0
$\begingroup$

I have read the comments about inferential statistics on census data. I am doing PhD and also facing the same problem with a little difference.

I have collected census data from the population of deans, directors,chairpersons and head of departments because of lack of availability on their part and can get about 64 % response rate.

can i take this 64% response as my sample and by applying inferential statistics eg t test and One way ANOVA draw conclusion regarding mean difference on the basis of gender, age, experience, administrative posts and generalize the results on the whole population. iif it is so then please send me the reference of its justification so that i can defend this conceptual point in my phD thesis. i have been trying a lot but can not find the reference. please is there any one who can help me i shall appreciate and acknowledge it in my dissertation.

$\endgroup$
1
$\begingroup$

You are "allowed" to do this. The problem is trying to justify that the 64% who responded are a random sample of the population. (This is the same problem as with all sampling).

While 64% is a pretty good response rate, compared with a lot of studies, the results could nevertheless be very misleading if there is bias.

If you have any information on the 36% who did not respond, you can use that to compare with the 64% who did. That doesn't eliminate the possibility of bias, but it does go some way towards limiting it.

$\endgroup$
  • $\begingroup$ Sir I have to use it in my phD thesis. I cant do any thing without reference. would you please give me a hint where i can find its reference so that i can quote and defend it before my thesis cometti $\endgroup$ – sadaf jabeen Dec 5 '13 at 7:31
  • $\begingroup$ I don't have a specific reference in mind. Sorry. $\endgroup$ – Peter Flom Dec 5 '13 at 11:11
  • $\begingroup$ @sadaf: A common technique is to approach some of the 36% of non-responders & cajole/bribe/menace them into responding - resource constraints won't allow you do this for all, but you then have a sample from the non-responder group - it's a way of getting the information on them that Peter mentioned $\endgroup$ – Scortchi Dec 5 '13 at 11:44

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.