Belief-based and analytic processing in transitive inference: Further evidence for the importance of premise integration

Transitive inference problems were presented in a modified conclusion evaluation procedure. Conclusions were believable or unbelievable and valid, invalid or indeterminate. The 67 undergraduate students read the premises, evaluated the conclusions (accept, reject, cannot tell), then provided confidence ratings. Fluid intelligence was also assessed. Acceptance of valid conclusions, rejection of invalid conclusions and cannot tell responses to indeterminate conclusions of non-integrable problems indicated use of analytic processing. Believability effects indicated the use of heuristic processing. Fluid intelligence and premise integration ability (non-integrable problems) predicted greater use of analytic processing on valid and invalid problems. Premise integration ability was associated with reduced belief bias on invalid problems. Premise integration ability appears to influence the extent of heuristic versus analytic processing. Confidence was sensitive to the presence of belief–logic conflict. Conflict detection scores reflecting this sensitivity were not associated with analytic processing suggesting that conflict detection occurs automatically and reflects an intuitive logic.     

Causal Inference from Descriptions of Experimental and Non-Experimental Research: Public Understanding of Correlation-Versus-Causation

The human tendency to conflate correlation with causation has been lamented by various scientists (Kida, 2006; Stanovich, 2009), and vivid examples of it can be found in both the media and peer-reviewed literature. However, there is little systematic data on the extent to which individuals conflate correlation with causation. In three experiments, we presented people with one of four research vignettes generated from the combination of two independent variables: whether the vignette described an experimental or non-experimental design, and whether it revealed a positive or negative association. Upon reading their vignette, participants selected inferences that could be drawn from the findings. Participants drew causal inferences from non-experimental vignettes as often as they did from experimental vignettes, and more frequently for causal statements and directions of association that fit with intuitive notions than for those that did not. We discuss our findings in relation to other biases in human thinking.     

Preverbal infants identify emotional reactions that are incongruent with goal outcomes

Identifying the goal of another agent’s action allows an observer to make inferences not only about the outcomes the agent will pursue in the future and the means to be deployed in a given context, but also about the emotional consequences of goal-related outcomes. While numerous studies have characterized the former abilities in infancy, expectations about emotions have gone relatively unexplored. Using a violation of expectation paradigm, we present infants with an agent who attains or fails to attain a demonstrated goal, and reacts with positive or negative affect. Across several studies, we find that infants’ attention to a given emotional display differs depending on whether that reaction is congruent with the preceding goal outcome. Specifically, infants look longer at a negative emotional display when it follows a completed goal compared to when it follows a failed goal. The present results suggest that infants’ goal representations support expectations not only about future actions but also about emotional reactions, and that infants in the first year of life can relate different emotional reactions to conditions that elicit them.     

The pseudodiagnosticity trap: Should participants consider alternative hypotheses?

The pseudodiagnosticity task has been used as an example of the tendency on the part of participants to incorrectly assess Bayesian constraints in assessing data, and as a failure to consider alternative hypotheses in a probabilistic inference task. In the task, participants are given one value, the anchor value, corresponding to P(D1|H) and may choose one other value, either P(D1|¬!H), P(D2|H), or P(D2|not;!H). Most participants select P(D2|H), or P(D2|¬!H) which have been considered inappropriate (and called pseudodiagnostic) because only P(D1|¬!H) allows use of Bayes' theorem. We present a new analysis based on probability intervals and show that selection of either P(D2|H), or P(D2|¬!H) is in fact pseudodiagnostic, whereas choice of P(D1|¬!H) is diagnostic. Our analysis shows that choice of the pseudodiagnostic values actually increases uncertainty regarding the posterior probability of H, supporting the original interpretation of the experimental findings on the pseudodiagnosticity task. The argument illuminates the general proposition that evolutionarily adaptive heuristics for Bayesian inference can be misled in some task situations.     

Intelligence and negation biases on the Conditional Inference Task: A dual-processes analysis

We examined a large set of conditional inference data compiled from several previous studies and asked three questions: How is normative performance related to intelligence? Does negative conclusion bias stem from Type 1 or Type 2 processing? Does implicit negation bias stem from Type 1 or Type 2 processing? Our analysis demonstrated that rejecting denial of the antecedent and affirmation of the consequent inferences was positively correlated with intelligence, while endorsing modus tollens inferences was not; that the occurrence of negative conclusion bias was related to the extent of Type 2 processing; and that the occurrence of implicit negation bias was not related to the extent of Type 2 processing. We conclude that negative conclusion bias is, at least in part, a product of Type 2 processing, while implicit negation bias is not.     

Some but not all dispreferred turn markers help to interpret scalar terms in polite contexts

In polite contexts, people find it difficult to perceive whether they can derive scalar inferences from what others say (e.g., does “some people hated your idea” mean that not everyone hated it?). Because this uncertainty can lead to costly misunderstandings, it is important to identify the cues people can rely on to solve their interpretative problem. In this article, we consider two such cues: Making a long Pause before the statement, and prefacing the statement with Well. Data from eight experiments show that Pauses are more effective than Wells as cues to scalar inferences in polite contexts—because they appear to give a specific signal to switch expectations in the direction of bad news, whereas Well appears to give a generic signal to make extra processing effort. We consider the applied value of these findings for human–human and human–machine interaction, as well as their implications for the study of reasoning and discourse.     

A computational-level explanation of the speed of goal inference

The ability to understand the goals that drive another person’s actions is an important social and cognitive skill. This is no trivial task, because any given action may in principle be explained by different possible goals (e.g., one may wave ones arm to hail a cab or to swat a mosquito). To select which goal best explains an observed action is a form of abduction. To explain how people perform such abductive inferences, Baker, Tenenbaum, and Saxe (2007) proposed a computational-level theory that formalizes goal inference as Bayesian inverse planning (BIP). It is known that general Bayesian inference–be it exact or approximate–is computationally intractable (NP-hard). As the time required for computationally intractable computations grows excessively fast when scaled from toy domains to the real world, it seems that such models cannot explain how humans can perform Bayesian inferences quickly in real world situations. In this paper we investigate how the BIP model can nevertheless explain how people are able to make goal inferences quickly. The approach that we propose builds on taking situational constraints explicitly into account in the computational-level model. We present a methodology for identifying situational constraints that render the model tractable. We discuss the implications of our findings and reflect on how the methodology can be applied to alternative models of goal inference and Bayesian models in general.     

From Blickets to Synapses: Inferring Temporal Causal Networks by Observation

How do human infants learn the causal dependencies between events? Evidence suggests that this remarkable feat can be achieved by observation of only a handful of examples. Many computational models have been produced to explain how infants perform causal inference without explicit teaching about statistics or the scientific method. Here, we propose a spiking neuronal network implementation that can be entrained to form a dynamical model of the temporal and causal relationships between events that it observes. The network uses spike‐time dependent plasticity, long‐term depression, and heterosynaptic competition rules to implement Rescorla–Wagner‐like learning. Transmission delays between neurons allow the network to learn a forward model of the temporal relationships between events. Within this framework, biologically realistic synaptic plasticity rules account for well‐known behavioral data regarding cognitive causal assumptions such as backwards blocking and screening‐off. These models can then be run as emulators for state inference. Furthermore, this mechanism is capable of copying synaptic connectivity patterns between neuronal networks by observing the spontaneous spike activity from the neuronal circuit that is to be copied, and it thereby provides a powerful method for transmission of circuit functionality between brain regions.     

The role of cues to differential absolute size in children's transitive inferences

To investigate the role that "nonlogical" cues might play in transitive inference, 6- and 7-year-olds were given a three-term transitive task in which perceptual cues to differential absolute size were either present or absent. Relationships between the taught premises and the relational information that was physically present were manipulated using four basic conditions: "congruent," "inverse," "pretended," or "persuaded." Both age groups showed identical overall premise memory, but the younger group tended to reason more on the basis of the perceptual information rather than on the successfully encoded premise information. Contrasts between the various conditions showed that categorical effects can be circumvented in three-term problems with appropriate controls, that there may be qualitative as well as quantitative differences in transitive inference with age, and that transitive inference is not based solely on memory. The findings also indicate that, although 7-year-olds are competent in "logic-based" transitive inference, they experience great difficulty on tasks involving pretend information.     

Causal inferences with group based trajectory models

A central theme of research on human development and psychopathology is whether a therapeutic intervention or a turning-point event, such as a family break-up, alters the trajectory of the behavior under study. This paper lays out and applies a method for using observational longitudinal data to make more confident causal inferences about the impact of such events on developmental trajectories. The method draws upon two distinct lines of research: work on the use of finite mixture modeling to analyze developmental trajectories and work on propensity scores. The essence of the method is to use the posterior probabilities of trajectory group membership from a finite mixture modeling framework, to create balance on lagged outcomes and other covariates established prior to t for the purpose of inferring the impact of first-time treatment at t on the outcome of interest. The approach is demonstrated with an analysis of the impact of gang membership on violent delinquency based on data from a large longitudinal study conducted in Montreal. The research has been supported by the National Science Foundation (NSF) (SES-99113700) and the National Institute of Mental Health (RO1 MH65611-01A2). It also made heavy use of data collected with the support from Québec’s CQRS and FCAR funding agencies, Canada’s NHRDP and SSHRC funding agencies, and the Molson Foundation. We thank Stephen Fienberg, Susan Murphy, Paul Rosenbaum, the editor, Paul De Boeck, and two anonymous reviewers for their insightful suggestions.