COM 633 Content Analysis Sampling and Units Kimberly
COM 633: Content Analysis Sampling and Units Kimberly A. Neuendorf, Ph. D. Cleveland State University Fall 2010
Defining the Population Researcher fiat!? Let’s discuss. “The set of units to which the researcher wishes to generalize. ” Is a full sampling frame available? (e. g. , population may be defined by a NEXIS search)
Population & Sample Population Sample The study of its units Census Survey, experiment, or content analysis The number of units N n A summary number about a variable and its distribution Parameter Statistic The mean of a variable µ M or X-bar The standard deviation of a variable σ sd The variance of a variable σ2 sd 2
Units Weyls (2001); study of adult entertainment Unit of sampling Unit of data collection Unit of analysis Lombard et al. (1996); study of TV production techniques
Units Weyls (2001); study of adult entertainment Unit of sampling News story Unit of data collection News story Unit of analysis Year Lombard et al. (1996); study of TV production techniques
Units Weyls (2001); study of adult entertainment Lombard et al. (1996); study of TV production techniques Unit of sampling News story Time, date, channel Unit of data collection News story Episode, time interval, etc. Unit of analysis Year Episode, time interval, etc.
Units National Cancer Janstova Institutes (2003); (2006); study of films study of news about mammography Unit of sampling Unit of data collection Unit of analysis by Jane Campion vs. others
Units National Cancer Janstova Institutes (2003); (2006); study of films study of news about mammography Unit of sampling News story Unit of data collection Story, attribution Unit of analysis News story by Jane Campion vs. others
Units National Cancer Janstova Institutes (2003); (2006); study of films study of news about mammography by Jane Campion vs. others Unit of sampling News story Film Unit of data collection Story, attribution Film, character, 5 -min. interval Unit of analysis News story Film, character, 5 -min. interval
Unitizing Unit—one decision rule might be to select the “smallest identifiable unit for which one can reliably code for the desired variables, thus providing maximal variance for these measures across the entire unit of sampling” (p. 72) Coder unitizing vs. a priori unitizing Unitizing reliability is necessary when coders unitize
Archives Unprecedented access to “archives” online! Comprehensive vs. Selective Electronic/online vs. non Free vs. costly
Medium Management Archiving messages Searching for messages Preparation of messages for coding/handling during coding Automated content analysis
Medium Management A changing area. . . More later in the term
Sampling Random or Probability Sampling Simple random sampling (SRS) Systematic sampling ----------------Cluster sampling Stratified sampling Multistage sampling
Sampling Non-probability (nonrandom) Sampling Convenience sampling Purposive or judgment sampling Quota sampling
Sample Size Rules of thumb At least 100 A “rich range” on the variables of interest Can use desired confidence interval to set sample size (but many formulae— Riffe, Lacy and colleagues have provided some examples)
- Slides: 16