Mid Semester Work

Chapter 1

6 - Compare and contrast the fast and frugal approach to judgment with the regression analysis approach.


Reasoning is one of the important ability that both humans and animals can acquire. However, this is considered being very complex and for many years it has generated numerous studies. From the minute we wake up until the minute we go to sleep we make judgments and decisions about things, peoples and situations. Gigerenzer and colleagues (1991) suggested that humans and animals make inferences about the world under limited time and knowledge. Rationality and optimality are the guiding idea of the probabilistic approach to cognition but they are not the only reasonable guiding ideas. They suggested that there are other concepts such as simplicity and frugality, which have also inspired models of cognition. Fast and frugal models are justified by their psychological plausibility and datedness to natural environment and also are easier to understand.

Gigerenzer and colleagues (1989) also suggested that these fast and frugal violate fundamental views of classical rationality; they neither look up nor integrate all information. Even though these fast and frugal heuristics are good they are however, under constant criticism for example, Hardman (2009) suggests that these fast and frugal heuristics does not tell us what is going on in the mind of decision maker during the process of judgment.

Gigerenzer et al (1996) suggested that many studies been conducted in order to test the validity of these two views by identifying conditions under which the human mind appears more rational or irrational. However, most of the works have mainly focused in simple situations, such as Bayesian inference with binary hypotheses, one single piece of binary data, and all necessary information conveniently laid out for the participant. In many real world situations, however, there are pieces of information which are not independent. Bayes’ theorem and other rational algorithms quickly become mathematically complex and computationally intractable, at least for human minds.


A number of heuristics or short cuts are used to make fast and frugal judgments such as recognition heuristics. According to the probabilistic mental model theory inferences about the unknown states of the world are based on probability cues, such as football team, universities and so on. Gigerenzer and colleagues (1996), created a scenario whereby the participants must make a choice between two alternatives, such as on a dimension of a city. The two alternatives choice tasks occur in different contexts in which inferences need to be made with limited time and knowledge. Their study of two alternative choice tasks in situations where a person has to make an inference based only on knowledge retrieved from memory, which implies that the participants need to use their declarative knowledge to perform the task.


According to Gigerenzer el at (1996), probabilistic mental model is an inductive device that uses limited knowledge to make fast references and they perform intelligent guesses about unknown features of the world, based on uncertain indicators. To make inferences about which one of the cities has higher population (a or b), then knowledge about the reference class, for example, cities in Germany could be searched. Limited knowledge indicates that the matrix of objects by cues has missing entries i.e. objects, or cues or even cue value which might be unknown.



On the other hand we have regression analysis approach, which according to studies takes more time and are cognitively more demanding. Regression analysis usually traces the distribution of the self-governing variable (y) or some characteristic of this distribution (x). However, according to Fox there are advantages of approaching regression analysis from a general perspective. On the one hand, the appreciation for the practical problems of fitting the general models implies the specification of more limiting models such as the use of the usual linear regression model with standard errors. On the other hand, modern methods of nonparametric regression, while not quite as general as models are emerging as the practical alternatives to the more traditional linear models.



Aminta