1. Values, errors, and precautions.
- Author
-
Needleman HL
- Subjects
- Humans, Risk Assessment, Risk Management, Sensitivity and Specificity, United States, Environmental Health, Epidemiologic Methods
- Abstract
In the environmental health literature, errors in interpreting studies or data are not infrequent. Many are of the Type II variety. Common solecisms of this type are: treating the criterion of p < 0.05 as a sacrament; demanding complete confounder control; arguing for the existence of phantom confounders; arguing that the effect size is trivial; building nonveridical models; arguing for no effect from inadequate sample size; demanding causal proof; arguing that causality is reversed; conducting a ballot of published studies. These are examined in this paper.
- Published
- 2004