ABSTRACT
A confounder is a factor that can either cause or prevent the outcome of interest, is not on the causal pathway, but is also statistically associated with the exposure of interest (Last, 2001). Failure to detect and control for confounders can lead to, or accentuate, a statistical association between an exposure and an outcome. A confounder may exaggerate or mask a true association. For example, in this study, I have demonstrated a statistical association between the school performance and psychosis, but that association may not be causal. If the real causal association is between, say, socioeconomic group and psychosis, and if socioeconomic group is statistically associated with school performance, then school performance will show a statistical association with psychosis, and thus masquerade as a cause. In this situation, socioeconomic group would be a confounder.
