confounding variable
Lurking third factor
A confounding variable is a factor in research that distorts the apparent relationship between the variables being studied.
If researchers want to know whether coffee drinking affects heart disease, age might be a confounding variable if older people both drink more coffee and have more heart disease—making coffee look guilty when age is the real culprit.
Good research design attempts to control for confounding variables through randomization or statistical adjustment.
Introducing chance systematically
- stratificationDivision into distinct levels
- matchingIdentical in appearance
- adjustmentAlteration to improve fit
- covariateVariable controlled in analysis
regression · counterbalancing · calibration · restriction · standardization
Cut through perpendicularly
- longitudinalAlong the longer dimension
- cohortGroup sharing circumstances
- exposureState of being unprotected
- risk factorCondition increasing harm probability
relative risk · incidence · prevalence · bias · selection bias
Neutral conflict resolver
- moderatorOne who presates proceedings
- interactionReciprocal action between parties
- counterfactualContrary-to-fact supposition
- colliderDual-beam particle accelerator
Imperfection in something
- oversightSupervision or inadvertent omission
- misjudgmentFaulty evaluation of something
- misinterpretationIncorrect understanding of meaning
- fallacyError in reasoning
miscalculation · misattribution · systematic error · artifact
Points sharing a line
- varianceSpread around the mean
- residualWhat remains afterward
- overfittingExcess adaptation to samples
- coefficientNumber multiplying a variable
randomization
Introducing chance systematically
- stratificationDivision into distinct levels
- matchingIdentical in appearance
- adjustmentAlteration to improve fit
covariate · regression · counterbalancing · calibration · restriction · standardization
cross-sectional
Cut through perpendicularly
- longitudinalAlong the longer dimension
- cohortGroup sharing circumstances
- exposureState of being unprotected
risk factor · relative risk · incidence · prevalence · bias · selection bias
mediator
Neutral conflict resolver
- moderatorOne who presates proceedings
- interactionReciprocal action between parties
- counterfactualContrary-to-fact supposition
flaw
Imperfection in something
- oversightSupervision or inadvertent omission
- misjudgmentFaulty evaluation of something
- misinterpretationIncorrect understanding of meaning
fallacy · miscalculation · misattribution · systematic error · artifact
collinearity
Points sharing a line
- varianceSpread around the mean
- residualWhat remains afterward
- overfittingExcess adaptation to samples
coefficient · estimate · standard error · confidence interval · p-value
- research methodsexperiment, hypothesis, deduction, experiments, induction
- variablesunknown, x, y, placeholder, alpha
- research validityvalidity, reliability, replicability, reproducibility, generalizability
- statistical fallaciesgambler's fallacy, survivorship bias, sampling bias, selection bias, law of averages
- researchstudy, research papers, researching, archive, peer review
- predictorsforecaster, prognosticator, prophet, seer, soothsayer
variable, invariably, invariable, confound, variability, confounded, variables, confounding, confounder