Recapitulation I Methodological Aspects: Prediction, diagnosis and explanation. Problem of establishing causal relation- ships. Problem of hidden common causes and confounders.
Recapitulation II Example of confounding: Full moon and car accidents:
Recapitulation III Common cause (factor analytic) model :
Recapitulation IV: Causal reasoning: Method Experimental method: placebo- controlled, double blind studies with random assignment of units to treatments.
Recapitulation V: Causal reasoning: Method Popper’s Approach: Causal theories and assumptions can be tested, since causal models imply specific patterns of associations (correlations). Problem of statistically equivalent causal models
Simpson’s Paradox: Example 1 Death sentence GroupYesNo %YesYule’s Q Black White
Simpson’s Paradox: Example 1 Color Death sentence VictimDelinquent YesNo %YesYule’s Q Black White WhiteBlack White
Simpson’s paradox: Example 1: Paik Diagram
Simpson’s Paradox: Example 2 Successful LocalityTreatmentYesNo% SuccessYule’s Q Goat-townNew %0.36 Old % Cow-cityNew %0.50 Old %
Simpson’s Paradox: Example 2 Successful TreatmentYesNo% SuccessYule’s Q New %-0.30 Old %
Simpson’s paradox: Example 2: Paik Diagram
Simpson’s Paradox: Example 3 Success FieldSexYesNo % SuccessYules Q Social workMan %-0.20 Woman % PsychologyMan %-0.14 Woman %
Simpson’s Paradox: Example 3 Success SexYesNo % SuccessYules Q Man %0.47 Woman %
Simpson’s paradox: Example 3: Paik Diagrams
Ecological Fallacy: Example Treatment of clustered (grouped) data
Regression to the mean
Regression artefact: Lord’s paradox
Memory Judgments Misleading information (Loftus) Misinformation: realistic examples Piaget Ingram Cognitive Mechanism: Encoding, retrieval, and forgetting of memory information
Memory Judgments Cognitive Mechanism: Reality monitoring Cognitive Mechanism: Source monitoring
Exercises: Exercise 2-2: Simpson’s Paradox: Exercise 2-4 Exercise 2-5