How deeply does Statistics Slam Dunk cover Simpson's paradox?

Introduces the phenomenon briefly and demonstrates visual detection, without developing a formal proof or causal explanation.

Simpson's paradox

As explained in Statistics Slam Dunk

Simpson's paradox is a statistical phenomenon in which an association observed in aggregated data reverses or conflicts with the associations found in relevant subpopulations.

Simpson's paradox occurs when an association in aggregated data conflicts with the associations found inside relevant subpopulations. A single overall correlation can therefore conceal a different pattern within meaningful groups. section 10.4.1

Consider NBA team payrolls and regular-season wins. Across the full 2012–13 data, payroll and wins have a positive association; after separating teams by season-ending result, the relationships are flat. Coloring the groups makes the hidden structure visible. section 10.4.1

Reveal the paradox by treating aggregation as an analytical choice:

Identify relevant subpopulations, such as playoff, non-playoff, and championship outcomes. section 10.4.1
Recalculate or visualize each subgroup separately rather than trusting the pooled result. section 10.4.1
Compare the directions and strengths of the pooled and subgroup relationships before drawing a conclusion. section 10.5

Payroll, wins, and season outcomes

A pooled NBA analysis reports a positive payroll–wins correlation, while grouping teams by season-ending factor produces flat within-group relationships. The result warns against interpreting one aggregate trend as the experience of every group. section 10.4.1 section 10.5

How Statistics Slam Dunk covers it

liveBook manning.com

Read Statistics Slam Dunk on liveBook Get the book at manning.com

Also explained in

liveBook manning.com

Causal Inference for Data ScienceSimpson’s paradox is a statistical phenomenon in which an association or trend observed across aggregated data reverses or differs after conditioning on a grouping variable.

Compare how 2 books explain Simpson's paradox →

Related concepts

correlationcontrasts with · Correlation is a statistical relationship describing how two or more variables or quantities vary together, including the strength and direction of their association, without by itself implying causation.
graphused by · A graph is a mathematical structure consisting of vertices connected by edges that represent relationships or connections.
causal adjustmentcontrasts with · Causal adjustment estimates a causal effect by conditioning on variables that block noncausal pathways and averaging over their target-population distribution.
subpopulationcontrasts with · A defined subset of a broader population that shares specified characteristics or conditions.
confounderaddresses · A confounder is a variable associated with both an explanatory or treatment variable and an outcome, whose influence can distort their apparent relationship or estimated causal effect.

This explainer was generated by Manning's knowledge graph from the text of Statistics Slam Dunk. The section links point to the parts of the book it draws on.