Matthew Leitao

Analyses

The statistical methods I use, grouped by purpose.

Analyses

53 methods across 11 groups. A filled dot marks a method I have used in my work. A ring marks one I am still building depth in.

Tests and group comparisons

  • t-tests (have)
  • ANOVA and mixed ANOVA (have)
  • Chi-square tests (have)
  • Planned contrasts (have)

Regression

  • Linear regression (have)
  • Logistic regression (have)
  • Ordinal regression (have)
  • Negative-binomial regression (have)
  • Forward stepwise regression (have)
  • Mixed-effects logistic regression (have)
  • Dominance and all-subsets analysis (have)

Multilevel and longitudinal

  • Multilevel (mixed-effects) models (have)
  • Person-mean centering (have)
  • Lagged predictors and week effects (have)
  • Diary and experience-sampling analysis (have)
  • Latent growth models (developing)

Mechanisms and subgroups

  • Moderation and interactions (have)
  • Causal mediation (ACME) (have)
  • Multilevel mediation (have)
  • Mediated moderation (have)

Causal inference

  • Randomized experiments (have)
  • Propensity-score matching (have)
  • Difference-in-differences (have)
  • Regression discontinuity (have)
  • Instrumental variables (developing)

Measurement

  • Exploratory factor analysis (have)
  • Confirmatory factor analysis (have)
  • Reliability and validity testing (have)
  • Structural equation and path models (have)
  • Multi-group models (developing)
  • Latent profile analysis (developing)

Planning and robustness

  • Power and sensitivity analysis (have)
  • Specification-curve analysis (have)
  • Bonferroni and Tukey corrections (have)
  • Variance inflation checks (have)
  • Intent-to-treat analysis (have)
  • Blind outcome rating (have)

Evidence for null results

  • Bayes factors (have)
  • Bayesian t-tests (have)

Text

  • Dictionary-based text analysis (LIWC) (have)
  • Text-similarity metrics (have)
  • TF-IDF and bag of words (developing)
  • Topic models and embeddings (developing)

Prediction

  • Classification trees (developing)
  • Support vector machines (developing)
  • Nearest neighbors (developing)
  • Naive Bayes (developing)
  • Cross-validation (developing)
  • ROC curves and confusion matrices (developing)

Data integration and place

  • Pooled raw-data analysis (have)
  • Data harmonization (have)
  • Geospatial mapping (developing)
  • Cost-benefit analysis (developing)