Single-variable blind raters guarded coding against hypothesis contamination

Evidence

All of the interviews were coded by two judges who were thoroughly familiar with the hypotheses and findings of the total research. In addition, however, there were seven control raters who were completely unacquainted with the research as a whole. Each control rater coded all interviews for just one variable. In the content analyses of the major interview study each rater coded all variables and that means that knowledge of one could easily have biased the coding of another. The study of psychiatric interviews was free of this flaw. — John T. Jost and Jim Sidanius, Political Psychology, 523