Excellent points, especially the tenet that we should be testing ideas, in my mind, hypotheses. To expand on your last argument a bit, it seems to me that these new techniques, while fundamentally Baysean in nature, are poorly enough understood that they invite skepticism. Some even use proprietary algorithms making it impossible to know the certain meaning of results. Still, I can't help thinking progress in studies using content analysis would accelerate by their adoption where the bottleneck is what to do with the volumes of raw data acquired as 'scrubbed' content from online sources. Sticking with manual coding means that only linear growth is possible. -----Original Message----- From: air-l-aoir.org-bounces@listserv.aoir.org [mailto:air-l-aoir.org-bounces@listserv.aoir.org] On Behalf Of Ellis Godard Sent: Tuesday, June 07, 2005 3:22 PM To: air-l@listserv.aoir.org Subject: RE: [Air-l] Technical competence Just because some folks are applying advanced techniques to data, doesn't mean that anyone (much less everyone) else needs to understand those techniques. Frequently, methods employed in the social sciences surpass the theoretical maturity available. Perhaps I'm archaic to think that techniques should test ideas, rather than generate them. But even factor analysis and stepwise regression give me pause - not because I lack the technical competence, but because sampling deviations may generate findings that won't hold beyond the available sample. -eg
-----Original Message----- From: air-l-aoir.org-bounces@listserv.aoir.org [mailto:air-l-aoir.org-bounces@listserv.aoir.org] On Behalf Of Cox Sent: Tuesday, June 07, 2005 4:23 AM To: air-l@listserv.aoir.org Subject: RE: [Air-l] Technical competence
The need for knowing about computer technologies in communications research is becoming greater than the rudiments of web composition and traffic analysis. Already, artificial intelligence is being applied to content analysis, as in the case of a number of papers published on the Enron email corpus. The skill sets involved fall outside those typically found among communications researchers. A principle researcher in one of these Enron studies is Andrew McCallum at UMass, who is a physicist iirc. Another physicist, Andrew Smith, is responsible for the Leximancer tool mentioned earlier by Thomas Koenig. Less abstract tools like structural equation modeling are common now, and require competence in computer technologies beyond SPSS.
Whether these technologies should be incorporated in curricula is maybe not the right question, as they are not the types of skills one gets in a course or two. Perhaps the field should recruit from among information science and computer science undergrads who come equipped with the skills already.
_______________________________________________ The Air-l-aoir.org@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/