Academic journal article Memory & Cognition

How Many Processes Underlie Category-Based Induction? Effects of Conclusion Specificity and Cognitive Ability

Academic journal article Memory & Cognition

How Many Processes Underlie Category-Based Induction? Effects of Conclusion Specificity and Cognitive Ability

Article excerpt

Two studies investigated participants' sensitivity to the amount and diversity of the evidence when reasoning inductively about categories. Both showed that participants are more sensitive to characteristics of the evidence for arguments with general rather than specific conclusions. Both showed an association between cognitive ability and sensitivity to these evidence characteristics, particularly when the conclusion category was general. These results suggest that a simple associative process may not be sufficient to capture some key phenomena of category-based induction. They also support the claim that the need to generate a superordinate category is a complicating factor in category-based reasoning and that adults' tendency to generate such categories while reasoning has been overestimated.

When we make an inductive inference we use what we already know about the world to make a prediction about some novel event or object. One frequently studied type of inductive inference exploits our tendency to group events and objects into categories. Based on our knowledge about the members of one category, we may be more or less confident that a feature known to be possessed by members ofthat category will also be possessed by members of another category. Thus, given that we know that dogs possess some property A, we might be relatively confident that wolves also possess the property but very unsure that goldfish have it There are a variety of factors known to affect the strength of an inductive inference, and a variety of accounts of the operation of these factors have emerged in the literature (for a review see Heit, 2000). In this paper I will describe two individual differences studies of people's sensitivity to some of these factors. The overall aim of these studies was to answer questions about how such sensitivity, where it exists, might best be captured by models of induction.

Category-based induction is most often studied using arguments such as 1 below. The statements above the line are the premises and the statement below the line is the conclusion. Typically, the predicate which participants are asked to project from the categories in the premises to the category in the conclusion is blank. The three characteristics of inductive arguments that I will consider here are the amount and diversity of evidence in the premises and the nature of the conclusion. Consider Arguments 1 and 2. Experimental evidence (e.g., Osherson, Smith, Wilkie, López, & Shafir, 1990; Heit & Feeney, 2005) shows that people consider arguments such as 2 to be stronger than arguments such as 1. This effect is known as the diversity effect and it occurs because people prefer arguments from diverse or dissimilar evidence.

It has also been shown that people have more confidence in conclusions that are supported by more evidence (e.g., Osherson et al., 1990; McDonald, Samuels, & Rispoli, 1996). Thus, arguments such as 3 below are perceived to be stronger than those such as 2 or 1. This effect is known as the monotonicity effect and is related to demonstrations in the literature of people's sensitivity to sample size (e.g., Nisbett, Krantz, Jepson, & Kunda, 1983). Finally, the nature of the conclusion may also have an effect on people's category-based inductive inferences. Consider Argument 4 below, which differs from Argument 2 by virtue of the greater specificity of its conclusion. Findings from adults (McDonald et al., 1996) and children (Lopez, Gelman, Gutheil, & Smith, 1992) suggest that the specificity of the conclusion of an argument affects people's inductive inferences. In particular, while Lopez et al. found no sensitivity to diversity and monotonicity in 5-year-old children, they observed sensitivity to these characteristics in 8-year-olds but only for arguments with general conclusions. More recent work (Heit & Hahn, 2001) has succeeded in demonstrating sensitivity to diversity in children as young as S. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.