Search

In describing how people generalize from observed samples of data to novel cases, theories of inductive inference have emphasized the learner's reliance on the contents of the sample. More recently, a growing body of literature suggests that different assumptions about how a data sample was generated can lead the learner to draw qualitatively distinct inferences on the basis of the same observations. Yet, relatively little is known about how and when these two sources of evidence are combined. Do sampling assumptions affect how the sample contents are encoded, or is any influence exerted only at the point of retrieval when a decision is to be made? We report two experiments aimed at exploring this issue. By systematically varying both the sampling cover story and whether it is given "before" or "after" the training stimuli we are able to determine whether encoding or retrieval issues drive the impact of sampling assumptions. We find that the sampling cover story affects generalization when it is presented before the training stimuli, but not after, which suggests that sampling assumptions are integrated during encoding.

An ongoing debate in the literature on human reasoning concerns whether or not the logical status (valid vs. invalid) of an argument can be intuitively detected. The finding that conclusions of logically valid inferences are liked more compared to conclusions of logically invalid ones--called the logic-liking effect--is one of the most prominent pieces of evidence in support of this notion. Trippas et al. (2016) found this logic-liking effect for different kinds of inferences, including conditional and categorical syllogisms. However, all invalid conclusions presented by Trippas et al. (2016) were also impossible given the premises and had a particular structure of surface features--that is, an incongruent atmosphere. We present new data from five preregistered experiments in which we replicate the effect reported by Trippas et al. (2016) for conditional and categorical syllogisms but show that this effect is eliminated when controlling for confounds in surface features. Moreover, we present evidence that there is a demand effect at play, which suggests that people are deliberately considering atmosphere cues of an argument to inform their liking ratings. Taken together, the findings of the present study cast doubt on the existence of logical intuitions.

The samples of evidence we use to make inferences in everyday and formal settings are often subject to selection biases. Two property induction experiments examined group and individual sensitivity to one type of selection bias: sampling frames - causal constraints that only allow certain types of instances to be sampled. Group data from both experiments indicated that people were sensitive to the effects of such frames, showing narrower generalization when sample instances were selected because they shared a target property (property sampling) than when instances were sampled because they belonged to a particular group (category sampling). Group generalization patterns conformed to the predictions of a Bayesian model of property induction that incorporates a selective sampling mechanism. In each experiment, however, there was considerable individual variation, with a nontrivial minority showing little sensitivity to sampling frames. Experiment 2 examined correlates of frames sensitivity. A composite measure of working memory capacity predicted individual sensitivity to sampling frames. These results have important implications for current debates about people's ability to factor sample selection mechanisms into their inferences and for the development of formal models of inductive inference.

Much recent research and theorizing in the field of reasoning has been concerned with intuitive sensitivity to logical validity, such as the logic-brightness effect, in which logically valid arguments are judged to have a "brighter" typeface than invalid arguments. We propose and test a novel signal competition account of this phenomenon. Our account makes two assumptions: (a) as per the demands of the logic-brightness task, people attempt to find a perceptual signal to guide brightness judgments, but (b) when the perceptual signal is hard to discern, they instead attend to cues such as argument validity. Experiment 1 tested this account by manipulating the difficulty of the perceptual contrast. When contrast discrimination was relatively difficult, we replicated the logic-brightness effect. When the discrimination was easy, the effect was eliminated. Experiment 2 manipulated the ambiguity of the perceptual task, comparing discrimination performance when the perceptual contrast was labeled in terms of rating "brightness" or "darkness". When the less ambiguous darkness labeling was used, there was no evidence of a logic-brightness effect. In both experiments, individual sensitivity to the perceptual discrimination was negatively correlated with sensitivity to argument validity. Hierarchical latent mixture modeling revealed distinct individual strategies: responses based on perceptual cues, responses based on validity or guessing. Consistent with the signal competition account, the proportion of those responding to validity increased with perceptual discrimination difficulty or task ambiguity. The results challenge explanations of the logic-brightness effect based on parallel dual-process models of reasoning.

Generalization of learning can arise from 2 distinct sources: failure to discriminate a novel test stimulus from the trained stimulus and active extrapolation from the trained stimulus to the test stimulus despite them being discriminable. We investigated these 2 processes in a predictive learning task by testing stimulus discriminability (identification of the trained stimulus) as well as generalization of learning (outcome expectancy). Generalization gradients were broader for expectancy than for identification, in both single cue and differential (discrimination) designs, implying a substantial extrapolation component for the most dissimilar stimuli. The shapes of the expectancy gradients were strongly determined by the training design (single cue vs. differential) and by the rules inferred by participants (similarity vs. linear). By contrast, the identification gradients were unaffected by the training design or inferred rules and were equivalent for predictive and nonpredictive stimuli. These results indicate that perceptual discriminability plays a substantial role in generalization, but it is largely unaffected by associative learning. Instead, learning appears to impact on generalization via an independent extrapolation component which involves cognitive processes such as inductive reasoning.

Four experiments examined the claims that people can intuitively assess the logical validity of arguments, and that qualitatively different reasoning processes drive intuitive and explicit validity assessments. In each study participants evaluated arguments varying in validity and believability using either deductive criteria (logic task) or via an intuitive, affective response (liking task). Experiment 1 found that people are sensitive to argument validity on both tasks, with valid arguments receiving higher liking as well as higher deductive ratings than invalid arguments. However, the claim that this effect is driven by logical intuitions was challenged by the finding that sensitivity to validity in both liking and logic tasks was affected in similar ways by manipulations of concurrent memory load (Experiments 1 and 2) and variations in individual working memory capacity (Experiments 3 and 4). In both tasks better discrimination between valid and invalid arguments was found when more working memory resources were available. Formal signal detection models of reasoning were tested against the experimental data using signed difference analysis (Stephens, Dunn, & Hayes, 2018b). A single-process reasoning model which assumes that argument evaluation in both logic and liking tasks involves a single latent dimension for assessing argument strength but different response criteria for each task, was found to be consistent with the data from each experiment (as were some dual-process models). The experimental and modeling results confirm that people are sensitive to argument validity in both explicit logic and affect rating tasks, but that these results can be explained by a single underlying reasoning process.

The extent to which we generalize a novel property from asample of familiar instances to novel instances depends on thesample used. In these experiments, we are interested in twosample characteristics: number of types (discrete entities) andnumber of tokens (copies of the same entity) that share a novelproperty. Existing studies permit separate and conditionalhypotheses about the effects of adding types and tokens, but nostudy has examined the effects of both variables ongeneralization stimuli varying in similarity. We find thatadding types broadens generalization to similar stimuli, buttightens generalization to dissimilar stimuli. Adding tokensdoes not affect generalization, but adding repetitions that areframed as types produces some tightening. Implications formodels of inductive reasoning are discussed.

We outline and test a Bayesian model of the effects of evidencesampling on property induction. Our model assumes thatpeople are sensitive to the effects of different sampling framesapplied to sampled evidence. Two studies tested the model bycomparing property generalization following exposure tosamples selected because they belong to the same taxonomiccategory or because they share a salient property. Both studiesfound that category-based sampling led to broadergeneralization than property-based sampling. In line withmodel predictions, these differences were attenuated when amixture of positive and negative evidence was presented(Experiment 1) and when category-property relations wereprobabilistic rather than deterministic (Experiment 2).

When asked to assess the deductive validity of an argument,people are influenced by their prior knowledge of the content.Recently, two competing explanations for this belief biaseffect have been proposed, each based on signal detectiontheory. Under a response bias explanation, people set morelenient decision criteria for believable than for unbelievablearguments. Alternatively, believable and unbelievablearguments may differ in subjective argument strength for bothvalid and invalid items. Two experiments tested theseaccounts by asking participants to assess the validity ofcategorical syllogisms and rate their confidence. Conclusion-believability was manipulated either within- or between-groups. A two-step signal detection model was applied toexamine the effects on the relative location of the decisionthreshold and the distributions of argument strength.Equivalent belief bias effects were found when believabilitywas manipulated within- and between-groups, supporting theview that the belief bias effect is due to response bias.

When asked to determine whether a syllogistic argument is deductively valid, people are influenced by their prior beliefs about the believability of the conclusion. Recently, two competing explanations for this belief bias effect have been proposed, each based on signal detection theory (SDT). Under a response bias explanation, people set more lenient decision criteria for believable than for unbelievable arguments. Under the alternative argument strength explanation, believability affects the reasoning stage of processing an argument, with believable and unbelievable arguments differing in subjective strength for both valid and invalid items. Two experiments tested these accounts by asking participants to make validity judgments for categorical syllogisms and to rate their confidence. Conclusion-believability was manipulated both within group (Experiment 1) and between groups (Experiment 2). A novel two-step version of the signal detection model was fit to receiver operating characteristic (ROC) curves for believable and unbelievable arguments. Model fits confirmed that in both experiments there was a shift in decision criterion but not argument discriminability as a function of argument believability. Crucially, when believability is manipulated between groups, this shift is expected under the response bias account but not under the argument strength account. Therefore, the results support the view that belief bias primarily reflects changes in response bias: people require less evidence to endorse a syllogism as valid when it has a believable conclusion. This has important implications for theories of deductive reasoning.

Two experiments tested whether a peak-shifted generalization gradient could be explained by the averaging of distinct gradients displayed in subgroups reporting different generalization rules. Across experiments using a causal judgment task (Experiment 1) and a fear conditioning paradigm (Experiment 2), we found a close concordance between self-reported rules and generalization gradients using a continuous stimulus dimension (hue). Both experiments also showed an overall peak-shifted gradient after differential conditioning, but not after single cue conditioning. Importantly, the peak shift could be decomposed into linear and peaked gradients when participants were divided into rule subgroups. Our results highlight the need to consider individual differences in the rules that participants derive in human generalization studies and suggest that in some situations, peak shift may be a consequence of averaging across diverse rule subgroups.

Two experiments examined theoretical accounts of whypeople fail to consider alternative hypotheses in judgmentsunder uncertainty. Experiment 1 found that a majority ofparticipants failed to spontaneously search for informationabout an alternative hypothesis, even when this requiredminimal effort. This bias was reduced when a specificalternative was mentioned before search. Experiment 2showed that when participants were given the likelihoods ofthe data given a focal hypothesis p(D|H) and an alternativehypothesis p(D|¬H), they gave estimates of p(H|D) that wereconsistent with Bayesian principles. The results show thatneglect of the alternative hypothesis typically occurs at theinitial stage of problem representation. However judgmentsare more consistent with Bayesian norms when they involveutilizing information about a given alternative

Three-experiments examined the number of qualitatively different processing dimensions needed to account for inductive and deductive reasoning. In each study, participants were presented with arguments that varied in logical validity and consistency with background knowledge (believability), and evaluated them according to deductive criteria (whether the conclusion was necessarily true given the premises) or inductive criteria (whether the conclusion was plausible given the premises). We examined factors including working memory load (Experiments 1 and 2), individual working memory capacity (Experiments 1 and 2), and decision time (Experiment 3), which according to dual-processing theories, modulate the contribution of heuristic and analytic processes to reasoning. A number of empirical dissociations were found. Argument validity affected deduction more than induction. Argument believability affected induction more than deduction. Lower working memory capacity reduced sensitivity to argument validity and increased sensitivity to argument believability, especially under induction instructions. Reduced decision time led to decreased sensitivity to argument validity. State-trace analyses of each experiment, however, found that only a single underlying dimension was required to explain patterns of inductive and deductive judgments. These results show that the dissociations, which have traditionally been seen as supporting dual-processing models of reasoning, are consistent with a single-process model that assumes a common evidentiary scale for induction and deduction.

Four experiments examined the locus of impact of causal knowledge on consideration of alternative hypotheses in judgments under uncertainty. Two possible loci were examined; overcoming neglect of the alternative when developing a representation of a judgment problem and improving utilization of statistics associated with the alternative hypothesis. In Experiment 1, participants could search for information about the various components of Bayes's rule in a diagnostic problem. A majority failed to spontaneously search for information about an alternative hypothesis, but this bias was reduced when a specific alternative hypothesis was mentioned before search. No change in search patterns was found when a generic alternative cause was mentioned. Experiments 2a and 2b broadly replicated these patterns when participants rated or made binary judgments about the relevance of each of the Bayesian components. In contrast, Experiment 3 showed that when participants were given the likelihood of the data given a focal hypothesis p(D|H) and an alternative hypothesis p(D|¬H), they gave estimates of p(H|D) that were consistent with Bayesian principles. Additional causal knowledge had relatively little impact on such judgments. These results show that causal knowledge primarily affects neglect of the alternative hypothesis at the initial stage of problem representation.

In 2 studies, the authors examined the development of the relationship between inductive reasoning and visual recognition memory. In both studies, 5- to 6-year-old children and adults were shown instances of a basic-level category (dogs) followed by a test set containing old and new category members that varied in their similarity to study items. Participants were given either recognition instructions (memorize study items and discriminate between old and new test items) or induction instructions (learn about a novel property shared by the study items and decide whether it generalizes to test items). Across both tasks, children made a greater number of positive responses than did adults. Across both age groups, a greater number of positive responses were made in induction than in recognition. The application of a mathematical model, called GEN-EX for generalization from examples, showed that both memory and reasoning data could be explained by a single exemplar-based process that assumes task and age differences in generalization gradients. These results show considerable developmental continuity in the cognitive processes that underlie memory and inductive reasoning. (Contains 2 tables, 3 figures and 2 footnotes.)

Two studies examined whether adults and children could learn to make context-dependent inferences about novel stimuli and the role of awareness of context cues in such learning. Participants were trained to match probes to targets on the basis of shape or color with the relevant dimension shifting according to item context. A selective induction test then examined context-dependent responding in a more complex matching task. Awareness of the role of context was assessed using a behavioral task and explicit questions. Experiment 1 showed that after training with the procedure described by Sloutsky and Fisher (2008), only a minority of adults showed evidence of context-dependent responding in the selective induction test. Experiment 2 used a modified training protocol that promoted attention to context cues. This led to reliable selective induction in a majority of adults and a sizeable proportion of 4- to 6-year-olds. Crucially, in both age groups, selective induction was dependent on awareness of context. Hence, children as young as 4 can learn to make selective inferences about novel stimuli, but only when they are aware of the relevant context cues. These results challenge previous claims that selective induction in children is the product of implicit learning.

Two experiments examined the impact of causal relations between features on categorization in 5- to 6-year-old children and adults. Participants learned artificial categories containing instances with causally related features and noncausal features. They then selected the most likely category member from a series of novel test pairs. Classification patterns and logistic regression were used to diagnose the presence of independent effects of "causal coherence," "causal status," and "relational centrality." Adult classification was driven primarily by coherence when causal links were deterministic (Experiment 1) but showed additional influences of causal status when links were probabilistic (Experiment 2). Children's classification was based primarily on causal coherence in both cases. There was no effect of relational centrality in either age group. These results suggest that the generative model (Rehder, 2003a) provides a good account of causal categorization in children as well as adults. (Contains 5 notes, 3 tables, and 2 figures.)

Previous research has suggested that when feature inferences have to be made about an instance whose category membership is uncertain, feature-based inductive reasoning is used to the exclusion of category-based induction. These results contrast with the observation that people can and do use category-based induction when category membership is known. The present experiments examined the conditions that drive feature-based and category-based strategies in induction under category uncertainty. Specifically, 2 experiments investigated whether reliance on feature-based inductive strategies is a product of the lack of coherence in the categories used in previous research or is due to the use of a decision-only induction procedure. Experiment 1 found that feature-based reasoning remained the preferred strategy even when categories with relatively high internal coherence were used. Experiment 2 found a shift toward category-based reasoning when participants were trained to classify category members prior to feature induction. Together, these results suggest that an appropriate conceptual representation must be formed through experience with a category before it is likely to be used as a basis for feature induction. (Contains 9 figures, 2 tables and 1 footnote.)

In an effort to assess the relations between reasoning and memory, in 8 experiments, the authors examined how well responses on an inductive reasoning task are predicted from responses on a recognition memory task for the same picture stimuli. Across several experimental manipulations, such as varying study time, presentation frequency, and the presence of stimuli from other categories, there was a high correlation between reasoning and memory responses (average r = 0.87), and these manipulations showed similar effects on the 2 tasks. The results point to common mechanisms underlying inductive reasoning and recognition memory abilities. A mathematical model, GEN-EX (generalization from examples), derived from exemplar models of categorization, is presented, which predicts both reasoning and memory responses from pairwise similarities among the stimuli, allowing for additional influences of subtyping and deterministic responding. (Contains 3 footnotes, 9 tables, and 8 figures.)

Three studies examined the development of category-based induction using an induction then recognition (ITR) procedure in which participants make category-based predictions about study items and are then given a surprise recognition test that requires discrimination between old and new category members. Exposure duration for study items was either self-paced (Experiment 1) or fixed for 5-year-olds and adults (Experiments 2a-b). Adults always showed a decrement in recognition performance following induction. Children showed the same decrement when exposure duration was equated across age groups. These results show that both young children and adults spontaneously access category-level information during induction. When study exposure time is self-paced, however, children may process additional, noncategorical aspects of study stimuli. (Contains 4 figures, 1 table, and 1 footnote.)

Four experiments examined the development of property induction on the basis of causal relations. In the first 2 studies, 5-year-olds, 8-year-olds, and adults were presented with triads in which a target instance was equally similar to 2 inductive bases but shared a causal antecedent feature with 1 of them. All 3 age groups used causal relations as a basis for property induction, although the proportion of causal inferences increased with age. Subsequent experiments pitted causal relations against featural similarity in induction. It was found that adults and 8-year-olds, but not 5-year-olds, preferred shared causal relations over strong featural similarity as a basis for induction. The implications for models of inductive reasoning and development are discussed.

V. M. Sloutsky and A. V. Fisher reported 5 experiments documenting relations among categorization, induction, recognition, and similarity in children as well as adults and proposed a new model of induction, SINC (similarity, induction, categorization). Those authors concluded that induction depends on perceptual similarity rather than conceptual knowledge. Despite the useful contributions of this work, there are some important limitations. The experimental designs examined a limited range of phenomena that are not the most revealing about the use of nonperceptual information. The main results involved a simple triad task, for which the SINC model's predictions are equivalent to the predictions of previous models of inductive reasoning. It is also unclear whether the SINC model can account for the observed relations between similarity and recognition. Implications for future work on induction and related cognitive activities are discussed.

Two experiments examined how 5- and 10-year-olds revised their category representations when exposed to exemplars that were congruent or incongruent with existing knowledge. Findings indicated that judgments about feature co-occurrent within the learned category were influenced by both stereotypical beliefs and exemplar observation. Stereotypical beliefs had a greater influence on co-occurrent judgments in the knowledge-subtyping than in standard conditions. Effects generalized to judgments about features not presented during training. (Author/KB)

Reviews empirical findings that misinformation effects in children are the product of automatic or unconscious and intentional or conscious processes. Outlines findings that show developmental change in cognitive processes underlying acceptance of misinformation in the absence of overall changes with age in the probability of reporting a suggestion. (Author)

Two experiments investigated the contribution of automatic and intentional memory processes to 5- and 8-year-olds' acceptance of misinformation either read to them or self-generated from semantic and perceptual hints. Results from recognition tests conducted under two instructional conditions suggested that both automaticity and recollection influenced misinformation acceptance, but the role of automatic processes declined with age. (Author/KB)

A study investigated effects of variations in the number of instances comprising a category on concept acquisition by 31 children (ages 9-14) with mild intellectual disability and 19 controls. Intellectual disability had little effect on ability to abstract a category prototype but did reduce use of exemplar-specific information for recognition. (Contains references.) (Author/CR)