[Air-L] Sampling size for content analysis of comments on weblogs

Saphron Hastie s.hastie at auckland.ac.nz
Tue Jul 5 16:03:36 PDT 2011


Kia ora John,


I have done some online content analysis and it really depends on your methodology more generally and your sampling method in particular. I would recommend that you go back to basics and just treat it like you would any quantitative sample. And if you're doing a qualitative analysis then the same thing - maybe aim for saturation point? Depending on the blog I don't see any reason why you can't treat it in a similar manner to a study of letters to the editor or of headlines or any other traditional content analysis text?

Assuming you're doing a straightforward quantitative content analysis with a random sample; and have a complete population (say you have access to all comments on one weblog) then I would probably go for a percentage of the total that, if proportional, will give you a specific minimum value of the smallest relevant variable of interest. So, say you have 1000 comments in total and you were interested in gender. You have assumed or know that 40% of comments are from women. Then you decide based on your project resources that you want a minimum value of 40 comments for each variable (in this case male and female). In that case you would go with, say, 10% of all comments which should give you a random sample of 100 comments including, if proportional, about 40 comments from women. Although in reality 100 cases might be a better minimum value although, again, it depends on how many variables you are looking at and what and how you're counting within each case.




Kind regards,

Saphron Hastie

-----Original Message-----
From: air-l-bounces at listserv.aoir.org [mailto:air-l-bounces at listserv.aoir.org] On Behalf Of John McAuley
Sent: Tuesday, 5 July 2011 4:26 a.m.
To: air-l at listserv.aoir.org
Subject: [Air-L] Sampling size for content analysis of comments on weblogs

Hey all,

I am conducting a content analysis of comments on a weblog but am unsure of
how to determine sample size?
Has anyone run into similar issues and, if so, could they point me to a
suitable paper or two?


Cheers,

John


-- 
John McAuley
------
Knowledge & Data Engineering Group
Trinity College Dublin
Ireland
------
_______________________________________________
The Air-L at listserv.aoir.org mailing list
is provided by the Association of Internet Researchers http://aoir.org
Subscribe, change options or unsubscribe at: http://listserv.aoir.org/listinfo.cgi/air-l-aoir.org

Join the Association of Internet Researchers:
http://www.aoir.org/



More information about the Air-L mailing list