This Appendix provides methodological details for the analyses discussed in Chapters 6 and 7. “Meta-analysis generally refers to the statistical integration of the results of independent studies,” according to Mullen (1989, p. 1). As simple as this sounds, in the present case it involved literally hundreds of decisions about the coding of thousands of findings to produce a dataset that reflects Mullen's definition. This Appendix describes the process and the decisions.

Prior to our article in Communication Yearbook 20, no one had undertaken a systematic meta-analysis of the body of cultivation data, although several reviews had appeared along the way. Despite the widespread and growing interest in meta-analysis in the communication field, some serious difficulties were awaiting anyone who chose to undertake a meta-analysis of cultivation results. One problem is that so many individual cultivation studies have been done. Another problem has to do with the fact that many cultivation studies report many different results using different techniques of data analysis; although from a meta-analytic perspective, “the more, the merrier” (since accumulating larger samples of results helps differentiate random from systematic error), the problem is that cultivation data have been reported in many inconsistent and often incompatible forms. Moreover, unlike some other areas of research in which results are reported in fairly simple terms, the essence of cultivation work within a given study or sample could not always be easily “boiled down” to a single coefficient.