers.

How can I demonstrate or reject the theory that having a college degree among the population of the study, ensures a better income?

$\endgroup$
2
  • 1
    $\begingroup$ Depends on how sophisticated you'd like to go. The median is a good measure for getting rid of outliers, but if you want to be extra technical, you can always do an EM algorithm en.wikipedia.org/wiki/… $\endgroup$ – Landmaster Aug 16 '17 at 0:22
  • $\begingroup$ Thanks, at first we only need something simple, in the future, we might need something more elaborated $\endgroup$ – ulitosCoder Aug 17 '17 at 19:58
1
$\begingroup$

I assume you are performing an independent samples t-test. That the Ns are different isn't necessarily a problem -- the mean is an unbiased estimator -- but just how different are the sample sizes? As you describe, you are probably violating the homogeneity of variance assumption.

T-tests tend to be fairly robust to this violation, but you might consider having some outlier treatment and/or using a nonparametric test.

$\endgroup$
1
  • $\begingroup$ Do you mean that I could run t-test between samples of the two groups? Probably that could work. I'd like to see some examples. $\endgroup$ – ulitosCoder Aug 22 '17 at 15:30

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.