I am, therefore, not necessarily referring to qualitative researchers or experimental researchers in discussing the issue of respect for survey research. Rather, the sense of disillusionment that can also come from the perspective of active survey researchers. For example, at times, I am asked by colleagues or Ph.D. students to consult on how best to apply the latest second generation multivariate data analytic techniques (supposedly due to my background knowledge in Partial Least Square or LISREL) to make sense of some survey data they just gathered and entered into computer accessible form. My standard process involves, among other things, delving into their theoretical model (if it exists, otherwise the grand tour question), their perspective on the ontological and epistemological relationship among theory, data and the material world, and the goal or focus of their survey. Needless to say, the process can take several days if not longer. Quite frequently, the final analytical results come out meaningless or statistically insignificant. Of course, their intent was to use the latest “sophisticated” technique to bolster their survey data. My general impression for those situations is that, contrary to the cynical view, techniques such as PLS or LISREL tend to magnify or expose problems inherent in the entire survey process. Thus, one minor claim I will make right now is that surveys will likely command more respect in the future since it will tend to be more difficult to hide poorly designed research from newer, sophisticated, analytical techniques. Ideally, this should not be the case at all - that being the “defect” checking procedure near the end of the survey research process. Instead, respect should not even be the issue. Rather it is more critical to focus on how we can adopt a viewpoint towards using the survey research process to evoke meaningful understanding of the phenomena of interest.

Poor results doesn’t necessarily imply insignificant results in a statistical sense. It can mean a degree of understanding, less rich relative to qualitative approaches. But I actually believe that well done survey-based studies can provide a very deep degree understanding that may not be accomplished by other means. Further, that survey research can be used primarily in a hypothetico-deductive research mode. It can also be used in an inductive, theory exploratory sense. Only under the very beginnings where the need to understand the problem-structure or phenomenon would survey be of little value. But to accomplish a meaningful review, I’d like to highlight where the pitfalls are in the survey research process.

Before I start, it should be clear that survey research as a method represents one of several alternate methods for empirical data collection. Others techniques not considered here, for example, would be direct measurement (as in testing physiological response), observation, or secondary data research. Yet survey research is undoubtedly the best for determining, with a known level of accuracy, information about large populations. In other words, surveys should be considered if you believe and wish to uncover certain generalizations in a nomothetic sense for a population of interest. [Note that: generalization on a mere fraction of the total population (a sample) did not obtain acceptance until the beginning of the twentieth century when W. S. Gossett under the pseudonym of Student wished to determine the quality of the liquor coming out of a distillery in England.] Surveys can also be administered in a timely fashion relative to say experimental studies. They can also be quite amenable to quantification and statistical analysis. Finally, it should be clear that survey information can be collected by various means: mail-out questionnaires, telephone, in-person surveys, the Internet. I won’t talk about the advantages or disadvantages at this level. Rather, let’s get back to the entire survey research process.