PhD / Research · Biostatistics and Data Science

Linear and logistic regression

Regression describes relationships between predictors and outcomes under assumptions that must be checked. The choice of model depends chiefly on the nature of the outcome.

Open interactive lesson & self-check ↗

Linear and logistic regression — learning map

Original conceptual SVG · scalable
LABELED RELATIONSHIP MAP · SCHEMATIC, NOT TO SCALELinear regressioncontinuous outcomeLogistic regressionbinary outcomeConfounding and selectiondesign firstModel checksperformance and uncertaintyLinear and logistic regres…KEY RELATIONSHIPS

Original schematic relationship map for this topic; relationships are organized for study, not intended as an anatomical depiction or a diagnostic algorithm.

Concept 01

Linear regression

A typical linear model estimates the conditional mean of a continuous response as a function of predictors and an error term.

Concept 02

Logistic regression

A logistic model relates predictors to the log-odds of a binary event and can produce adjusted odds ratios.

Concept 03

Confounding and selection

Including a variable in a model does not automatically remove confounding; causal structure, measurement quality and selection affect interpretation.

Concept 04

Model checks

Inspect residual patterns where relevant, linearity assumptions, influential observations, calibration, missingness and uncertainty in estimates.

Education / safety note: An association in a fitted model is not automatically a causal effect or an individualized clinical prediction.

Reference and next reading

Explore the original curriculum and publisher-hosted resources for full-depth reading; this note is an original schematic introduction, not an exhaustive chapter.

Official / publisher source: PubMed ↗Official / publisher source: PubMed ↗Official / publisher source: PubMed ↗