you may have many.. but i would be good tips for you to take care of these things while selecting.

1. look at availability of university.  

2.keep your resources in mind.

3.keep the stander. select international one rather than any one.

Dear Emmanuel,

Attached find two texts that may help to guide your research.

First, consider and clarify what  your central research questions are. In most countries, the student populations are very specific with regard to some socio-demographic traits, i.e., they will be quite different from the average population. I therefore assume that you are not trying to find a student sample that somehow reflects the average population of the country but rather to find universities with students that somehow reflect the average of all students in that respective country. In that case, I would recommend a so-called "purposeful sampling" (also known as "judgment sampling").

Best,

Pradeep

Emmanuel -

I'm not sure what kind of survey you are doing.  The description does not sound like distance or behavioral sampling from definitions I checked, though those are listed as two of your 'topics,' but it does sound like a possible multistage random survey design such as you might find in a survey sampling book such as Cochran(1977):

Cochran, W.G(1977), Sampling Techniques, 3rd ed., John Wiley & Sons. 

You could randomly selected universities in one stage, and students in the next.

You might research "multistage sampling" and "subsampling" as keywords on the internet. 

To improve accuracy, if you could stratify your sample by common factors, that is generally helpful. But if you are to travel, then cluster sampling may be less expensive. You might consider the first stage, universities, to be sampled by PPS sampling, but simple or stratified random sampling may be good enough for both stages. 

The above assumes you want to do a quantitative study from which you can make inferences.  Variance and bias are important, and you not only need to deal with sampling error, but nonsampling error as well. 

Terms used above may be researched on the internet. For continuous data or yes/no questions (proportions), Cochran is very good, but there are a number of other good textbooks. Here are just a couple more you might consider:

Blair, E and Blair, J(2015), Applied Survey Sampling, Sage Publications

and

Lohr, S.L(2010), Sampling: Design and Analysis, 2nd ed., Brooks/Cole.

More advanced books may include

Särndal, C-E, Swensson, B, and Wretman, J(1992), Model Assisted Survey Sampling, Springer-Verlang,

and

Brewer, KRW (2002), Combined survey sampling inference: Weighing Basu's elephants, Arnold: London and Oxford University Press.  

But a successful survey always depends upon determining what kind of data you are seeking, and setting  your goals. Planning ahead, as you apparently are doing, is very important. You might also start with a small pilot study to see what issues you may incur, and for such a quantitative study,  to help estimate sample size needs (first estimating sigmas). 

The applicability of the above does depend upon the type of survey you are doing, but there are often analogous needs, to an extent, for some issues across different kinds of surveys, and different kinds of data. 

Best wishes - Jim