Ugh. More sampling issues. I just finished polling editorials from ad-Difa` in post-war 1967. Virtually none of them deal with domestic issues. And ad-Difa` is the privately owned newspaper. Of course, it is a privately owned newspaper that survived the post-’67 War establishment of martial law by toeing the government line. I did not anticipate this in my research design. Reality is fuckin’ messy.
This makes me wonder if I should tally editorials together with articles at all. Right now, I am sampling the top domestic policy article (usually short and buried somewhere after p. 3 of a six-page newspaper). If there are no domestic policy editorials to sample, should I allow them to be tallied together? My target was a combined 300 pieces per survey, as that is the most likely minimum number to give me p ≤ 0.05. If I can't combine them as a single survey, then I'd need to do a separate 400-piece survey for each. I really can’t envision doing 300 articles and 300 editorials for each period, with six total periods. I need to graduate. Should I just dump the editorials?
I guess I need to start looking at Lebanon. I need to know the look and feel of a Lebanese paper before I can make a decision.
No comments:
Post a Comment