Model to Meaning: How to Interpret Statistical Models with R and Python

Cover....1

Half Title....2

Title Page....3

Copyright Page....4

Contents....5

Author....9

1. Who is this book for?....11

1.1. The big picture....11

1.2. Software....13

1.3. Documentation....14

1.4. Data....15

I. Interpretation....17

2. Models and meaning ....19

2.1. Why fit a model?....19

2.2. What is your estimand?....23

2.3. Making sense of parameter estimates....25

3. Conceptual framework ....27

3.1. Quantity....28

3.2. Predictors....33

3.3. Aggregation....39

3.4. Uncertainty....40

3.5. Test....41

3.6. Summary....42

II. Quantities and tests....47

4. Hypothesis and equivalence tests ....49

4.1. Null hypothesis....52

4.2. Equivalence....57

4.3. Summary....60

5. Predictions ....62

5.1. Quantity....63

5.2. Predictors....66

5.3. Aggregation....72

5.4. Uncertainty....75

5.5. Test....76

5.6. Visualization....80

5.7. Summary....86

6. Counterfactual comparisons ....89

6.1. Quantity....90

6.2. Predictors....95

6.3. Aggregation....103

6.4. Uncertainty....108

6.5. Test....108

6.6. Visualization....110

6.7. Summary....112

7. Slopes ....115

7.1. Quantity....116

7.2. Predictors....122

7.3. Aggregation....124

7.4. Uncertainty....125

7.5. Test....126

7.6. Visualization....127

7.7. Summary....129

III. Case studies....133

8. Causal inference with G-computation ....135

8.1. Treatment effects: ATE, ATT, ATU....136

8.2. Conditional treatment effects: CATE....144

9. Experiments ....146

9.1. Regression adjustment....146

9.2. Factorial experiments....148

10. Interactions and polynomials ....153

10.1. Multiplicative interactions....154

10.2. Polynomial regression....172

11. Categorical and ordinal outcomes ....177

11.1. Predictions....179

11.2. Counterfactual comparisons....183

12. Multilevel regression with poststratification ....185

12.1. Multilevel models....186

12.2. Frequentist....187

12.3. Bayesian....189

12.4. Poststratification....196

13. Machine learning ....200

13.1. tidymodels and mlr3....200

13.2. Predictions....202

13.3. Counterfactual comparisons....205

14. Uncertainty ....207

14.1. Delta method....207

14.2. Bootstrap....216

14.3. Simulation....218

14.4. Conformal prediction....220

IV. Back matter....229

Appendix I: Online content ....231

Appendix II: Python ....232

1. Who is this book for?....232

3. Conceptual framework....233

4. Hypothesis and equivalence tests....234

5. Predictions....235

6. Counterfactual comparisons....238

7. Slopes....242

8. Causal inference with G-computation....244

9. Experiments....245

10. Interactions and polynomials....246

13. Machine learning....248

Roadmap....250

Bibliography....251

Index....261

Our world is complex. To make sense of it, data analysts routinely fit sophisticated statistical or machine learning models. Interpreting the results produced by such models can be challenging, and researchers often struggle to communicate their findings to colleagues and stakeholders. Model to Meaning is a book designed to bridge that gap. It is a practical guide for anyone who needs to translate model outputs into accurate insights that are accessible to a wide audience.

Features:

• Presents a simple and powerful conceptual framework to interpret the results from a wide variety of statistical or machine learning models.
• Features in-depth case studies covering topics such as causal inference, experiments, interactions, categorical variables, multilevel regression, weighting, and machine learning.
• Includes extensive practical examples in both R and Python using the marginal effects software.
• Accompanied by comprehensive online documentation, tutorials, and bonus case studies.

Model to Meaning introduces a simple and powerful conceptual framework to help analysts describe the statistical quantities that can shed light on their research questions, estimate those quantities, and communicate the results clearly and rigorously. Based on this framework, the book proposes a consistent workflow that can be applied to (almost) any statistical or machine learning model. Readers will learn how to transform complex parameter estimates into quantities that are readily interpretable, intuitive, and understandable.

Written for data scientists, researchers, and students, the book speaks to newcomers seeking practical skills, and to experienced analysts who are ready to adopt new tools and rethink entrenched habits. It offers useful ideas, concrete workflows, powerful software, and detailed case studies, presented using real-world data and code examples.