Conjoint Analysis : Data Quality Control

With 30 years of development and innovation, conjoint analysis has grown to become the foremost technique in quantifying and measuring consumers’ preferences during purchase decisions that involve multiattribute products. Researchers in conjoint analysis generally agree that one of the most significant developments occurred when commercial conjoint computer packages were introduced. Software packages bring mixed blessings to researchers and marketers. They enable widespread usage of conjoint applications in consumer behavior, while at the same time canned programs allow managers and marketers to employ the technique blindly without truly understanding the technique or making sure that the data is consistent. This often leads to poor data quality and hence less than desirable results that have huge impact on managerial decisions. The purpose of this paper is to investigate various kinds of inconsistencies that exist in conjoint survey data, in order to develop and recommend a framework for managers and marketers to identify signs of possibly inconsistent data. We then discuss some ideas we discovered that can be used to improve data quality control. Next, we illustrate the problem of poor data quality and the application of our recommendations with a detailed case study related to a wireless telecommunications service provider T-Mobile. The paper concludes with a brief discussion of some promising areas of future research in the area of data quality control in conjoint analysis.