1
$\begingroup$

In general, how does the total sample size for stratified random sampling compare with sample size for simple random sampling?

The variable involved is binary: yes-no. The population parameter to estimate is the proportion of Yes. (For example, say there are 800 cases in 25 cities. I could form 25 strata or group the cities into 5 geographic areas to form 5 strata. (The parameter of interest is the proportion of Yes, nationwide.)

I am thinking of using strata to avoid my random sample from being strange or lopsided.

$\endgroup$
1
$\begingroup$

It depends on the homogeneity of your strata. If units are homogeneous within your strata and heterogeneous between strata, then stratified sampling will be more efficient than simple random sampling. You should look into optimal allocation for stratified sampling (specifically Neyman allocation if your sampling costs are equal between strata).

$\endgroup$
  • $\begingroup$ I have no information about the homogeneity of the potential strata. I am thinking of using strata to avoid my random sample from being strange or lopsided. $\endgroup$ – Joel W. Jul 12 at 22:48

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.