schemaGlm

Logistic-regression association schema

Description

schemaGlm() defines the result and artifact contract for binomial logistic regression fits produced by associate$regression(model = “glm”).

Usage

schemaGlm()

Value

A schema definition list.

Result semantics

Logistic-regression rows represent coefficient-level tests on the log-odds scale.

  • outcome: binary endpoint being modeled.

  • predictor: tested predictor.

  • term: coefficient term retained for the tested predictor or requested interaction.

  • term.type: main for predictor coefficient rows or interaction for predictor-by-modifier coefficient rows.

  • effect.scale: log.odds.

  • estimate: log-odds coefficient. Forest plots exponentiate this to an odds ratio.

  • se, lower, upper: standard error and confidence interval on the log-odds scale.

  • statistic: z statistic.

  • p.value, adj.p.value: raw and adjusted coefficient p-values.

  • n, n.cases, n.controls: complete observations and binary endpoint counts used by the model.

  • formula: display formula used for the fitted model.

Artifacts

The optional prediction.grid artifact contains fitted probabilities over representative predictor values and covariate settings. It is aggregate plot support and should not contain individual-level phenotype rows.