Deductive rationality in validating and testing conditional inferences.

We asked people to validate conditional inferences (e.g., "A, therefore C" with "if A then C"). People are more likely to look for falsifications ("A and not-C") versus confirmations ("A and C") given a forced choice. Second, falsification rates are lower for logically valid versus invalid inferences. Logically valid inferences are inferences that follow necessarily. Experiment 1 (N = 96) shows that emphasising this logicality constraint increases falsification rates in the validation task and corroborates that validation-by-falsification increases logically correct inference evaluations. Experiment 2 (N = 41) corroborates the other way round that people who are more likely to make logically correct evaluations, show higher falsification performance in the validation task. The results support mental-models theory and suggest alternative theories similarly need to specify how people would go about looking for counterexamples. We proffer such a specification for two alternatives to the model theory. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Can any ostrich fly?: some new data on belief bias in syllogistic reasoning

According to one version of the mental models theory (Oakhill, J.V., Johnson-Laird, P.N., Garnham, A., 1989. Believability and syllogistic reasoning. Cognition 31, 117-140) beliefs exert their influence on reasoning in three ways. First they can affect the interpretation of the premises, for example by conversion. Second, they can curtail the search for alternative models of the premises, if an initial model supports a believable conclusion. Third, they can act as a filter on any conclusion that is eventually generated. This last influence is important in explaining the effects of belief bias in one-model syllogisms with no convertible premises, since such syllogisms, by definition, have no alternative models. However, the most natural interpretation of such a filter is that it filters out conclusions and leads to the response 'no valid conclusion'. The present study, which was conducted with groups of both British and Italian subjects, looked at the effect of prior knowledge on syllogistic reasoning, and showed that: (1) invalid conclusions for such one model syllogisms, either thematic or abstract, are typically not of the type 'no valid conclusion', but state invalid relations between the end terms; (2) belief-bias is completely suppressed when previous knowledge is incompatible with the premises, and therefore the premises themselves are always considered. The results are compatible with a version of the mental models theory in which a representation of prior knowledge precedes modelling of the premises, which are then incorporated into the representation of this knowledge. The relation between this theory and other accounts of belief bias in syllogistic reasoning, and the implications of these findings for reasoning more generally, are considered.     

Conditional reasoning, causality, and the structure of semantic memory: strength of association as a predictive factor for content effects

Available evidence indicates that responses to conditional inferences using concrete causal premises is affected by the relative number of available alternate causes (Cummins, D.D., 1995. Memory and Cognition 23 (5), 646-658). We propose that another important factor that may influence the kinds of inferences made to causal conditionals is the relative strength of association between such causes and the consequent term. We present a study with adult participants that examines the effect of strength of association on performance on a conditional reasoning task using causal premises for which there exist one highly associated potential cause for the given consequent term. We predicted that adults would produce a greater proportion of biconditional responses to invalid forms with strongly associated premises than weakly associated ones, while valid forms would not be affected by strength of association. The results are consistent with this hypothesis.     

Representations as mediators of adolescent deductive reasoning

In Experiment 1, preadolescents, middle adolescents, and late adolescents were presented 3 deductive reasoning tasks. With some important exceptions, conditional reasoning improved with age on problems containing permission conditional relations, and reasoning fallacies increased with age on problems containing causal conditional relations. The results of Experiments 2a and 2b indicated that problem type (i.e., permission or causal) does not mediate the activation of conditional reasoning skills. Rather, valid conditional inferences are more common on problems for which plausible alternative antecedents can be generated than on problems for which alternative antecedent generation is difficult. Conditional rules for which alternative antecedent generation is difficult may be misrepresented as biconditionals, resulting in biconditional rather than conditional reasoning.     

Hypothetical and certain inferences from conditional arguments read in texts.

In 3 experiments, affirmative and hypothetical probes were presented after narrative texts containing conditional arguments. According to the data, readers represented modus ponens deductions as certain, except when it was only a weakly necessary cause of a given effect. They represented any logically invalid inferences resulting from affirming the conditional consequent as hypothetical, except when it was the effect of strongly sufficient cause. Accordingly, readers must be processing conditional syntax as an asymmetric constraint. However, the underlying causal knowledge can be sufficient either to discredit or warrant the inferences. Thus according to the theory of natural logic, readers can draw formal deductions and be convinced of their necessity. This provides further evidence that readers can represent their inferences as hypothetical (N. Campion, 2004). (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Reasoning about certainty and uncertainty in concrete, causal, and propositional contexts.

Examined children's comprehension of certainty and uncertainty within the context of concrete and propositional reasoning tasks. 69 1st, 3rd, and 5th graders were given G. Pieraut-LeBonniec's (1980) box task and a multisufficient causality task to assess reasoning about certainty and uncertainty in concrete contexts. Ss were also given conditional syllogisms to assess this ability in a propositional context. Half of the Ss at each grade were given contramanded syllogism task statements intended to block erroneous conversational inferences made about these conditional statements. Results indicate that there were no developmental differences in reasoning about concrete certainty, but significant improvement occurred with age in reasoning about concrete uncertainty. On syllogisms, only the 5th graders benefited from contramanding and thus demonstrated an understanding of propositional uncertainty. Correlational and error analyses showed that the discrimination between certainty and uncertainty was mastered in concrete contexts prior to the time when this discrimination occurred in propositional contexts. It is concluded that reasoning about concrete certainty and uncertainty requires a different competence than that required for reasoning about propositional certainty and uncertainty. (17 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Nature of life events and difficulties in depressed adolescents

OBJECTIVE: To examine the significance of acute life events and ongoing difficulties in adolescents with a recent major depressive disorder. METHOD: Adolescents (aged 13-18 years) with a recent episode of major depressive disorder based on DSM-III-R (n = 26) and normal controls free of any Axis I lifetime psychiatric disorder (n = 15) were assessed using the investigator-based Life Events and Difficulties Schedule (LEDS). RESULTS: Traditionally defined severe events were more likely to occur in the year prior to onset among depressed adolescents (46%) than in a comparable period among normal controls (20%), but these differences did not reach statistical significance. Expanding the definition of severe events to include those events focused on others important to the adolescent resulted in a significantly higher percentage of depressed adolescents having one or more refined "severe" events in the year prior to onset (62%) compared with normal controls (27%) (p < or = .02). It is interesting that one half of the depressed adolescents had two or more refined severe events occur during the year prior to onset compared with none of the normal controls (p < or = .01). Further analyses showed that depressed adolescents were significantly more likely to have a major difficulty precede the onset of their depression (27%) compared with normal controls (0%) (p < or = .04). CONCLUSIONS: The results suggest that depressed adolescents are exposed to high levels of stress prior to becoming depressed. Future investigations might benefit from using the LEDS with adolescents to assess acute and ongoing stressors.     

Biomechanical analysis of masticatory system configuration in Neandertals and Inuits

Johnson-Laird and Byrne (1991) present a theory of conditional inference based upon the manipulation of mental models. In the present paper, the theory is critically examined with regard to its ability to account for psychological data, principally with respect to the rate at which people draw the four basic inferences of modus ponens (MP), denial of the antecedent (DA), affirmation of the consequent (AC) and modus tollens (MT). It is argued first that the theory is unclear in its definition and in particular with regard to predictions of problem difficulty. Clarification and specification of principles are consequently provided here. Next, it is argued that there are a number of phenomena in the conditional reasoning literature for which the theory cannot account in its present form. Specifically, (a) the relatively frequency of DA and AC inferences on affirmative conditional is not as predicted by the theory, (b) differences occur between inferences on if then and only if rules beyond the capacity of the theory to explain and (c) there is no account of the "negative conclusion bias" observed when negated components are introduced into the rules. A number of revisions to the mental model theory of conditional reasoning are proposed in order to account for these findings.     

Reasoning With Premises That Are Not Empirically True: Evidence for the Role of Inhibition and Retrieval.

Two studies examined conditional reasoning with false premises. In Study 1, 12- and 16-year-old adolescents made "if-then" inferences after producing an alternative antecedent for the major premise. Older participants made more errors on the simple modus ponens inference than did younger ones. Reasoning with a false premise reduced this effect. Study 2 examined the relation between performance on a negative priming task (S. P. Tipper, 1985) and reasoning with contrary-to-fact premises in 9- and 11-year-olds. Overall, there was a correlation between the relative effect of negative priming on reaction times and the number of knowledge-based responses to the reasoning problems. The results of these studies are consistent with the idea that reasoning with premises that are not true requires an interaction between information retrieval and inhibition. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

On the probability of making Type I errors.

A statistical test leads to a Type I error whenever it leads to the rejection of a null hypothesis that is in fact true. The probability of making a Type I error can be characterized in the following 3 ways: the conditional prior probability, the overall prior probability, and the conditional posterior probability. In this article, we show (a) that the alpha level can be equated with the 1st of these and (b) that it provides an upper bound for the second but (c) that it does not provide an estimate of the third, although it is commonly assumed to do so. We trace the source of this erroneous assumption first to statistical texts used by psychologists, which are generally ambiguous about which of the 3 interpretations is intended at any point in their discussions of Type I errors and which typically confound the conditional prior and posterior probabilities. Underlying this, however, is a more general fallacy in reasoning about probabilities, and we suggest that this may be the result of erroneous inferences about probabilistic conditional statements. Finally, we consider the possibility of estimating the (posterior) probability of a Type I error in situations in which the null hypothesis is rejected and, hence, the proportion of statistically significant results that may be Type I errors. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Theories of Conditional Reasoning: A Developmental Examination of Competing Hypotheses.

Children and adolescents were presented with problems that contained deontic (i.e., if action p is taken, then precondition q must be met) or causal (i.e., if event p occurs, then event q will transpire) conditionals and that varied in the ease with which alternative antecedents could be activated. Results showed that inferences were linked to the availability of alternative antecedents and the generation of "disabling" conditions (claims that the conditionals were false under specific circumstances). Age-related developments were found only on problems involving indeterminate inferences. Correlations among inferences differed for children and adolescents. The findings provide stronger support for domain-general theories than for domain-specific theories of reasoning and suggest, under some conditions, age-related changes in the roles of implicit and explicit processing. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Working memory and bias in reasoning across the life span.

Age differences in syllogistic reasoning in relation to crystallized and fluid ability were studied in 278 adults from 19 to 96 yrs of age. Two reasoning tasks, the evaluation and the construction of conclusions for syllogisms of varying complexity and believability, a vocabulary test, and 3 tasks of working memory were administered. The magnitude of age-related variance on selected reasoning tasks was only partially reduced by statistically controlling measures of both working memory and vocabulary. Additional age-related effects on reasoning were found to be significantly associated with number of mental models and bias produced by conflict between belief and logic. A significant bias was also found toward acceptance of invalid syllogisms as valid, even when contents were abstract. These sources of error in logic are discussed in relation to P. N. Johnson-Laird's (1983) theory of mental models and J. St. B. T. Evans's (1989) account of bias in human reasoning. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Reading and arithmetic in adolescents with autism spectrum disorders: Peaks and dips in attainment.

In describing academic attainment in autism spectrum disorders (ASD), results are typically reported at the group mean level. This may mask subgroups of individuals for whom academic achievement is incommensurate with intellectual ability. The authors tested the IQ, literacy, and mathematical abilities of a large group (N = 100) of adolescents (14–16 years old) with ASD. Seventy-three percent of the sample had at least one area of literacy or mathematical achievement that was highly discrepant (approximately 14 standard score points) from full-scale IQ (FSIQ). The authors focused on four subgroups with either word reading (“Reading Peak” and “Reading Dip”) or arithmetic (“Arithmetic Peak” and “Arithmetic Dip”) higher or lower than FSIQ. These subgroups were largely mutually exclusive and were characterized by distinct intellectual profiles. The largest was the “Arithmetic Peak” subgroup of participants, who presented with average intellectual ability alongside superior arithmetic skills and who were predominantly in a mainstream educational setting. Overall, the most pervasive profile was discrepantly poor reading comprehension, which associated with severity of social and communication difficulties. The high rate of uneven academic attainment in ASD has implications for educational practice. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The impact of print exposure quality and inference construction on syllogistic reasoning.

This study extended the work of S. Siddiqui, R. F. West, and K. E. Stanovich (1998), who studied the link between general print exposure and syllogistic reasoning. It was hypothesized that exposure to certain text structures that contain well-delineated logical forms, such as popularized scientific texts, would be a better predictor of deductive reasoning skill than general print exposure, which is not sensitive to the quality of an individual's reading activity. Furthermore, it was predicted that the ability to generate explanatory bridging inferences while reading would also be predictive of syllogistic reasoning. Undergraduate students (N = 112) were tested for vocabulary, nonverbal cognitive ability, exposure to general print, exposure to popularized scientific literature, and the ability to comprehend texts distinguished by the number of inferences that must be generated to support comprehension. Hierarchical multiple regression analyses showed that a combined measure of exposure to general and scientific literature was a significant predictor of syllogistic reasoning ability. Additionally, the ability to comprehend high-inference-load texts was related to solving syllogisms that were inconsistent with world knowledge, indicating an overlap in deductive reasoning skill and text comprehension processes. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Conversational comprehension processes are responsible for reasoning fallacies in children as well as adults: If is not the biconditional.

Proposes that children do poorly on reasoning from premises of the form if p then q not because they construe if as a biconditional but rather because they use discourse comprehension processes that lead them to accept the invited inferences if not p then not q and if q then p. This hypothesis predicts that children should respond appropriately to premises in which the invited inferences are countermanded. In Exp I, 24 undergraduates and 44 10-yr-olds were given conditional reasoning problems. Some of these had a major premise consisting of a single if–then sentence, while others had a more elaborate major premise in which the invited inferences were explicitly countermanded. In Exp II, Ss were 24 undergraduates, 20 10-yr-olds, and 34 7-yr-olds. In some problems the major premise consisted of a single if–then statement; in others, the major premise consisted of 3 such statements, 2 of which shared the same consequent, thus implicitly countermanding the invited inferences. In both experiments, all age groups committed the fallacies in the simple condition but not in the more complex condition. It is concluded that children's representation of if distinguishes necessary from merely invited inferences. Data suggest a collection of countermandable context-dependent inferences of varying degrees of invitingness associated with if. (19 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Diagnostic, demographic, memory quality, and cognitive variables associated with acute stress disorder in children and adolescents.

To date, no studies have investigated factors associated with acute stress disorder (ASD) in children and adolescents. Relationships between ASD and a number of demographic, trauma, cognitive, and trauma memory variables were therefore investigated in a sample (N=93) of children and adolescents involved in assaults and motor vehicle accidents. Several cognitive variables and the quality of trauma memories, but not demographic or trauma variables, were correlated with ASD and also mediated the relationship between peritraumatic threat and ASD. Finally, nosological analyses comparing ASD with indexes of posttraumatic stress disorder in the month posttrauma revealed little support for the dissociation mandate that uniquely characterizes ASD. The results are discussed with respect to assessment and treatment for the acute traumatic stress responses of children and young people. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Adolescents' reasoning about exclusion from social groups.

Ninth- and 11th-grade students (N=379) were surveyed regarding their evaluations of excluding someone from a social group solely on the basis of his or her social reference group membership. Individuals evaluated exclusion in ambiguous and nonambiguous situations. Judgments and reasoning about exclusion were compared with judgments and reasoning about a more prototypically moral situation (denial of resources). Overall, participants evaluated exclusion as less wrong than denial of resources and used fewer moral and more conventional reasons to justify their judgments. Participants relied more on their group knowledge or stereotypes in evaluating ambiguous situations and more on their personal knowledge in evaluating nonambiguous situations. Age- and gender-related differences in evaluations, reasoning, and use of stereotypes were also found. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The role of classification on testing conditional sentences in the context of an empirical task.

An analysis of a disjunctive concept based on 2 binary dimensions reveals that the relationships between the individual stimuli and classification are implicatory; these relationships may be expressed in conditional sentences of the form "if p then q." A 4-phase experiment with 40 students (average age 17 yrs 6 mo) was conducted to determine the effect on testing the truth value of conditional sentences of (a) prior experience in classifying the instances of the conditional, and (b) knowledge that such classifications would be of empirical value in a subsequent concept learning task. Phase I involved classifying individual stimuli for a subsequent concept learning problem; Phase II was the problem itself; Phase III was testing the truth value of 4 conditional sentences which, depending on whether they were true or false, described respectively the implicatory or tautologous relationships inherent in a concept learning task that followed (Phase IV). It is concluded that both prior experience with classifying the components of an implicatory rule and knowledge of the empirical relevance of such classifications to a subsequent task were necessary to the emergence of logical reasoning in the Phase III selection task. (14 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Use of a sequencing task designed to stress the supervisory system in schizophrenic subjects

BACKGROUND: We investigated whether schizophrenic subjects are impaired in non-routine behaviour because of the dysfunction of a general executive component labelled, in neuropsychological terms, the supervisory system. METHODS: A specific verbal sequencing test was designed for this purpose. Subjects had to perform sequential reasoning with verbal material. Each test sequence consisted of a series of words presented in jumbled order. The construction of some sequences had to be done using familiar routine associations (valid conditions). In contrast, some other sequences required the overriding selection of familiar routine associations, which were inappropriate within the general context of the task (invalid conditions). Twenty verbal sequences (10 valid-10 invalid) were administered. Thirty-seven DMS-IV schizophrenic patients and 21 normal volunteers matched for age and educational level were recruited. RESULTS: Compared to the control group the schizophrenic group was impaired in both valid and invalid conditions. The number of 'capture errors' specific to supervisory system failure was significantly higher in the schizophrenic group and only the schizophrenic patients had significantly fewer correct sequences in invalid conditions than in valid conditions. Poor performance in invalid conditions alone was observed only among the schizophrenic subjects without a general cognitive defect. CONCLUSIONS: These findings suggest that sequencing procedures requiring an executive input are impaired in schizophrenia.