HKUST Institutional Repository >
MARK Working Papers >
Please use this identifier to cite or link to this item:
|Title: ||A method for improved learning from meta-analysis research|
|Authors: ||Vanhonacker, Wilfried R.|
Price, Lydia J.
Lehmann, Donald R.
Response surface extrapolation
|Issue Date: ||Oct-1995 |
|Series/Report no.: ||Marketing Working Paper Series ; MKTG 95.051|
|Abstract: ||Meta-analysis has been adopted in many disciplines as a methodology to take stock of large numbers of independently-executed empirical studies to see what knowledge, if any, has accumulated. Although the objectives of meta-analysis include 1) estimating average population parameters, 2) determining the effects of population characteristics and contextual factors on these parameters and 3) determining the effects of methodological factors on study estimates (Jackson 1980), this research demonstrates that most applications in marketing place disproportionate emphasis on the third objective without fully recognizing its relation to the first two. Specifically, it is shown analytically and empirically that the outcome of many meta-analyses is the identification of biases in empirical estimates of prior marketing studies. Although a significant finding in itself, such results add little additional insight into the phenomenon of interest. More importantly, potentially interesting insights are obscured because they are conditional on the methodological constraints of the prior studies. This paper addresses the problems raised by overemphasis on methodological variables in meta-analysis designs. The objective is to encourage greater emphasis on understanding the phenonmena at the center of meta-analysis studies in addition to understanding the methodological difficulties in researching them. Implications for meta-analysis design, measurement and estimation are discussed and illustrated.|
|Appears in Collections:||MARK Working Papers|
Files in This Item:
All items in this Repository are protected by copyright, with all rights reserved.