2009 Annual SAPOR Conference
Short Course Offering
Practical Content Analysis for Survey Research
By: Professor Mark D. West, Ph.D.
As individuals have greater access to the Internet, survey researchers have ready access to huge ad hoc databases of textual data concerning consumer preferences, voting intentions, and other sorts of subjective perceptions of the world. The techniques for gathering and analyzing this data, however, are unfamiliar.
The initial analysis of textual data -- computer content analysis --requires the use of new tools, including both 'off the shelf' software packages and 'hand-coded' programs using a new generation of textually-oriented programming languages. The gathering of this data requires substantive knowledge of how to 'scrape' web sites, and of the implications such techniques have for sampling and weighting schemes. Further, the statistical analysis of the data gathered using such methods involves familiarity with distributions such as the Yule and Poisson distributions.
This course will deal with the basic definitions and concepts of content analysis, and with methods of gathering data, the initial analysis of the textual materials -- the conversion of the raw text into numerical data, and analysis of data employing a variety of techniques.
The course will be centered on a project involving the analysis of the Presidential radio addresses of George W. Bush, the changes in the verbal tone of those addresses over time, and the statistical relationship of those changes to 'real-world' events. This course will also look at perceptions of health care providers as described in client 'blogs' for North Carolina hospitals.
|
