0
$\begingroup$

I am doing a study to try to explain certain behaviour using 6 predictors. I have used a questionnaire to gather the data. The questionnaire included a demographics section eg age, gender, job position.

I an not sure whether I should include the demographic variables in the regression or not. I didn't originally plan to since my literature review focused on the 6 predictors and not on differences in demographics. I have found that two of the demographics are in fact significant in the GLM. What should I do? Any feedback is appreciated.

Improve this question
$\endgroup$
3
  • 2
    $\begingroup$ Rather relevant to answering this will be what your regression's for, how the fitted model will be used. $\endgroup$
    – Scortchi - Reinstate Monica ♦
    Feb 25, 2014 at 13:36
  • $\begingroup$ I will just be discussing the results- which predictors contribute to certain unethical behaviour occuring at the work place. But since I didn't discuss the possible effect of differences in eg gender on this behavioiur in my literature review, I do not know if I should include it or not. $\endgroup$
    – dissertationhelp
    Feb 25, 2014 at 13:38
  • 5
    $\begingroup$ Start drawing or visit your proposed causal pathway and examine the plausibility that demographics can confound with the six predictors or cause the outcome. Don't just throw them into the regression and see what sticks. If you indeed can build a good causal reasoning, then I would consider building the original one (for the sake of comparing to others' findings), and attach a further adjusted one that include demographics. $\endgroup$
    – Penguin_Knight
    Feb 25, 2014 at 13:45

0

Reset to default

Your Answer

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

Browse other questions tagged or ask your own question.