Info

r

No ^ Are they different?

No ^ Are they different?

Choose the indicated alternative

Figure 1 A flowchart of one-reason decision making. First, select a cue dimension and ascertain the corresponding cue values for each alternative; next, check whether the values for that cue discriminate between the alternatives. If so, then choose the indicated alternative; if not, select another cue dimension and repeat this process. Random choice can be used if no more cues are available.

stop cue search when a cue is found that discriminates between the two cities, and then choose the city that has the cue value 1 when the other city has cue value 0. For instance, when inferring whether New York or Los Angeles has a higher homelessness rate, the unemployment cue might be the first cue randomly selected, and the cue values are found to be 1 for both cities. Because this cue does not discriminate between the cities, search is continued, the public housing cue is randomly selected, and the cues values are 0 for New York and 1 for Los Angeles. Search is stopped at this discriminating cue and the inference is made that Los Angeles has a higher homelessness rate, as it indeed does.

The Take The Best heuristic is exactly like Minimalist except that it considers cues in order of their validity from highest to lowest. If the highest validity cue does not discriminate, the next best cue is tried, and so forth. Thus, Take The Best differs from Minimalist only in the information search rule, but it has the same stopping and decision rule. Take The Best (unlike the Minimalist, Take The Last, and recognition heuristics) is an instance of the class of lexicographic decision strategies. This term signifies that the cues are looked up in a fixed order of validity, and the first cue where choices differ is used alone to make the decision, like the alphabetic order used to arrange words in a dictionary. The Arabic number system is also lexicographic. To determine which of two numbers with equal digit length is larger, one has to start by examining the first (leftmost) digit: If this digit is larger in one of the numbers, the whole number is larger. If they are equal, one has to examine the second digit, and so on (a simple method that is not possible for Roman numbers). There is growing empirical evidence that people actually use lexicographic heuristics such as Take The Best, particularly when time is limited.

How well do these one-reason decision heuristics perform? Table I compares the performance of three fast and frugal heuristics (Minimalist, Take The Best, and Take The Last) to that of multiple regression and Dawes's and Franklin's rule. Unlike the heuristics, multiple regression is a computationally expensive linear strategy that calculates weights that reflect the covariances between predictors or cues. When the task is merely fitting the given data set, multiple regression is the most accurate strategy, by two percentage points, followed by Take The Best. However, when the task is to generalize from a training set to a test set, a simple heuristic such as Take The Best can outperform multiple regression (note that multiple regression has

Table I

Performance of Three Fast and Frugal Heuristics (Take The Best, Minimalist, and Take The Last) and Three Linear Strategies (Dawes's rule, Franklin's rule, and multiple regression) Averaged across 20 Empirical Data Sets"

Accuracy

Table I

Performance of Three Fast and Frugal Heuristics (Take The Best, Minimalist, and Take The Last) and Three Linear Strategies (Dawes's rule, Franklin's rule, and multiple regression) Averaged across 20 Empirical Data Sets"

Accuracy

Heuristic/strategy

Frugality

Fitting

Generalization

Understanding And Treating Autism

Understanding And Treating Autism

Whenever a doctor informs the parents that their child is suffering with Autism, the first & foremost question that is thrown over him is - How did it happen? How did my child get this disease? Well, there is no definite answer to what are the exact causes of Autism.

Get My Free Ebook


Post a comment