Correct Answer: C. It is distributed equally in both study and control groups
A confounding factor is a variable that distorts the true association between an exposure and an outcome. The defining characteristic of a confounder is that it is unequally distributed between study and control groups, creating a spurious or masked association. By definition, confounders are NOT equally distributed—this unequal distribution is precisely what makes them confounders. If a factor were equally distributed across both groups, it would not confound the relationship because its effect would be balanced and would not bias the results. This is why matching (pairing subjects with similar confounder values) and stratification are effective control methods in epidemiological studies. In Indian public health surveillance (e.g., RNTCP studies on TB risk factors, or NFHS data analysis), confounders like socioeconomic status, nutritional status, and comorbidities are carefully assessed because they cluster differently in exposed versus unexposed populations. The false statement is that confounders are distributed equally—they are inherently unequally distributed, which is the source of confounding bias.
Why the other options are wrong
A. It is associated with the exposure of the study — This is TRUE and a core criterion for confounding. A confounder must be associated with the exposure (e.g., smoking is associated with alcohol use). Without this association, the variable cannot distort the exposure-outcome relationship. This is a defining feature, not a false statement. B. It can be reduced by matching — This is TRUE and a standard epidemiological control method. Matching pairs subjects in study and control groups on confounder values (e.g., age-matching in case-control studies), thereby reducing confounding bias. This is taught in all Indian PSM curricula and is widely used in Indian epidemiological research. D. It is associated individually with both cause and effect — This is TRUE and the second core criterion for confounding. A confounder must be independently associated with both the exposure (cause) and the outcome (effect). For example, age confounds the smoking-lung cancer relationship because age is linked to both smoking prevalence and cancer risk. This is fundamental to confounder definition.
High-Yield Facts
- Confounders are unequally distributed between exposed and unexposed groups—equal distribution would eliminate confounding bias.
- Two criteria for confounding: (1) associated with the exposure, (2) independently associated with the outcome.
- Matching in case-control studies and stratification in analysis are primary methods to control confounding in Indian epidemiological studies.
- Socioeconomic status, age, and nutritional status are common confounders in Indian public health research (NFHS, RNTCP, ICMR studies).
- Confounding bias differs from selection bias and information bias—it arises from unequal distribution of a third variable, not from study design flaws.
Mnemonics
CONFOUNDER Criteria (CAE) Causal pathway (associated with exposure), Associated with outcome independently, Equal distribution = NO confounding. If unequally distributed → confounding present. Control Methods: MAS Matching (pair on confounder value), Analysis stratification (separate by confounder level), Statistical adjustment (multivariate analysis). All reduce confounding.
NBE Trap
NBE pairs "confounding" with "equal distribution" to trap students who confuse confounding bias (which requires UNEQUAL distribution) with random error (which is equally distributed). The trap is the word "equally"—confounders by definition are NOT equally distributed.
Clinical Pearl
In Indian TB surveillance (RNTCP), socioeconomic status confounds the relationship between smoking and TB incidence because poverty is associated with both smoking prevalence and TB risk. Unequal distribution of poverty between smokers and non-smokers creates confounding—matching or stratification by socioeconomic status is essential to isolate the true smoking-TB effect.
_Reference: Park's Textbook of Preventive and Social Medicine, Ch. 8 (Epidemiology: Study Design and Bias); Robbins & Cotran Pathologic Basis of Disease, Ch. 1 (General Principles of Pathology)_