Glossary

APPENDIX · GLOSSARY

Glossary

Plain-English definitions of the terms this curriculum uses the most.

Each entry links to the first lab in which the term appears.

A

α (alpha). The probability of a type I error — rejecting a true null hypothesis. Conventional values are 0.05 and 0.01. Course 1 W3 S5.

ANCOVA. Analysis of covariance: a linear model combining categorical predictors and continuous covariates, commonly used to adjust RCT analyses for baseline. Course 2 W3 S2.

ANOVA. Analysis of variance: a linear model with categorical predictors. Course 2 W2 S1.

Assumption. A condition the chosen test needs to be interpretable — normality, equal variance, independence, linearity. Course 1 W4 S1.

B

Bias. Systematic deviation of an estimator from the quantity being estimated. Course 1 W3 S1.

Bootstrap. Resampling with replacement to approximate the sampling distribution of a statistic. Course 1 W3 S2.

Brier score. Mean squared error between predicted probabilities and observed outcomes. Course 2 W3 S5.

C

CI (confidence interval). A range of plausible values for a parameter; in repeated sampling, 95% CIs cover the true value 95% of the time. Course 1 W3 S2.

Cohen’s d. Standardised difference in means; a common effect size for two-group comparisons. Course 1 W4 S1.

Competing risk. An event whose occurrence precludes or alters the probability of the event of interest. Course 3 W3 S2.

Confounding. A third variable distorting the association between exposure and outcome. Course 2 W1 S3.

Cox model. A proportional-hazards regression for time-to-event outcomes. Course 2 W4 S3.

CV (cross-validation). Splitting data into folds to estimate generalisation error. Course 4 W1 S1.

D

DAG (directed acyclic graph). A graphical representation of causal assumptions. Course 3 W3 S3.

E

Effect size. A standardised measure of the magnitude of an effect, independent of sample size. Course 1 W4 S1.

F

FDR (false discovery rate). The expected proportion of false positives among rejected nulls. Course 4 W4 S4.

Fisher’s exact test. Exact test of independence for 2×2 tables, appropriate when expected cell counts are small. Course 1 W4 S2.

G

GAM (generalised additive model). A regression with smooth non-linear terms, fitted via penalised splines. Course 2 W2 S4.

GEE (generalised estimating equations). A marginal-model approach for clustered or repeated data. Course 3 W2 S4.

GLM (generalised linear model). A regression with a link function and exponential-family error. Course 2 W3.

H

Hazard ratio. Ratio of hazard rates between two groups in a survival model. Course 2 W4 S3.

I

ICC (intraclass correlation). Proportion of variance attributable to clustering. Course 2 W4 S2.

Interaction. The effect of one predictor depends on the value of another. Course 2 W1 S3.

IPTW (inverse-probability-of-treatment weighting). Propensity-score method that reweights observations to emulate a trial. Course 3 W3 S4.

K

Kaplan-Meier. Non-parametric estimator of the survival function. Course 2 W4 S3.

Kruskal-Wallis. Non-parametric one-way ANOVA on ranks. Course 1 W4 S4.

L

Lasso. L1-regularised regression; produces sparse coefficient estimates. Course 4 W1 S2.

Likelihood. The probability of the observed data under a model, viewed as a function of the parameters. Course 1 W3 S3.

Linear model. A model of the form \(y = X\beta + \varepsilon\). Course 2 W1 S2.

Logistic regression. GLM with a logit link for binary outcomes. Course 2 W3 S1.

M

MAR (missing at random). Missingness depends on observed variables only. Course 3 W2 S1.

MCAR (missing completely at random). Missingness is independent of all variables. Course 3 W2 S1.

MCMC. Markov-chain Monte Carlo; the Bayesian posterior-sampling workhorse. Course 4 W3 S2.

MDE (minimum detectable effect). Smallest effect your study has power to detect. Course 3 W1 S5.

Meta-analysis. Combining effect estimates across studies. Course 3 W4 S2.

Mixed model. Regression combining fixed and random effects. Course 3 W2 S3.

MNAR (missing not at random). Missingness depends on unobserved values. Course 3 W2 S1.

Multiple imputation. Imputing missing values several times and pooling the analyses. Course 3 W2 S2.

N

Non-parametric. A test or estimator that makes weak distributional assumptions. Course 1 W4 S4.

O

Odds ratio. Ratio of odds between two groups; the natural scale for logistic regression. Course 2 W3 S1.

Outlier. An observation far from the bulk of the data. Course 2 W1 S4.

Overdispersion. Variance exceeding the model’s nominal variance; common in Poisson regression. Course 2 W3 S4.

P

p-value. Probability of data as or more extreme than observed, assuming the null. Course 1 W3 S5.

Paired test. A test comparing matched observations rather than independent samples. Course 1 W4 S1.

PCA (principal components analysis). Linear dimension reduction by orthogonal projection onto directions of maximum variance. Course 4 W1 S3.

Permutation test. Inference by shuffling labels to build a null distribution. Course 1 W3 S2.

Poisson regression. GLM with a log link for counts. Course 2 W3 S4.

Power. Probability of detecting an effect if it exists; 1 − β. Course 1 W4 S5.

Pre-registration. A timestamped record of the research plan before data analysis. Course 3 W4 S5.

Propensity score. Probability of treatment given covariates. Course 3 W3 S4.

Pseudoreplication. Treating correlated observations as independent replicates. Course 3 W1 S4.

R

Random effect. A model coefficient treated as a draw from a distribution. Course 3 W2 S3.

Randomisation. Allocating units to arms by a chance mechanism. Course 3 W1 S2.

Regression to the mean. Tendency of extreme values to be closer to the mean on remeasurement. Course 2 W4 S1.

Reliability. Consistency of repeated measurements. Course 2 W4 S2.

Resampling. Bootstrap, permutation, and cross-validation collectively. Course 1 W3 S2.

Residual. Observed minus predicted. Course 2 W1 S4.

Risk ratio. Ratio of event probabilities between two groups. Course 1 W4 S2.

Robust SE. A standard error computed without assuming homoscedasticity. Course 2 W1 S5.

ROC / AUC. Receiver operating characteristic; area under it measures discrimination. Course 2 W3 S5.

S

SAP (statistical analysis plan). The formal plan for a trial’s analysis, written before data are seen. Course 3 W4 S5.

SE (standard error). Standard deviation of a sampling distribution. Course 1 W3 S1.

SEM (standard error of the mean). Standard error of a sample mean. Course 1 W3 S1.

Shapiro-Wilk. A test of normality, powerful in small samples. Course 1 W4 S1.

Spearman correlation. Rank-based correlation. Course 1 W4 S3.

Survival. Time-to-event analysis. Course 2 W4 S3.

T

Type I / II error. False-positive and false-negative errors of a test. Course 1 W3 S5.

V

VIF (variance inflation factor). Measure of collinearity among regression predictors. Course 2 W1 S4.

W

Wilcoxon test. Non-parametric test for paired (signed-rank) or unpaired (rank-sum) data. Course 1 W4 S4.