I am facing a problem that I had never encountered before, so I am relying on the collective intelligence in order to get at least a hunch on what to do.
In doing content analysis of news items in two different newspapers, whose coverage of a certain issue I am comparing, I have come across unequal population sizes. Newspaper A has got about 1,500 news items spanned over 11 years, whereas newspaper B, more prominent and older than A, has got over 16,000 over the same period.
Given this disproportion, and given common time and resource constraints that impose to keep the data amount manageable, is it legitimate to use different sampling methods for the two newspapers?
Also, would it make sense to use different proportions of the populations to make the sample size, e.g. one third for A and one tenth for B?
I realize the problem may not have a straight-forward answer and it definitely depends on several factors, starting from the research design, however since this is a problem that it seems it has not been dealt with at great extent in the literature, I thought it may have been interesting to share it.
Thank you in advance to all those who will want to contribute this discussion.