Heuristics & Biases (Tversky & Kahneman, 1974) Flashcards
Describe the representativeness heuristic
Mental shortcut used to make quick judgments and decisions by comparing a current situation to a prototype or stereotype in our minds. It’s a way of estimating the probability of an event or object belonging to a category based on how similar it is to a known or typical example of that category
Describe what insensitivity to prior probability of outcomes refers to
A bias resulting from the representativeness heuristic where people neglect the base-rate frequency (prior probability) of outcomes when evaluating probabilities based on representativeness. In an experiment, subjects largely ignored the differing proportions of engineers and lawyers when assessing the probability of a description belonging to either profession, focusing instead on the representativeness of the description to stereotypes
Describe worthless evidence and prior probabilities
While people correctly use prior probabilities when no specific evidence is given, they tend to ignore prior probabilities even when presented with totally uninformative evidence. For example, subjects judged the probability of Dick being an engineer as .5 regardless of the base rate of engineers, despite the description being irrelevant to the profession
Describe insensitivity to sample size
When evaluating the probability of a sample result, people using the representativeness heuristic often fail to appreciate the role of sample size. They judge the similarity of a sample statistic to a population parameter, which does not depend on sample size, leading to similar probability judgments for different sample sizes
Describe misconceptions of chance (local representativeness)
People expect a short sequence generated by a random process to represent the essential characteristics of that process locally. For example, H-T-H-T-T-H is seen as more likely than H-H-H-T-T-T or H-H-H-H-T-H. This belief leads to expecting too many alternations and too few runs in short random sequences
Describe gambler’s fallacy
A consequence of the belief in local representativeness, where people erroneously believe that after a long run of one outcome in a random process (e.g., red on a roulette wheel), the opposite outcome (black) is now “due” to make the sequence appear more representative of randomness. Chance is incorrectly viewed as a self-correcting process
Describe law of small numbers
Even experienced researchers sometimes hold a belief that small samples are highly representative of the populations from which they are drawn. This leads to overconfidence in the results of small samples and overestimation of their replicability
Describe insensitivity to predictability
When making numerical predictions based on a description, people often predict solely based on the favorableness of the description, ignoring the reliability of the description and the expected accuracy of the prediction. Predictions become insensitive to the actual predictability of the outcome
Describe the illusion of validity
The unwarranted confidence in a prediction that is produced by a good fit (high representativeness) between the predicted outcome and the input information, even if the input is scanty, unreliable, or outdated. This illusion persists even when the judge is aware of factors limiting predictive accuracy
Describe confidence in redundant input variables
People tend to have great confidence in predictions based on highly consistent or redundant input variables. However, statistically, a prediction based on several independent inputs can be more accurate than one based on redundant inputs of the same validity. Redundancy increases confidence but can decrease accuracy
Describe misconceptions of regression toward the mean
People often fail to expect regression toward the mean, a phenomenon where if one selects individuals whose average score on one measure deviates from the mean, their average score on a second, equivalent measure will usually deviate less from the mean. When regression is recognized, people tend to invent spurious causal explanations
Describe regression and reward/punishment
Failure to understand regression can lead to overestimating the effectiveness of punishment and underestimating the effectiveness of reward. Improvement often follows poor performance and deterioration often follows outstanding performance due to regression alone, which can be misattributed to the effects of punishment and reward
Describe the availability heuristic
The availability heuristic involves assessing the frequency of a class or the probability of an event by the ease with which instances or occurrences can be brought to mind. While generally useful, availability is affected by factors other than frequency and probability, leading to biases
What are some biases due to retrievability of instances?
When judging the size of a class by the availability of its instances, a class whose instances are easily retrieved will appear more numerous than a class of equal frequency whose instances are less retrievable. Familiarity and salience are factors affecting retrievability
What are some biases due to effectiveness of a search set?
The ease with which we can search for examples can influence frequency judgments. For instance, people judge words starting with a particular letter as more numerous than words having that letter in the third position, simply because it’s easier to search by the first letter
What are some biases of imaginability?
When assessing the frequency of a class whose instances are not stored in memory, people generate instances and evaluate frequency by the ease with which relevant instances can be constructed. However, ease of construction does not always reflect actual frequency, leading to biases
Describe illusory correlation
An overestimation of the frequency with which two events co-occur, often based on the strength of an associative bond between them or prior beliefs, even when the correlation is weak, nonexistent, or negative. Availability can explain this effect, as strong associates are more easily brought to mind together
Describe the adjustment and anchoring heuristic
In many situations, people make estimates by starting from an initial value (anchor) and adjusting it to yield the final answer. Adjustments are typically insufficient, meaning different starting points lead to different estimates biased toward the initial values
Describe insufficient adjustment
A key characteristic of the anchoring heuristic is that the adjustments made from the initial anchor are usually too small, resulting in final estimates that remain too close to the starting point. This occurs even when the anchor is arbitrary
What are some biases in evaluation of conjunctive events
People tend to overestimate the probability of conjunctive events (where multiple events must occur). This bias can be explained by anchoring on the probability of the individual elementary events, with insufficient downward adjustment for the need for all events to occur
What are some biases in evaluation of disjunctive events
People tend to underestimate the probability of disjunctive events (where at least one of multiple events must occur). This bias can also be explained by anchoring on the probability of the individual elementary events, with insufficient upward adjustment for the increased chance of at least one event occurring
Describe anchoring in subjective probability distributions
When constructing subjective probability distributions, people often anchor on their best estimate of the quantity and make insufficient adjustments when selecting values for different percentiles (e.g., X10, X90). This leads to overly narrow confidence intervals and a miscalibration of subjective probabilities
Describe calibration of subjective probabilities
Proper calibration means that for a set of assessed quantities, the true values should fall below the stated Xn value exactly n percent of the time. Studies show that subjective probability distributions are often poorly calibrated, with overly narrow confidence intervals, indicating overconfidence
What are some limitations of internal consistency for probability judgments
While modern decision theory emphasizes internal consistency (coherence) in subjective probability judgments, this is not the only criterion for adequate or rational judgments. Judgments must also be compatible with the individual’s broader web of beliefs and knowledge, which internal consistency alone does not guarantee
