1

You are predicting a binary outcome (win = 1 / lose = 0), one observation per team, no hierarchical/repeated structure.

• ❌ Linear regression model

• ✅ Logistic regression model

• ❌ Multilevel linear regression model with random team effects

• ❌ Multilevel logistic regression model with random team effects

• ❌ Marginal linear model, fitted using GEE

• ❌ Marginal logistic model, fitted using GEE

Explanation:
Outcome is binary and observations are independent across teams → use a standard logistic regression (subject-level binary model).

---

2

Binary DV (ever had MDD), area cluster sampling, and you want to estimate between-cluster variance and explain it with cluster-level covariates.

• ❌ Linear regression model

• ❌ Logistic regression model

• ❌ Multilevel linear regression model with random cluster effects

• ✅ Multilevel logistic regression model with random cluster effects

• ❌ Marginal linear model, fitted using GEE

• ❌ Marginal logistic model, fitted using GEE

Explanation:
Binary outcome + clustering + interest in estimating between-cluster variance (and explaining it) → a multilevel (mixed-effects) logistic model with random cluster effects is appropriate.

---

3

Continuous outcome (birth weight), simple random sample (no clustering), predictors at individual level.

• ✅ Linear regression model

• ❌ Logistic regression model

• ❌ Multilevel linear regression model with random hospital effects

• ❌ Multilevel logistic regression model with random hospital effects

• ❌ Marginal linear model, fitted using GEE

• ❌ Marginal logistic model, fitted using GEE

Explanation:
Continuous outcome + simple random sample → standard linear regression (OLS) is appropriate for predicting birth weight.

---

4

Now data from many hospitals, and you want to estimate variance in expected birth weight between hospitals and explain it with hospital-level covariates.

• ❌ Linear regression model

• ❌ Logistic regression model

• ✅ Multilevel linear regression model with random hospital effects

• ❌ Multilevel logistic regression model with random hospital effects

• ❌ Marginal linear model, fitted using GEE

• ❌ Marginal logistic model, fitted using GEE

Explanation:
Interest in estimating between-hospital variance (a variance component) and explaining it → use a multilevel linear model with random hospital effects (continuous outcome).

---

5

Binary forced-choice outcome, multiple respondents per neighborhood, not interested in cluster variance but want population-averaged effect and to account for within-neighborhood correlation.

• ❌ Linear regression model

• ❌ Logistic regression model

• ❌ Multilevel linear regression model with random neighborhood effects

• ❌ Multilevel logistic regression model with random neighborhood effects

• ❌ Marginal linear model, fitted using GEE

• ✅ Marginal logistic model, fitted using GEE

Explanation:
You want the population-averaged (marginal) relationship and only need to account for within-neighborhood correlation (not estimate variance components) → use a GEE logistic model (marginal logistic model).

---

🧾 Summary Table