multivariate analysis
A set of statistical techniques (and a way of thinking) for disentangling how many simultaneous factors jointly influence an outcome.
As defined in Twenty Years of Freakonomics (with Stephen Dubner)
Where it comes up · 1
Most important claims · 3 of 3Search all claims for “multivariate analysis” →
- Good, attentive, loving parenting probably does not matter nearly as much as parents wish, but bad parenting is a substantial disadvantage; who a person becomes is a complex multivariate outcome shaped by family and environment, so a child's starting conditions can heartbreakingly determine their fate.Twenty Years of Freakonomics (with Stephen Dubner)
- Economics papers, especially in better journals, follow a template that distinctively requires the author to propose and rule out alternative explanations for their thesis before defending it, which other social sciences do far less consistently, making economics resemble the hard sciences.Twenty Years of Freakonomics (with Stephen Dubner)
- The difference between good and bad journalism is that bad journalism is storytelling without data, which makes the anomaly seem normal; good storytelling must include time (over what period a change occurred), magnitude, how representative or anomalous it is, and what incentives may have changed the behavior.Twenty Years of Freakonomics (with Stephen Dubner)