Type III Error

Type III error has been defined in two primary ways. First, researchers have described a Type III error as occurring when a research study provides the right answer but for the wrong question or research hypothesis. For example, in public health research, when a research hypothesis predicts risk [Page 1580]differences between groups or across time periods, the study design requires adequate characterization of all relevant groups and time periods. If discrepancies are found among the research hypothesis, time periods, and the methods used to test the hypothesis, these discrepancies can lead to what seems to be answers to a specific question, when in fact the results support a different question. Second, other researchers have stated that in statistical tests involving directional decisions, a Type III ...