The Misuse of Statistical Techniques in Evaluating Sampling Data
The article points out that a large proportion of primary research being carried out in at least one functional area of business may be suspect because of the casual procedures being used to gather data. Specifically, many researchers in their attempts to facilitate the problems and costs of amassing data are abandoning procedures that guarantee randomness of the elements being selected for investigation. The article reviews the more typical types of errors being made, and concludes with an example of how such procedural errors can lead to large statistical biases and unjustified conclusions.
Gentry, Dwight L., and John Hoftyzer. "The misuse of statistical techniques in evaluating sampling data." Journal of the Academy of Marketing Science 5, no. 1-2 (1977): 106-112.