Meta-analysis and science

Journal/Book: Theor Psychol. 1996; 6: 2455 Teller Rd, Thousand Oaks, CA 91320. Sage Publ Inc. 229-246.

Abstract: Meta-analytic writers have argued that psychological science is not working and that the solution to the problem is meta-analysis (MA), quantitative summaries of the research literature. This paper assesses these claims by evaluating MA as a means of scientific discovery and by considering the nature of effective science. It is shown that MA should not be credited as a means of discovery because its database, the research literature, is an untrustworthy source of information about nature. The difference between scientific activity, on the one hand, and literature study, on the other, is highlighted in the analyses. Both MA and empirical science are said to be misguided by a hypothesis-testing approach to science, which approach leads to an obliviousness with respect to mechanisms that underlie phenomena.

Note: Article D Sohn, Univ N Carolina, Dept Psychol, Charlotte, NC 28223 USA

Keyword(s): PUBLICATION BIAS; METAANALYSIS

Zurück | Weiter