Quality issues with survey research

This paper identifies some of the many decisionmaking stages in the analysis of survey data and some of the ways in which results are affected by analysis decisions. It argues for greater transparency in reporting the details of both the data and the methods of analysis used. This not only enables replication but also allows a much more critical appraisal of the research than is usually possible. This is particularly important where research makes claims about causal processes that may influence policy decisions. Achieving this requires a greater willingness by some journal editors to allow details of data collection and analysis to be fully reported, for example by greater use of associated websites. Appendix 1 lists a proposed set of reporting criteria that could be adopted by editors.