Undisclosed Flexibility in Computing and Reporting Structural Equation Models in Communication Science |
| |
Authors: | Christopher Scott Seaman Rene Weber |
| |
Affiliation: | 1. University of California, Santa Barbararenew@comm.ucsb.edu;3. University of California, Santa Barbara |
| |
Abstract: | The use of structural equation modeling (SEM) in communication research has become increasingly popular in recent years, with SEM being commonly used as the primary method of data analysis in studies published within the field’s journals. However, despite its widespread use, there are persistent misapplications of SEM in published empirical studies in the communication sciences. Of these types of misuse, the most troubling are those related to the exploitation of flexibilities in data collection and analysis. Given the number of decisions that must be made when analyzing and reporting structural equation models, decisions which often remain undisclosed, the potential for researchers to exploit flexibilities in SEM research is high. To assess the extent of this problem, as well as the more longstanding types of misuse (i.e., omission of information and disclosed “bad practice”) identified by prior studies, we undertake a methodical review of studies employing structural equation modeling within major journals in the field of communication between the years 2007 and 2011. Articles were coded on the characteristics of the sample, characteristics of the model tested, as well as for various potential problems that we defined a priori. The most serious issues are identified, and recommendations for future standards in SEM research are suggested. |
| |
Keywords: | |
|
|