The robust beauty of improper linear models in decision making.
1979/07/01 by Robyn M. Dawes · 1,996 citations
Decision Sciences · Mathematics · Medicine · Psychology · #Computer science #Econometrics #General linear model #Intuition #Linear discriminant analysis #Linear model #Linear regression #Mathematics #Medicine #Psychology #Regression analysis #Reliability and Agreement in Measurement #Statistical Methods in Clinical Trials #Statistics #Statistics Education and Methodologies #Weighting
paper · doi:10.1037/0003-066x.34.7.571
published in American Psychologist 34(7), 571-582 (American Psychological Association)
openalex publication_date 1979/07/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/04
Abstract
ABSTRACT: Proper linear models are those in which predictor variables are given weights in such a way that the resulting linear composite optimally predicts some criterion of interest; examples of proper linear models are standard regression analysis, discriminant function analysis, and ridge regression analysis. Research summarized in Paul Meehl's book on clinical versus statistical prediction—and a plethora of research stimulated in part by that book—all indicates that when a numerical criterion variable (e.g., graduate grade point average) is to be predicted from numerical predictor variables, proper linear models outperform clinical intuition. Improper linear models are those in which the weights of the predictor variables are obtained by some nonoptimal method; for example, they may be obtained on the basis of intuition, derived from simulating a clinical judge's predictions, or set to be equal. This article presents evidence that even such improper linear models are superior to clinical intuition when predicting a numerical criterion from numerical predictors. In fact, unit (i.e., equal) weighting is quite robust for making such predictions. The article discusses, in some detail, the application of unit weights to decide what bullet the Denver Police Department should use. Finally, the article considers commonly raised technical, psychological, and ethical resistances to using linear models to make important social decisions and presents arguments that could weaken these resistances.
Cited by
- People expect artificial moral advisors to be more utilitarian and distrust utilitarian moral advisors
- The hot hand in basketball: On the misperception of random sequences
- Artificial Intelligence in Utilitarian vs. Hedonic Contexts: The “Word-of-Machine” Effect
- Overcoming Algorithm Aversion: People Will Use Imperfect Algorithms If They Can (Even Slightly) Modify Them
- On the Selection of Item Scores or Composite Scores for Clinical Prediction
- Heuristic Decision Making
- Evolutionary Personality Psychology
- Resistance to Medical Artificial Intelligence
- Constructive Consumer Choice Processes
- Scalable Bayesian Rule Lists
- Models, Markets, and the Forecasting of Elections
- Factors influencing the use of a decision rule in a probabilistic task
- Signed, Sealed, Counted? An Experimental Study of Mail-in Ballot Signature Verification
- The Paradox of Algorithms and Blame on Public Decision-makers
- Understanding People’s Preferences for Predictions: People Prioritize Being Right over Minimizing How Wrong They Are in Expectation
- Interpretable Classification Models for Recidivism Prediction
- When Algorithms Err : Differential Impact of Early vs. Late Errors on Users’ Reliance on Algorithms
- The Pick-the-Winner-Picker Heuristic: Preference for Categorically Correct Forecasts
- Individual differences in reasoning: Implications for the rationality debate?
- Psychological Entitlement: Interpersonal Consequences and Validation of a Self-Report Measure
- Conceptual Complexity and Construct Validity
- A systematic review of algorithm aversion in augmented decision making
- Choosing between human and algorithmic advisors: The role of responsibility sharing
- AI can help humans find common ground in democratic deliberation
- Decision authority and the returns to algorithms
- Perceived Fairness of Human Managers Compared with Artificial Intelligence in Employee Performance Evaluation
- Generating Conditional Probabilities for Bayesian Networks: Easing the Knowledge Acquisition Problem
- Simple Rules to Guide Expert Classifications
- Algorithm aversion: People erroneously avoid algorithms after seeing them err.
- Argumentation as a General Framework for Uncertain Reasoning
- Algorithm appreciation: People prefer algorithmic to human judgment
- Making sense of recommendations
- The role of trust and algorithms in consumers’ front-of-pack labels acceptance: a cross-country investigation
- When do you trust AI? The effect of number presentation detail on consumer trust and acceptance of AI recommendations
- Reimagining work in the age of AI: a multilevel review of individual, organizational, and societal dynamics
- The description–experience gap in risky choice: the role of sample size and experienced probabilities
- Learning Optimized Or's of And's
- The median split: Robust, refined, and revived
- An analysis of machine learning approaches for enhancing decision-making in complex discrete choice tasks
- Why Heuristics Work
- From Atoms to Bits and Back: A Research Curation on Digital Technology and Agenda for Future Research
- The Usefulness of Unit Weights in Creating Composite Scores
- In Top-Down Decisions, Weighting Variables does Not Matter: A Consequence of Wilks' Theorem
- Methodology Review: Estimation of Population Validity and Cross-Validity, and the Use of Equal Weights in Prediction
- CART analysis with unit-weighted regression to predict suicidal ideation from Big Five traits
- On the sensitivity of regression and regressors.
- Things I have learned (so far).
- Understanding Misunderstanding: Social Psychological Perspectives
- Inside the Planning Fallacy: The Causes and Consequences of Optimistic Time Predictions
- Algorithmic Decision Making in the Presence of Unmeasured Confounding
- Seven deadly sins of contemporary quantitative political analysis
- International reputation and alliance portfolios: How unreliability affects the structure and composition of alliance treaties
- Extensional versus intuitive reasoning: The conjunction fallacy in probability judgment.
- Reasoning the fast and frugal way: Models of bounded rationality.
- Comparative efficiency of informal (subjective, impressionistic) and formal (mechanical, algorithmic) prediction procedures: The clinical–statistical controversy.
- Support theory: A nonextensional representation of subjective probability.
- Like Goes with Like: The Role of Representativeness in Erroneous and Pseudo-Scientific Beliefs
- Heuristics made easy: An effort-reduction framework.
- Situation Awareness Is Adaptive, Externally Directed Consciousness
- Clinical Versus Actuarial Judgment
- Heuristics and Biases
- Optimal incentives for collective intelligence
- When fairness works: Toward a general theory of uncertainty management
- When Predictions Fail: The Dilemma of Unrealistic Optimism
- Ambiguity and self-evaluation: The role of idiosyncratic trait definitions in self-serving assessments of ability.
- A response-time approach to comparing generalized rational and take-the-best models of decision making.
- Simple heuristics and rules of thumb: Where psychologists and behavioural biologists might meet
- Sympathetic Magical Thinking: The Contagion and Similarity “Heuristics”
- Predicting Wimbledon 2005 tennis results by mere player name recognition
- Expert intuitions: How to model the decision strategies of airport customs officers?
- Risk and Rationality in Adolescent Decision Making
- Development of a Scale to Identify the Fall-Prone Patient
- Robust regression [wikipedia]
- Unit-weighted regression [wikipedia]
- Ecological rationality: a framework for understanding and aiding the aging decision maker. [europepmc]
- The current and future status of the concealed information test for field use. [europepmc]
- Controlled information integration and bayesian inference. [europepmc]
- Prospective cohort study of the relationship between neuro-cognition, social cognition and violence in forensic patients with schizophrenia and schizoaffective disorder. [europepmc]
- Predicting Performance in Higher Education Using Proximal Predictors. [europepmc]
- Neural Underpinnings of Decision Strategy Selection: A Review and a Theoretical Model. [europepmc]
- Fast or frugal, but not both: Decision heuristics under time pressure. [europepmc]
- Choosing Prediction Over Explanation in Psychology: Lessons From Machine Learning. [europepmc]
- Local Choices: Rationality and the Contextuality of Decision-Making. [europepmc]
- Prediction of Violence, Suicide Behaviors and Suicide Ideation in a Sample of Institutionalized Offenders With Schizophrenia and Other Psychosis. [europepmc]
- Early-Life Adversity and Dysregulation of Adult Diurnal Cortisol Rhythm. [europepmc]
- A Potential Bias in Subjective Ratings of Mental Effort. [europepmc]
- Prefrontal networks dynamically related to recovery from major depressive disorder: a longitudinal pharmacological fMRI study. [europepmc]
- Leeds-Genoa Non-Union Index: a clinical tool for asessing the need for early intervention after long bone fracture fixation. [europepmc]
- Predicting the replicability of social science lab experiments. [europepmc]
- A Tutorial on Mechanical Decision-Making for Personnel and Educational Selection. [europepmc]
- The Treatment-induced Neuropathy Assessment Scale (TNAS): a psychometric update following qualitative enrichment. [europepmc]
- A framework for building cognitive process models. [europepmc]
- Integrating social and cognitive aspects of belief dynamics: towards a unifying framework. [europepmc]
- Humans rely more on algorithms than social influence as a task becomes more difficult. [europepmc]
- Evaluating the Effectiveness of Personalized Medicine With Software. [europepmc]
- Fidelity and the impact of patient safety huddles on teamwork and safety culture: an evaluation of the Huddle Up for Safer Healthcare (HUSH) project. [europepmc]
- Exploring Hepatocellular Carcinoma Mortality Using Weighted Regression Estimation for the Cohort Effect in Taiwan from 1976 to 2015. [europepmc]
- Taking the Shortcut: Simplifying Heuristics in Discrete Choice Experiments. [europepmc]
- Incomparability and Incommensurability in Choice: No Common Currency of Value? [europepmc]
Related