Perceiving a strong causal relation in a weak contingency: Further investigation of the evidential evaluation model of causal judgement

Contingency information is information about the occurrence or nonoccurrence of a certain effect in the presence or absence of a candidate cause. An objective measure of contingency is the δP rule, which involves subtracting the probability of occurrence of an effect when a causal candidate is absent from the probability of occurrence of the effect when the candidate is present. Causal judgements conform closely to δP but deviate from it under certain circumstances. Three experiments show that such deviations can be predicted by a model of causal judgement that has two components: a rule of evidence, that causal judgement is a function of the proportion of relevant instances that are judged to be confirmatory for the causal candidate, and a tendency for information about instances in which the candidate is present to have greater effect on judgement than instances in which the candidate is absent. Two experiments demonstrate how this model accounts for some recently published findings. A third experiment shows that it is possible to use the model to predict the occurrence of high causal judgements when the objective contingency is close to zero.     

Judgement of two causal candidates from contingency information: II. Effects of information about one cause on judgements of the other cause

When judgements are being made about two causes there are eight possible kinds of contingency information: occurrences and nonoccurrences of the outcome when both causes are present, when Cause 1 alone is present, when Cause 2 alone is present, and when neither cause is present. It is proposed that contingency information is used to some extent to judge proportionate strength, which is the proportion of occurrences of the outcome that each cause can account for. This leads to a prediction that judgements of one cause will be influenced by information about occurrences, but not nonoccurrences, of the outcome when only the other cause is present. In six experiments consistent support was found for this prediction when the cause being judged had a positive relation with the outcome, but no consistent tendency was found when the cause being judged had a negative relation with the outcome. The effects found for causes with positive contingency cannot be explained by the Rescorla-Wagner model of causal judgement nor by the hypothesis that causal judgements are based on conditional contingencies.     

Judgement of two causal candidates from contingency information: Effects of relative prevalence of the two causes

In four experiments participants made judgements about two possible causes of an effect. The prevalence of the causes was manipulated independently of their degree of contingency with the effect. Significant effects of the prevalence manipulation were obtained: In particular, ratings of the unmanipulated candidate tended to decline as the prevalence of the other candidate increased, and there was also a significant but smaller effect on judgements of the latter. These tendencies were independent of the order in which the two candidates were judged. These results were replicated under two stimulus presentation procedures, the instance list procedure (Experiments 1 and 2) and the trial-by-trial procedure (Experiment 3). It was hypothesized that people judge, to some degree, the proportionate strength of the candidates, in other words the proportion of occurrences of the effect in the presence of each, and that the cause prevalence effect is a consequence of this tendency. This hypothesis was supported by the results of Experiment 4: Those participants whose judgements of one candidate were negatively correlated with the frequency of occurrence of the effect in the presence of the other candidate showed a significantly stronger cause prevalence effect than the remainder.     

Causal judgment from contingency information: A systematic test of the<Emphasis Type="Italic">p</Emphasis>CI rule

Contingency information is information about the occurrence or nonoccurrence of an effect when a possible cause is present or absent. Under the evidential evaluation model, instances of contingency information are transformed into evidence and causal judgment is based on the proportion of relevant instances evaluated as confirmatory for the candidate cause. In this article, two experiments are reported that were designed to test systematic manipulations of the proportion of confirming instances in relation to other variables: the proportion of instances on which the candidate cause is present, the proportion of instances in which the effect occurs when the cause is present, and the objective contingency. Results showed that both unweighted and weighted versions of the proportion-of-confirmatory-instances rule successfully predicted the main features of the results, with the weighted version proving more successful. Other models, including the power PC theory, failed to predict the results.     

Cue interaction effects in causal judgement: An interpretation in terms of the evidential evaluation model

In judging the extent to which a cue causes an outcome, judgement can be affected by information about other cues that are correlated with the one being judged. These cue interaction effects have usually been interpreted in terms of associative learning processes. I propose that a different model of causal judgement, the evidential evaluation model, offers a viable alternative interpretation of cue interaction phenomena. Under the evidential evaluation model, instances of contingency information are interpreted as evidence, which is confirmatory, disconfirmatory, or irrelevant for the cue being judged. When two cues co-occur in a set of instances the evidential value of the instances for one of them is determined by three factors: the proportion of confirming instances in the set; disambiguation value, which concerns the relation between the set of information and prior beliefs about the co-occurring cue; and confirmation value, which concerns the relation between the set of information and prior beliefs about the cue being judged. Any previous judgement of the cue is then modified in the light of these. It is shown that this model can account for all the cue interaction phenomena that have been investigated in studies of human causal judgement. The model also generates novel predictions, and the results of three experiments give support to these predictions. It is also shown that several other current models of causal judgement fail to predict a key result from Experiment 3.     

Inferences about unobserved causes in human contingency learning

Estimates of the causal efficacy of an event need to take into account the possible presence and influence of other unobserved causes that might have contributed to the occurrence of the effect. Current theoretical approaches deal differently with this problem. Associative theories assume that at least one unobserved cause is always present. In contrast, causal Bayes net theories (including Power PC theory) hypothesize that unobserved causes may be present or absent. These theories generally assume independence of different causes of the same event, which greatly simplifies modelling learning and inference. In two experiments participants were requested to learn about the causal relation between a single cause and an effect by observing their co-occurrence (Experiment 1) or by actively intervening in the cause (Experiment 2). Participants' assumptions about the presence of an unobserved cause were assessed either after each learning trial or at the end of the learning phase. The results show an interesting dissociation. Whereas there was a tendency to assume interdependence of the causes in the online judgements during learning, the final judgements tended to be more in the direction of an independence assumption. Possible explanations and implications of these findings are discussed.     

Causal judgement as evaluation of evidence: The use of confirmatory and disconfirrnatory information

There are four kinds of contingency information: occurrences and nonoccurrences of an effect in the presence and absence of a cause. In two experiments participants made judgements about sets of stimulus materials in which one of these four kinds had zero frequency. The experiments tested two kinds of predictions derived from the evidential evaluation model of causal judgement, which postulates that causal judgement depends on the proportion of instances evaluated as confirmatory for the cause being judged. The model predicts significant effects of manipulating the frequency of one kind of contingency information in the absence of changes in the objective contingency. The model also predicts that extra weight will be given to one kind of confirmatory information when the other kind has zero frequency, and to one kind of disconfirmatory information when the other kind has zero frequency. Results supported both sets of predictions, and also disconfirmed predictions of the power probabilistic contrast theory of causal judgement. This research therefore favours an account of causal judgement in which contingency information is transformed into evidence, and judgement is based on the net confirmatory or disconfirmatory value of the evidence.     

Launching at a distance: The effect of spatial markers

It has been widely reported that spatial contiguity is important to judgements of causality involving one object launching another [Michotte's “launching effect” (1963, 1991)]. The present study examined the impact of different types of spatial markers on causal judgements of a distal launch (one object approaching other, stopping short of it, and the second object subsequently moving along the same trajectory). The spatial markers were objects that either partially or completely bridged the spatial gap between two objects (Experiment 1), or they were dashed lines that marked the stopping location of the first object or the starting location of the second object (Experiment 2). The presence of either objects or dashed lines could produce higher causal ratings, but the location of the marker mattered. The results suggest that altering a cause's ability to predict when the effect would occur (via a spatial marker) and the presence of a conduit for energy transmission have independent effects on causal judgements of object interaction.     

Causal attribution from covariation information: the evidential evaluation model

It is hypothesized that causal attributions are made by transforming covariation information into evidence according to notions of evidential value, and that causal judgement is a function of the proportion of instances that are evaluated as confirmatory for the causal hypothesis under test: this is called the evidential evaluation model. An experiment was designed to test the judgemental rule in this model by setting up problems presenting consensus, distinctiveness, and consistency information in which the proportion of confirmatory instances varied but the objective contingency did not. It was found that judgements tended to vary with the proportion of confirmatory instances. Several other current models of causal judgement or causal attribution fail to account for this result. Similar findings have been obtained in studies of causal judgement from contingency information, so the present findings support an argument that the evidential evaluation model provides a unified account of judgement in both domains. Copyright © 2002 John Wiley & Sons, Ltd.     

Metacomprehension judgements reflect the belief that diagrams improve learning from text

In two experiments we systematically explored whether people consider the format of text materials when judging their text learning, and whether doing so might inappropriately bias their judgements. Participants studied either text with diagrams (multimedia) or text alone and made both per-paragraph judgements and global judgements of their text learning. In Experiment 1 they judged their learning to be better for text with diagrams than for text alone. In that study, however, test performance was greater for multimedia, so the judgements may reflect either a belief in the power of multimedia or on-line processing. Experiment 2 replicated this finding and also included a third group that read texts with pictures that did not improve text performance. Judgements made by this group were just as high as those made by participants who received the effective multimedia format. These results confirm the hypothesis that people's metacomprehension judgements can be influenced by their beliefs about text format. Over-reliance on this multimedia heuristic, however, might reduce judgement accuracy in situations where it is invalid.     

Is subjective shortening in human memory unique to time representations?

Three experiments compared forgetting of the duration of a bar-like visual stimulus with forgetting of its length. The main aim of the experiments was to investigate whether subjective shortening (a decrease in the subjective magnitude of a stimulus as its retention interval increased) was observable in length judgements as well as in time judgements, where subjective shortening has been often observed previously. On all trials of the three experiments, humans received two briefly presented coloured bars, separated by adelay ranging from 1 to 10 s, and the bars could differ in length, duration of presentation, or both. In Experiment 1 two groups of subjects made either length or duration judgements, and subjective shortening-type forgetting functions were observed only for duration. Experiments 2 and 3 used the same general procedure, but the stimuli judged could differ both in length and duration within a trial, and different subject groups (Experiment 2) or the same subjects in two conditions (Experiment 3) made either length or duration judgements of stimuli, which were on average physically identical. Subjective shortening was only found with duration, and never with length, supporting the view that subjective shortening may be unique to time judgements.     

Discounting in causality and covariation judgements

In two experiments, humans were asked to judge the strength of a moderate contingency between a cue and an outcome in the presence of a second strong contingency between another cue and the outcome. The first experiment replicated the discounting effect whereby a strong contingency causes subjects to reduce or discount judgements of a moderate contingency. This experiment used a video-game procedure in which subjects camouflaged a tank to make it safe from mines. The second causal cue was the presence or absence of a spotter plane. Experiment 1 also ruled out the possibility that judgements might be determined by the number of co-occurrences of the cue and outcome as opposed to the level of contingency. The second experiment used an abstract scenario in which discounting was demonstrated when subjects were asked to judge the relationship between the occurrence of geometric objects. The instructions were neutral to discourage causal hypotheses. These results support the notion that judgements result from associative or connectionist processes and not from a two-step cognitive retrospective process in which an estimate of covariation is calculated between cue and outcome and then in a second step this information is used in a normative manner when a decision is made.     

Valuing intervention and observation

Understanding causal relations is fundamental to effective action but causal data can be confounded. We examined the value that participants placed on data derived from a hypothetical intervention or observation. Our materials involved a possible cause ("bottled water"), a possible confound ("food"), and a context ("a restaurant"). We supposed that participants seek to draw as specific a causal inference as possible from presented data and value information sources more highly that allow them to do so. On this basis, we predicted that in circumstances where an intervention removed the confounding causal factor but observation did not, participants would prefer data derived from an intervention when the possible cause was present (the bottled water was drunk) but show the reverse preference when the possible cause was absent (the bottled water was not drunk). Experiment 1 confirmed this prediction. Using a between-subjects design, Experiment 2 tested for a difference in confidence in causal judgements given identical data, including data on the confound, as a function of method of data collection (intervention or observation). There was no significant difference in confidence ratings between the two methods but confidence ratings were sensitive to the probability of an effect (illness) given the cause. Using a within-subjects design, Experiment 3 revealed systematic individual differences in preference for the two methods. Participants were divided between those who considered intervention more confounded and those who considered observation more confounded. Our experiments point to the subtleties of participants' evaluation of data from studies of human beings.     

Evidence for a distinction between judged and perceived causality

Two experiments investigated Michotte's launch event, in which successive motion of two objects appears to evoke an immediate perception that the first motion caused the second, as in a collision. Launching was embedded in event sequences where a third event (a colour change of the second object) was established as a competing predictor of the second motion, in an attempt to see whether subjects' learning of alternative predictive relationships would influence their causal impressions of launch events. In Experiment 1 subjects saw launch events in which temporal contiguity at the point of impact was varied so that an impact was varied so that an impact itself did not reliably predict when the second object would move. Half of these scenes, however, contained a colour change of the second object which did reliably predict when it would move. In accordance with Michotte's theory, subjects' ratings of the degree of perceived causality were not affected by the colour change. In Experiment 2 subjects saw scenes that contained launch events with or without temporal contiguity and a colour change. These were interspersed with events in which a colour change alone did or did not precede the second motion. Thus, movement of the second object was either contingent on or independent of the impact. Subjects repeatedly (a) rated perceived causality in single launch events and (b) judged the necessity of collisions for movement in the overall set of events. These responses dissociated, in that ratings type (a) showed only a substantial contiguity effect, whereas judgements of type (b) showed both a contingency and a much smaller contiguity effect. These results appear to support a distinction between judged and perceived causality and are discussed with respect to Michotte's theory of direct causal perception.     

Valuing intervention and observation

Understanding causal relations is fundamental to effective action but causal data can be confounded. We examined the value that participants placed on data derived from a hypothetical intervention or observation. Our materials involved a possible cause (“bottled water”), a possible confound (“food”), and a context (“a restaurant”). We supposed that participants seek to draw as specific a causal inference as possible from presented data and value information sources more highly that allow them to do so. On this basis, we predicted that in circumstances where an intervention removed the confounding causal factor but observation did not, participants would prefer data derived from an intervention when the possible cause was present (the bottled water was drunk) but show the reverse preference when the possible cause was absent (the bottled water was not drunk). Experiment 1 confirmed this prediction. Using a between-subjects design, Experiment 2 tested for a difference in confidence in causal judgements given identical data, including data on the confound, as a function of method of data collection (intervention or observation). There was no significant difference in confidence ratings between the two methods but confidence ratings were sensitive to the probability of an effect (illness) given the cause. Using a within-subjects design, Experiment 3 revealed systematic individual differences in preference for the two methods. Participants were divided between those who considered intervention more confounded and those who considered observation more confounded. Our experiments point to the subtleties of participants' evaluation of data from studies of human beings.     

经验信息与共变信息对判断两原因共同作用的影响

本研究考察了被试在使用经验信息和共变信息进行两原因共同作用因果判断时所具有的特点。研究结果表明:(1)两原因的性质对于被试的选择有着显著影响。当两原因的可信度相等时,更多被试认为两原因都可以引起结果。(2)两原因都不出现时结果出现的概率P(e/~i~j)对于被试的选择没有显著影响。(3)两原因单独出现时结果出现的概率对于被试的选择有着显著的影响,但是这一变量是与两原因的可信程度共同产生影响的。     

Strategy selection in causal reasoning: when beliefs and covariation collide.

The present study investigated how people combine covariation information (Cheng & Novick, 1990, 1992) with pre-existing beliefs (White, 1989) when evaluating causal hypotheses. Three experiments, using both within- and between-subjects designs, found that the use of covariation information and beliefs interacted, such that the effects of covariation were larger when people assessed hypotheses about believable than about unbelievable causal candidates. In Experiment 2, this interaction was observed when participants made judgments in stages (e.g., first evaluating covariation information about a causal candidate and then evaluating the believability of a candidate), as well as when the information was presented simultaneously. Experiment 3 demonstrated that this pattern was also reflected in participants' metacognitive judgments: Participants indicated that they weighed covariation information more heavily for believable than unbelievable candidates. Finally, Experiments 1 and 2 demonstrated the presence of individual differences in the use of covariation- and belief-based cues. That is, individuals who tended to base their causality judgments primarily on belief were less likely to make use of covariation information and vice versa. The findings were most consistent with White's (1989) causal power theory, which suggests that covariation information is more likely to be considered relevant to believable than unbelievable causes.     

Contribution of perceptual fluency to recognition judgments.

Following a shallow (count vowels) or deep (read) study task, old and new words were tested for both fluency of perception and recognition memory. Subjects first identified a test word as it came gradually into view and then judged it as old or new. Old words were identified faster than new words, indicating implicit, perceptual memory for old words. Independently of this effect, words judged old were identified faster than words judged new, especially after shallow study. Eight experiments examined the possible causal relationship between perceptual fluency and recognition judgements. Experiments 1 to 4 showed that fast identifications per se do not promote old judgments. Accelerating the identification of test items by semantically priming them or making them come more quickly into view did not affect recognition judgments. Experiment 5 showed that the usual association of fast identifications with old judgments is not an artifact of item selection because the association disappeared when the identifications and judgements were segregated into different phases of the test task. Experiments 6 and 7 showed tha the likelihood of old judgments increases directly with the pretested perceptibility of test words, but only after shallow study. Experiment 8 showed that the dependency of recognition judgments on perceptual fluency continues to hold when the requirement to identify the words before judging them is eliminated. We conclude that fluency of perception contributes to recognition judgments, but only when the fluency is produced naturally (e.g., through perceptual memory) and explicit memory is minimal.