Download Engineering Psychology: Human Performance and Cognitive Heuristics in Decision Making - Pr and more Assignments Psychology in PDF only on Docsity! 1 1 Engineering Psychology & Human Performance Review: – Human Memory – Training Outline of Lecture 17 – Decision Making 2 Assignment 3 Read “Double Vision” from Casey’s (1993) Set Phasers on Stun (download if necessary). Demonstrate understanding of decision making concepts in a critique of the diagnosis process – Concepts: inference (hypothesis testing), representativeness heuristic, availability heuristic, bayes theorum, prior probabilities (base rates), anchoring and adjustment heuristic, confirmation bias, absence of cues, as -if heuristic, effect of information saliency/simplicity/expensiveness, elimination by aspect heuristic 3 Decision Making Overview of Decision Making Task – operator samples incomplete information from the environment (cues are filtered passively and actively) – taken together the cues provide incomplete information about the true state of the world – operator must use inference (hypothesis testing) to determine true state of the world 4 Decision Making Overview of Decision Making Task (cont.) – Often, the operator is overwhelmed with information for making a decision memory and attentional limits are reached information triage – These limitations lead the operator to non-optimal decision making strategies heuristics: rules of thumb, cognitive shortcuts that work most of the time, but not all of the time 5 Decision Making Cognitive Heuristics – What are they? Easy or natural ways of thinking that are often useful & powerful Oversimplifications that can lead to systematic biases or misperceptions often used by experts in technical domains very difficult to overcome 6 Cognitive Heuristics – Which of the following events is most likely? That a man is under 55 and has a heart attack That a man has a heart attack That a man smokes and has a heart attack that a man is over 55 and has a heart attack 2 7 Cognitive Heuristics – Representativeness Heuristic tendency to judge an event as likely if it “represents” the typical features of its category substitution of similarity for probability Conjunction Fallacy: probability of conjunction of events cannot be greater than the probability of any single event Base Rate Fallacy: e.g., crosswalk study, medical diagnosis Misconceptions of chance – perceiving nonrandomness from randomness – Gambler’s Fallacy: belief that prior outcomes can influence the outcome of an independent probabilistic event 8 Cognitive Heuristics How many English words contain this pattern? _ n _ i n g Availability Heuristic – tendency to form judgment on the basis of what is easily brought to mind – if memory retrieval is unbiased then the heuristic works – memory retrieval is never unbiased -- its based on retrieval cues! 9 Cognitive Heuristics Availability Heuristic (cont.) – “solo” or token members of a set and availablility Von Restorff Effect: distinctive members of a set are more likely to be remembered Taylor, Fiske, Etcoff, & Ruderman (1978) – Uniqueness of race and perception of activeness, influence, and valence in interpersonal relations – unique members of a group are seen as more active and influential than others and their contributions are better remembered – Implications for perception of frequency of accidents and judgments of risk 10 Forming Inferences Representativeness & Availability and human inferential abilities – These heuristics affect the manner in which hypotheses and inferences are formed – Hypotheses are chosen based on Bottom-up processing – current data suggest or deny a hypothesis – representativeness heuristic affects this process Top-down processing – previous experience (knowledge of prior probabilities) helps determine relevant hypotheses – affected by both representativeness & availability 11 – Model of optimal inference: Bayes’ Theorum Given 2 competing hypotheses H1 and H2, what is the probability of either hypothesis being correct given the data (D); i.e., P(H1/D) and P(H2/D) Forming Inferences odds prior likelihood odds ratio Human (non-optimal) inference – Representativeness affects perception of likelihood ratio – Representativeness and availability affect perception of prior odds )/( )/( )( )( )/( )/( 2 1 2 1 2 1 HDP HDP HP HP DHP DHP ×= 12 Evaluating Inferences Which sequence produces the larger #? 8 x 7 x 6 x 5 x 4 x 3 x 2 x 1 1 x 2 x 3 x 4 x 5 x 6 x 7 x 8 Anchoring and Adjustment Heuristic 8 x 7 x 6 x 5 x 4 x 3 x 2 x 1 (median est.= 2250) 1 x 2 x 3 x 4 x 5 x 6 x 7 x 8 (median est.= 512) (both are 40320) – We tend to emphasize the first information we receive when making decisions -- it anchors us – adjustments are made upon receiving additional information, but not large enough to compensate