1
$\begingroup$

I have a population size of about 600 people whom I need to survey(online) for career options they chose after college and was trying to see what my options for sampling were?

Option 1: Complete survey with all people(census).

Option 2: Pick a random sample, but assuming a response rate of about 20%, it seems like the marginal error rises to about 8% which would make generalizations difficult.

Any help on what a good sampling strategy will help. Ideally, I would like to also make cuts on the population to see how gender,age,location,salary etc. have also made an influence on their decisions.

Well, the population size comes from the number of graduates we have had over the last 10 years and the survey is meant only for them. In that case, would a census be better than a random sample given what i have said in the first post.

Any help will be greatly appreciated.

|cite|improve this question
$\endgroup$
  • $\begingroup$ I think you are a bit confused between population and sample. If your whole population is 600 you are dealing with a finite population which will need its own specific sampling requirements and analysis. So it's best to start with an understanding of what is your target population, it's size and then decide on sampling... if your population is those who have completed a colleague, its going to be an infinite population. So you may want to include more info on where you have got this 600 population size. hope this helps $\endgroup$ – RomRom Oct 24 '16 at 2:54
  • $\begingroup$ Well, the population size comes from the number of graduates we have had over the last 10 years and the survey is meant only for them. In that case, would a census be better than a random sample given what i have said in the first post. $\endgroup$ – AdiR Oct 24 '16 at 13:19
  • $\begingroup$ This would be students of a specific university in a specific program $\endgroup$ – AdiR Oct 24 '16 at 21:52
  • $\begingroup$ I updated the answer below $\endgroup$ – RomRom Oct 24 '16 at 23:38
  • $\begingroup$ Are there monetary or analysis constraints that are causing you to limit the sample? If not, use everyone available. The only reason you would leave out some of the graduates is if it costed too much money or time. $\endgroup$ – Pierre Lafortune Oct 24 '16 at 23:42
0
$\begingroup$

If you can afford, do a census. By definition it will be superior to a sample survey, even if there is nonresponse (which you will have in both). If you can't afford you have to do a sample survey leaving the issue of sample size, which again is a cost issue. Maxmize you sample size on available budget.

|cite|improve this answer
$\endgroup$

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.