How to create a representative sample?

#1
Hello to all! :)
I have recently launched a survey to students of specific universities among European countries and got nearly 3000 replies.
However my sample is not balanced, meaning that some countries provided many replies, while some others just few.

I want to analyse my results at European basis, therefore I think that this lack of balance will harm the representativeness of my sample.

I tried to balance the sample by putting weights according to number of students per university but it didnt work, cause big universities didnt provide equally big input etc.:confused:

So I am now considering to follow a "flat" approach, meaning taking equal number of replies per university. This will drop my sample to 700 replies.
Do you think this approach makes sense?
Can I test somehow if my sample is ok without cutting off so many replies? And if yes, how?

Thank you in advance!
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
Did you collect any demographic or characteristic data?

You you may want to further investigate weighting approaches.
 
#3
Thank you for the reply.

Yes, I have collected:
- age
- gender
- country
- university
- field of study

Age and gender are very well balanced. My concern is: if I generalise the outcomes at European level, will the students' input be representative, or it will be biased according to their country? Can I check somehow if origin plays a significant role in the replies?