The process of recognition is usually conceptualized as finding a match between a perceptual representation of a given stimulus and the representations (or traces) in memory of previously encountered stimuli. Thus, this notion of recognition is intimately connected to the notion of representation, “the most important concept ever evoked in explaining the mind” (p. 1).

The book starts off rather philosophically, from Edelman’s unshakeable belief in the veridicality of shape perception and his optimism regarding the plausibility of a particular formal theory of it. He quotes classic philosophers like Berkeley, Hume, Locke, Wittgenstein, … and situates his approach in the present-day cluster of philosophy of mind, philosophy of language, and theory of knowledge (e.g., Clark, Cummins, Dretske, Millikan, Putnam). Starting from the age-old problem of understanding what it means to see a cat on a mat, the book is explicitly written by a cognitive scientist, a non-philosopher, “to meet the philosophers (at least the more empirically minded of them) halfway down the road” (p. 2).

Chapter 1 (pp. 1–10) reviews the theoretical and practical difficulties with the attempt to base recognition on geometrically reconstructed representations of distal objects (like Marr’s 2,5-D sketch): Representation by reconstruction first leads to the homunculus problem, and, second, has been shown to be practically impossible, even using the most sophisticated computer vision techniques. Any alternative approach must then state what representation is, if not reconstruction, and, in what sense, if not geometric, it can be veridical. Edelman’s core principle, introduced in Chapter 2, is that representation is representation of similarities.

Chapter 2 (pp. 11–42) contains three rather different sections. Section 2.1 distinguishes four recognition tasks in which representations are needed: (1) identification (i.e., recognition of a previously seen view of an object), (2) generalization (i.e., recognition of an object despite a change in its appearance due to some transformation), (3) categorization (i.e., attribution of an object to a class of similar objects), and (4) analogy (i.e., drawing a parallel between transformations of distinct objects). Section 2.2 attempts to define and formalize the notion of representation. Borrowing from Shepard, Edelman distinguishes between two types of mappings: first-order isomorphism, between objects in the world and their corresponding internal representations, and second-order isomorphism, between the relations among several external objects and the relations among their corresponding internal representations. Because first-order isomorphism, or representation by similarity, leads to the problematic reconstructionist approach, Edelman proposes to follow the formally just as good second-order isomorphism, which he calls “representation of similarity”. Section 2.3 reviews several problems with current computational theories of recognition, such as Biederman’s “Recognition by Components” (RBC) theory, Ullman’s alignment theory, and multidimensional feature spaces.

Chapter 3 (pp. 43–74) introduces the notion of a shape space, a formalism borrowed from mathematical statistics (e.g., Kendall). Shape space is a metric space in which each point corresponds to a particular shape and in which geometric similarity between shapes can be defined rigorously as proximity, a quantity inversely related to distance. Edelman argues that shape similarity can be made not to suffer from the objections commonly raised against a metric-space approach to similarity (e.g., arbitrary choice of features, context dependence, and asymmetries of comparisons). Although a large number of parameters is needed to fully represent distal shape space, fewer may be sufficient if, as in second-order isomorphism, only relations between objects must be represented (e.g., a blending coefficient when morphing the shape of a cow and a pig). To assure a proper mapping from distal to proximal shape space, a perceptual system must carry out many measurements (M) and then reduce the dimensionality of the resulting space (R), while getting rid of the extrinsic variables such as pose and illumination. The perceptual mapping must satisfy the following requirements: (1) the mapping should not collapse any behaviorally relevant dimension of shape variation (i.e., regularity), (2) small changes in object geometry should be mapped onto small changes in R (i.e., smoothness), (3) R should be locally decomposable into view space and shape space components, and (4) R should be locally low-dimensional. With “locally” Edelman refers to a small part of R that contains a class of intrinsically similar shapes.

To realize in practice this theoretically possible veridicality of representation one must find a way to identify the relevant low-dimensional structure within the high-dimensional measurement space. This is done in Chapter 4 (pp. 75–110) for the various recognition tasks distinguished earlier (identification, generalization, categorization, and analogy). The mechanisms that are needed to solve these tasks are based on “navigation” in a shape-space “landscape” with “landmarks”. A novel shape (or a novel view of a previously seen shape) can be localized in shape space based on its similarity to known objects (or stored views of known objects). A tuned unit (or module of units) that responds optimally to some shape (i.e., the landmark) and progressively less to progressively less similar shapes, is enough to implement this mechanism. Edelman uses a radial basis function (RBF) approximation network that can be trained to generalize its response from a series of given views of an object to other views of that object. As a by-product of this learning, this network will also respond progressively less to progressively less similar objects, precisely what is needed for it to work as an active landmark in internal shape space. This scheme not only works for known objects (as in the case of identification and recognition) but also for novel objects (as in the case of categorization and analogy). The trick is that a new object must be represented by a vector of proximities to several reference shapes or landmarks. For example, even if one has never seen a giraffe before, it is possible to compute the proximities to known animals such as a camel, a goat, a pig or even a leopard.

Edelman calls this system a “Chorus of Prototypes” to stress that the prototypes act together in representing shape, in contrast to a Winner-Take-All scheme such as Selfridge’s Pandemonium. In Chapter 5 (pp. 111–143) Edelman provides simulation results of his implemented scheme (called “SiC”, for the combination of second-order isomorphism with a Chorus of Prototypes). He uses a simple system composed of ten prototype modules, each trained on a different reference object (e.g., cow, cat, robot, fly, tuna, Landrover, F16) and then tested on different recognition tasks for a database of 43 additional objects (e.g., another cow, an ox, a calf, a buffalo, and other quadrupeds, along with other fishes, other cars, other airfighters, etc.). The simulation results with SiC are encouraging, especially in light of the wide variety of tasks: identification of novel views of familiar objects, categorization of novel object views, discrimination among views of novel objects, local viewpoint invariance for novel objects, recovery of a standard view and of pose for novel objects, prediction of a novel view for a novel object, etc. None of its theoretical competitors (pictorial alignment, structural matching, or feature spaces) can deal with this variety of tasks.

When the nature of an action pattern is such that instruction focuses on developing a single recommended movement pattern, learning is often stimulated by providing the learner with demonstrations from a skilled model performing the movement pattern. The principal theoretical influence of modeling is to leave the learner with a conception of the way a skill is to be performed, which can serve as a guide for action. In this way, the learner is spared from creating a cognitive conception of the action pattern gradually as a result of trial-and-error experiences. Thus, modeling can serve to increase the efficiency of skill acquisition.

A popular theoretical account of the cognitive mechanisms underlying observational learning presumes that vicarious observational experiences can initiate formation of a cognitive representation in memory that can be enacted and refined during overt practice. According to Bandura, the observer must experience four constituent subprocesses as a result of modeling for observational learning to be successful. The learner must: selectively attend to relevant information in the demonstration (attention subprocess), retain the relevant information (retention subprocess) for eventual imitation, have the capability for using the relevant information for imitation (production subprocess), and have the desire to imitate the modeled action (motivation subprocess). Several techniques are commonly employed to assist the learner in experiencing each subprocess so that information present in a demonstration is consolidated into a memory representation. For example, the use of multiple demonstrations provides repeated exposure to the goal performance in an effort to assist the attention and retention subprocesses. Another technique involves the use of performance cues provided verbally in conjunction with the demonstration to guide the learner’s attention to task-relevant cues present in the demonstration on the assumption that the learner will display elements of these cues in overt performance. Both the use of multiple demonstrations and verbal cues have been shown to enhance performance following demonstration.

Another factor which may potentially influence the observational learning subprocesses is the timing of the demonstration in relation to overt practice. Arguably, the most common time that a learner is exposed to a demonstration of a goal movement pattern is prior to overt practice. A presumed advantage of this method would be that observational experiences would allow consolidation of as much information as possible into a preliminary memory representation before enactment. An alternate method of combining demonstration and overt practice involves interspersing demonstrations throughout the practice sequence. This method would presumably allow an assimilation of information into the representation as practice and demonstration progressed, so that observational learning would interact with overt practice. In this manner, the trial-and-error learning process in overt practice would be augmented by frequent opportunities to reinforce the representation for the movement pattern with observational experiences.

The limited research examining the demonstration–overt practice relationship renders equivocal conclusions. For example, McGuire (1961) compared a condition in which modeling was provided exclusively before practice to a condition in which an expert model was provided both pre-practice, and then again at mid-acquisition for learning recommended technique on the pursuit-rotor tracking task. Groups were not significantly different on either ratings of form or time-on-target, prompting McGuire to conclude that pre-practice modeling was sufficient to maximize observational learning for this task.

Landers (1975) and Thomas, Pierce and Ridsdale (1977) have also examined the question about the best time to introduce a model in the learning sequence to children learning balance tasks. Landers compared a group receiving modeling exclusively before practice to groups receiving modeling either before and at the mid-point of practice, or exclusively at the mid-point of the practice sequence. In the first block of acquisition trials, the groups seeing the model before practice outperformed the group not introduced to the model until mid-practice. The group that viewed the model on a second occasion at mid-practice performed significantly better than the group that received modeling only at the practice mid-point, but only on the trial block immediately following mid-practice modeling. By the end of acquisition, no significant differences among groups existed. Thus, modeling before practice elevated initial performance; however, providing a second opportunity for exposure to the model further benefited performance on this task in the short-term. In addition, modeling was not as effective for elevating performance if it was administered for the first time in mid-acquisition. Thomas et al. (1977) compared pre-practice modeling to mid-practice modeling or no modeling at all in seven and nine year-old children. Modeling prior to practice was beneficial for both age groups compared to no modeling. However, mid-practice modeling was only effective for performance of the nine year-old children; seven year-old children performed less effectively when viewing a model in mid-acquisition relative to no modeling or pre-practice modeling. Thus, the ability to benefit from a model that was introduced after substantial experience had accrued depended on the age of the observer. Neither the Landers (1975) study nor the Thomas et al. (1977) study included retention tests to determine the potential that one of the modeling conditions would impact learning.

Sidaway and Hand (1993) examined the interspersing of modeling throughout the acquisition phase by comparing groups receiving modeling either prior to each attempt in acquisition, or at ratios of one demonstration to every five practice attempts or every 10 practice attempts. Only accuracy in achieving the external goal was assessed; that is, form of the learner was not assessed. No differences existed in accuracy scores between groups in acquisition or in a post-acquisition transfer test. However, the group receiving modeling prior to each trial in acquisition was significantly more accurate in a no modeling retention test. Unfortunately, because form was not assessed, it was not possible to determine whether modeling was effective in promoting form differences among groups. In addition, a group receiving all pre-practice modeling was not included to determine the effectiveness of this modeling schedule relative to schedules with modeling interspersed throughout acquisition.

In the studies of McGuire, 1961 and Landers, 1975, and Thomas et al. (1977), mid-practice modeling was inserted into the practice sequence only after participants had engaged in a considerable amount of overt practice. It is possible that mid-practice demonstration may be ineffective if participants have settled into a movement pattern that is contrary to that depicted in the demonstration. Receiving modeling for the first time at mid-practice could lead to a dissonant learning condition in which the learner would be prompted to readjust the movement pattern to conform to the goal response. However, this shortcoming of mid-practice demonstration might be alleviated if, as was done in the study of Sidaway and Hand (1993), demonstrations were introduced earlier in the overt practice sequence when an appreciable amount of learning was left to occur. To investigate this possibility, the aim of this study was to investigate various methods of combining practice with demonstration for learning a discrete action pattern, the overhand volleyball serve. One group experienced 10 pre-practice demonstrations of an expert performing the discrete movement pattern prior to overt practice. A comparison group viewed a single pre-practice demonstration, followed by three practice attempts, then another demonstration, with this pattern of interspersing a demonstration every three practice attempts continued throughout acquisition. A third group receiving a modified combination of each schedule viewed five pre-practice demonstrations followed by five more demonstrations provided one at a time following every three practice attempts. Thus, modeling for this “combination” group was completed by mid-acquisition. Inclusion of this group was predicated on the expectation that a compromise between the two extremes of all-pre-practice and interspersed demonstration might take advantage of the benefits of each schedule.

All groups received the same frequency of videotape demonstrations in an acquisition phase. Participants received only pre-response information about appropriate form; that is, no form-related prescriptive feedback was given to preclude the possibility of confounding modeling effects with performance gains expected from receiving augmented feedback about form. Participants were, however, able to see the outcome of each serve; thus, feedback relative to achievement of the external goal was available on each attempt. With the exception of McGuire (1961), previous studies combining modeling schedules with overt practice have examined the “outcome” of performance as opposed to the “product” of performance (movement pattern form). Because the information resident in demonstrations pertains to form, both ratings of form and outcome were assessed in the present study to measure change due to modeling and overt practice. To determine if the effects due to modeling and overt practice persisted over time, an immediate no modeling retention test, and a 48-h no modeling retention test were administered.

It was hypothesized that the combination group would obtain higher form scores than the all-pre-practice or interspersed demonstration groups in acquisition and retention. This prediction was based on the notion that several pre-practice opportunities to see “what to do” combined with additional opportunities to enrich the memory representation with observational experiences early in overt practice would be most beneficial for immediate performance and retention of skill. It was also hypothesized that because all groups received outcome feedback on each trial, accuracy scores among groups would not differ significantly. That is, achievement of the external goal of the task was predicted to improve in each group as a function of overt practice with outcome feedback whether appropriate form was used or not. Improvement in accuracy as a function of practice with outcome feedback has been a long-reported phenomena in the motor learning literature. However, like many discrete action patterns, it was emphasized to the learners that recommended form would enhance the probability that the external goal would be achieved from trial-to-trial.

The participants were 30 right-handed undergraduate students (15 male and 15 female, M=20.2, S.D.=1.9) recruited from a general university population. To ensure unfamiliarity with the movement pattern, the overhand volleyball serve, participants were required to meet one of two inclusion criteria: either they had never performed the serve, or had not practiced the overhand serve in the previous three years. Participants signed an institutionally-approved informed consent form prior to participation. To randomly distribute pre-practice differences in skill proficiency, participants were randomly assigned to one of the three groups such that each group consisted of five males and females.

Animal evidence on the organization of frontal lobe and limbic structures and pathways and evidence from frontal lobe patients have led to cognitive models which attempt to explain human-motivated abstract reasoning abilities. These models relate motivated information processing to the activation of specific frontal brain areas and circuits that are further related to ongoing autonomic changes.

The Wisconsin Card Sorting Test (WCST) has been generally considered a prototype examination of abstract reasoning and frontal lobe function. In the WCST, the subject has to match geometrical shapes on a target card to four standard cards with different geometrical shapes. The subject has to find a rule on the basis of positive or negative feedback upon the response. When the rule has been found, application of this rule provides positive feedback until (after 10 trials) the rule is suddenly changed. From this point on, the sequence of rule search followed by rule application is repeated. Perseverative responses occur when a subject is unable to switch to a new rule, but continues to apply the old rule.

Evidence has established that subjects with frontal lobe damage, elderly adults, and Attention Deficit Disordered Hyperactive children perform relatively poorly on the WCST. These subject groups detect fewer rules, and exhibit more perseveration than control subjects. From experimental studies in patients with prefrontal lesions, Dubois et al. (1995) concluded that two dysfunctions emerge in these patients: (1) a disruption of rule finding, because of the emergence of non-cognitive, rigid patterns of behavior and (2) inadequate feedback integration in the selection process of the appropriate response.

Support, in particular, for the first dysfunction has come from PCA studies of WCST scores in frontal patients, schizophrenic patients and control subjects. These studies have consistently extracted a main factor, which explained most of the variance. Scores such as conceptual level response and number of correct categories had negative loadings, while perseverative error scores had positive loadings on this factor. Moreover, this factor clearly discriminated between control and frontal lobe patients, or schizophrenic patients. Interpretations such as ‘perseveration’, ‘abstract thinking ability’, ‘concept formation’, ‘undifferentiated executive function’ and ‘flexibility’ have been proposed for this PCA component.

Other studies have also indicated that a deficit in frontal executive functions may deteriorate WCST performance. Neurophysiological studies, using PET scan techniques, indicated that WCST problem solving engages, in particular, the dorsolateral prefrontal cortex . Also Ragland et al. (1997) reported more selective dorsolateral prefrontal and inferior frontal regional cerebral blood flow activation in good WCST performers than in middle and bottom WCST performers. This evidence suggests that WCST problem solving is related to activation in left frontal, left prefrontal, and left temporal cortical areas.

In the literature, the WCST has often been conceptualized as a working memory task that invokes specific intermediary operations, including switching cognitive sets according to changing contingencies, rule learning, formation of conceptual sets, application of detected concepts, maintaining sets, and use of error information. Close inspection of the test trials reveals that the WCST requires two distinct types of problem-solving strategies: rule search and (after the rule has been found) rule application for 10 consecutive trials. In rule search, the subject has to deduce the relationship between the standard and target cards. This demands a complex, ‘open ended’ set of processes, such as, the inhibition of an earlier rule, testing of alternative rules, and updating after feedback. In contrast, after the rule has been found, rule application only requires a sequence of relatively simple, pre-planned operations, the subject only matches one dimension of the geometrical shapes of the target and standard cards and ignores the other dimensions.

Theoretical models assume that in novel or complex tasks, such as the WCST, a supervisory controller co-ordinates the selection of the intermediary operations. According to Norman and Shallice (1986) a supervisory attentional system allows for conscious attentional control to modulate performance. It operates by the allocation of attentional resources, which add to the activation values of selected operations and decrease the activation values of earlier operations. Norman and Shallice suggested that WCST perseverative responses in frontal lobe patients result from a deficit in the allocation and supervision of attention. This model is consistent with results of frontal patients which indicated that attention to and integration of feedback information in these patients is inadequate. Also Dunbar and Sussman (1995) concluded from experiments with normal subjects that WCST perseveration may result from an inability to exert attentional control resulting in a failure to update feedback information. The theory of Norman and Shallice attributes a main regulatory frontal lobe function to attentional processes during WCST rule search. Moreover, it suggests a complementary process that is sufficient for relatively simple or well-learned acts: automatic contention scheduling. This mechanism involves the automatic execution of an action sequence (schema) that does not require attention. As argued above, WCST rule application demands a similar process. It requires the pre-planned execution of a series of processing steps which together result in the application of a rule. Following Norman and Shallice’s theory, this rule application process may be automatically executed and may not require attentional resources.

Based on more recent animal brain research, Gray (1995a, b) has proposed a system of frontal/septo-hippocampal pathways that controls the automatic, pre-planned execution of step-by-step motor programs: the behavioral approach system. When a comparator process triggers a mismatch, this ongoing behavior is inhibited by the activation of a behavioral inhibition system which also increments arousal and attention. From a related animal research perspective, Pribram (1991) has argued that attentional processes in the frontal lobes and in the limbic system are closely connected to habituation, familiarization and orientation. A novel stimulus or a reinforcing stimulus such as positive feedback may disrupt habituation and elicit an orienting response. Moreover, it is well known that the frontal lobes have output pathways that project to the brain-stem, to control autonomic physiological support for anticipated behaviors and orienting responses. Hence, Pribram (1991) argued that an important part of the cortically and sub-cortically elicited orienting response is a set of viscero-autonomic responses, such as brief heart rate change, skin conductance change and change in respiratory rate. His research suggested to us that WCST rule application and rule search might also elicit viscero-autonomic responses. Below we will argue on the basis of the literature how such viscero-autonomic responses, in particular cardiac responses, may be associated with WCST performance.

Age-of-acquisition (AoA) is cited increasingly as an important variable in verbal tasks, largely due to the work of Ellis and his co-workers. The notion that words learned earlier in life are faster to name than later-acquired words was first addressed experimentally by Carroll and White (1973), who examined picture-naming latencies. For a long time, however, the interest in AoA was limited to a few researchers, predominantly from the United Kingdom. The vast majority of researchers did not take the variable into account, and considered it as a confound of word frequency (in that earlier-acquired words tend to occur more frequently in adulthood). In a provocative article, Morrison and Ellis (1995) reopened the issue and reported that word frequency no longer affected word naming times when AoA was controlled for, whereas AoA kept on having a strong impact when word frequency was controlled for. On the basis of these findings, Morrison and Ellis concluded that all reported effects of frequency in lexical tasks may be AoA effects in disguise. Although subsequent studies have shown that Morrison and Ellis’s claim was too strong because combined effects of frequency and AoA on word naming latencies have been obtained, a growing number of researchers become convinced that AoA plays a basic role in lexical tasks.

The consensus that seems to have emerged from recent studies is that AoA is the critical variable in word production; this is the so-called phonological completeness hypothesis. In general, reference is made to Brown and Watson’s (1987) idea that early-acquired words are stored in their entirety within the phonological output lexicon, but that the representations of late-acquired words may be more fragmented. The extra time required to assemble the dispersed representation of late-acquired words would account for their slower naming speed. Two reasons are given for situating the AoA effect at the speech output stage. First, AoA is a significant variable in all tasks that require the production of a word to describe the presented stimulus (i.e., visual word naming, picture naming, word finding problems in aphasia), but is not always a significant variable in binary, manual decision tasks (e.g., object classification; see the following). Second, Gerhand and Barry (1998) found a strong effect of AoA on pronunciation duration, with a smaller and less reliable effect of frequency. In this task, participants were presented with spoken words, one at a time, and were requested to repeat each word 10 times as fast as they could while still pronouncing each word correctly and clearly. The time taken to repeat each word 10 times was measured by the experimenter.

Van Loon-Vervoorn (1989), however, suggested another possible origin of the AoA effect. According to her, the order of acquisition is the most important organisational principle of the semantic system, with the meanings of later-acquired concepts being built on those of earlier-acquired concepts. Empirical evidence for her position was provided by van Loon-Vervoorn (1989, Chapter 10). She used a discrete word-associate generation task to tap into the semantic system. In this task, participants are asked to say the first word that comes to their mind when seeing a stimulus word. The task has also been used by Chumbley and Balota (1984) and de Groot (1989) to assess the nature of the semantic system. Van Loon-Vervoorn presented 60 one-syllable words that allowed her to assess the independent effects of AoA, word frequency and imageability (IMA). She obtained a reliable effect of AoA (earlier-acquired words: RT=1440 ms; later-acquired words: RT=1681 ms), IMA (high=1445 ms; LOW=1677 ms), and no effect of frequency (high=1539 ms; LOW=1583 ms). On the basis of her findings, van Loon-Vervoorn (1989) concluded that AoA is a semantic variable rather than a lexical variable.

Van Loon-Vervoorn’s work has not been incorporated in the recent discussion on the importance of AoA, partly because it was published in Dutch but also because Morrison et al. (1992) had failed to obtain an AoA effect in a semantic task in which participants classified pictures of objects as naturally-occurring (e.g., apple) or man-made (e.g., anchor).

On the other hand, the possibility of a semantic origin of the AoA effect is appealing, because it would explain a number of findings. First, though there is an important, negative correlation between AoA and frequency, nearly all studies have reported a more pronounced correlation between AoA and other semantic variables. Rubin (1980), for instance, reported a correlation of ?0.40 between AoA and frequency, together with a correlation of ?0.59 between AoA and IMA. The same was true for Whaley (1978) who reported correlations of, respectively, ?0.52 and ?0.69. In both studies, factor analysis indicated that AoA loaded most on a semantic factor that included variables such as imagery, concreteness and number of meanings. Using a more objective AoA-measure (obtained by asking children of different ages to name line drawings), Morrison, Chappell & Ellis (1997) obtained a correlation of ?0.47 between their real AoA measure and the logarithm of the Cobuild frequency, compared to a correlation of ?0.55 between AoA and imageability.

A second finding that is in line with a semantic interpretation of the AoA effect is the robust AoA effect in object naming latencies, as picture naming requires not only the correct name to be produced but also semantic activation to connect the pictorial input with the verbal output. Third, a semantic interpretation of the AoA effect may explain why the AoA effect in oral reading of visually presented words seems to be particularly strong when naming latencies are long, because there has been some speculation that semantic variables may affect word naming times when these are long enough. Finally, AoA has a strong effect on lexical decisions times and since the work of Chumbley and Balota (1984) it is known that lexical decisions involve semantics. Morrison and Ellis (1995) and Gerhand and Barry (1998) provided a phonological explanation of this AoA effect by assuming that lexical phonology contributes to the word/non-word decision, but further research (Gerhand & Barry, 1999) has shown that the AoA effect in the lexical decision task remains significant when efforts are made to interfere with the phonological processing (such as using only pseudo-homophone non-words or using articulatory suppression).
View Within Article

On the basis of these considerations, it occurred to us that researchers may have rejected van Loon-Vervoorn’s semantic interpretation of AoA too rapidly. At least, it seemed worthwhile to investigate the importance of AoA for a number of semantic tasks, and see how these findings relate to more “lexical” tasks, such as word naming and lexical decision. To do so, we made use of a set of six lists of 24 words recently assembled by Brysbaert et al. (2000) and validated in a naming and a lexical decision experiment. These lists consist of three pairs of lists that differ in AoA, frequency or IMA, and are matched on the other variables. We used these stimuli in two semantic tasks: a discrete word-associate generation task (Experiment 1) and a “word with a definable meaning” vs. “given-name” classification task (Experiment 2). The former is a replication of van Loon-Vervoorn (1989); the latter has been inspired by Taft and van Graan (1998).

In reading, the eyes progress on the line of text with forward saccades of variable sizes. Two out of three words are fixated with a forward saccade, some words being skipped during a first eye pass. The position where the eyes initially land in a word is generally located between the beginning and the middle of the word, but in some instances it can be located at the beginning or end of the word.

During the last 30 years, a large amount of research has been devoted to understanding what determines the variability in both word skipping rate and initial landing sites in words. Results as a rule show that visuomotor factors associated with saccadic programming (such as the length of the next word and the distance from the word where the eyes are launched from) contribute largely to this variability. In contrast, linguistic factors such as the processing difficulty of the words located in parafoveal vision seem to participate to a smaller extent in the variability. As noted by Brysbaert and Vitu (1998), most studies have failed to find an influence of ease of processing a parafoveal word on the likelihood of skipping it, and when effects are reported, they are either small in amplitude or can be attributed to some other variables.

Concerning initial landing sites in words, Beauvillain, Doré and Baudoin (1996) reported a difference in initial landing sites of less than a letter between words that differ in the frequency of their initial letter sequence. On the other hand, Everatt and Underwood (1992) showed an initial landing difference as a function of where the uniquely identifying information in the word is located. But this effect could never be repeated, and it has been attributed to either artifacts of the eye tracking system or confounding variables such as the luminance of the letters. Finally, Dubois and Sprenger-Charolles (1988) reported an effect of semantic context on initial landing sites, when paired comparisons were made between sentences that were identical except for the target word (that was or was not semantically related to a preceding prime word). Since target words differed in both the visual attributes and frequency of occurrence in the language, conclusions can hardly be drawn as to the role of semantic factors.

At the same time, there is evidence in the literature that processing of parafoveal words occurs at least up to a lexical level. Both naming time and gaze duration (or total time the eyes spend on the word) are shorter when the word was visible in parafoveal vision before being fixated than when it was masked. In addition, the resulting parafoveal preview benefit is larger for high- than low-frequency words and for words that are predictable from the prior semantic and/or syntactic context. Whether there is a semantic preprocessing of the parafoveal words is still an open question.

Thus, although some information is extracted from words located in parafoveal vision, it does not seem to have a clear effect on where to move the eyes next. According to O’Regan (1990), the paradox results from both the slowness of parafoveal word identification processes and the use of an autonomous scanning strategy that is independent of these processes and relies only on low-level visual processes. The eyes could be guided through the line of text by a strategy attempting to drive the eyes from the center of one word to the center of the next word. Variability in both word skipping rate and initial landing sites then result mostly from low-level visuomotor factors that affect saccadic accuracy; the rare instances in which ongoing linguistic processes influence the eye landing could correspond to cases where the prior fixation duration is longer than usual.

This oculomotor-control view is challenged by a more cognitive view, that suggests that parafoveal word processing and the programming of the next saccade overlap in time, and that ongoing linguistic processes can modify the planned saccade. According to this view, all inter-word saccades initially aim for the center of the next word, but they deviate from the target when processing of the word is terminated before the saccade is computed. Thus, the likelihood that a saccade is influenced by ongoing processes then depends on the ease of processing associated with the parafoveal word.

The purpose of the present study was to distinguish between oculomotor- and cognitive-control views of eye movements in reading, and to determine the extent to which initial landing sites in words are sensitive to ongoing perceptual and linguistic processes. In particular, the present experiments tested whether the position where the eyes initially land in a word depends on the word’s predictability from a prior semantic context and whether the likelihood of observing such an effect is a function of the ease of processing associated with the parafoveal word. If context effects are obtained in conditions that facilitate parafoveal word processing, then this argues against the hypothesis that inter-word saccades are determined by an autonomous scanning strategy that is independent of ongoing linguistic processes. At the same time, this would provide further evidence for the case of semantic parafoveal word processing, and it would bring some insights on the question of where in the time course of word identification semantic processes intervene. This question has indeed been extensively debated in the last 20 years, and it is still not clear whether semantic context effects emerge before or after lexical access has been completed. As a saccade is programmed within 250 ms of an average fixation duration in reading, linguistic effects on saccades then suggest that semantic context effects can be shown at an earlier stage than in lexical decision and naming tasks, which take about 500 ms and which, in addition, involve post-access strategies.

In the present experiments, eye movement data were recorded in two experiments that were part of a larger study. Participants were presented isolated sentences with one or two semantically (un)related prime(s) preceding a target word. In both studies, conditions were selected to maximize the occurrence of early context effects (i.e., on initial landing sites). First, we used for related prime(s) and target words, pairs of words that were strongly associated. Second, we used both high- (Experiments 1 and 2) and low-frequency target words (Experiment 1 only). It has been shown in prior studies that high-frequency words are more likely to be preprocessed in parafoveal vision, and therefore they may enhance the occurrence or the amplitude of context effects. Third, we determined a posteriori the launch site (i.e., the last position of the eyes before the target word), since context effects may emerge only for close launch sites that favor parafoveal preprocessing. In addition, to ensure that context effects could not be the result of the target words having different visual attributes, pairs of, respectively, related and unrelated sentences contained the same target word, but differed by the preceding prime being used.

Participants. Eighteen students (11 from the University of Leuven, Belgium, and seven from Paris, France) who were between 18 and 30 years old, were paid to participate in the experiment at the University of Leuven, Belgium. They were all native French speakers (first language) and had normal or corrected-to-normal vision (only with glasses).

Material. Two sets of 72 sentences were constructed. Each sentence contained a prime and a target word. In half the sentences, both words were semantically related, and in the other half, both words were unrelated. Both related and unrelated sentences were matched except for the prime word which was either semantically related to the target or not. The related and unrelated primes were matched in length up to a two-letter difference.

Facial attractiveness plays a major role in the formation of interpersonal social judgements. Many studies showed that attractive individuals enjoy advantages over less attractive people. They are perceived as more likeable, kind, and intelligent and are more likely to be professionally successful than their less attractive counterparts. Even children have been shown to be differentially evaluated on the basis of their physical appearance, both by peers and by adults. Facial attractiveness has been demonstrated to influence the status within the peer group as well as evaluation of school performance by teachers.

Such biases in favor of attractive individuals were for a long time thought to be based on prevalent socio-cultural norms. Judgements of attractiveness were viewed as depending largely on fashion and the underlying mechanisms were viewed as being only gradually acquired through internalization of prevalent socio-cultural stereotypes in the course of individual development. Over the past 10 years, however, evidence has accumulated which suggests that the perception of facial attractiveness may be remarkably similar across both different cultures and different age groups. The results of studies examining the influence of the beholder’s age on the perception of facial attractiveness in fact indicate that already children as young as three months old are able to discriminate between rather attractive and unattractive faces. The perception of facial attractiveness, therefore, seems to have a sizeable biological basis and to be not an arbitrary product of social norms. But while matters seem fairly clear at the extremes of the attractiveness continuum, it is much less clear to what degree people at different ages differ in the extent to which they perceive differences between less extreme stimuli and to what degree the type of stimulus and experiential or maturational factors play a role in mediating this ability.

Recent work on the neural mechanisms underlying face processing abilities indicates that maturational factors indeed play a role in several aspects of face processing. For instance, Taylor, McCarthy, Saliba and Degiovanni (1999) demonstrated a gradual maturation for a face-specific electrophysiological component throughout childhood and adolescence. Their study seems remarkable in that it did not involve a specific task which may have confounded face perception with other activities such as recognition or recall but simply compared neural responses when viewing faces with neural responses to other classes of objects in different age groups. On the basis of their result the authors suggest that face processing undergoes a gradual, quantitative maturation but no qualitative change during individual development.

With regard to the processing of facial expression, Kolb, Wilson and Taylor (1992) found that performance levels of eight- to thirteen-year-old children in an expression matching task was similar to the performance of adults with frontal lobe injury, implying that some frontal lobe regions involved in this task may not yet have matured in this age group. In this context, data implicating the left frontal lobe in judgements of facial attractiveness (Nakamura et al., 1998) provide circumstantial evidence to suggest a role of developmental factors in the perception of facial attractiveness, as the frontal lobes are areas in the human brain which are subject to great modifications in the course of individual development.

In cognitive research into the development of face recognition it is widely accepted that faces are a special class of visual objects from the very beginning in that they are preferred by infants over non-faces. However, it is also clear that there is a considerable amount of development involved in the ability to discriminate and recognize faces. A two-process theory has been put forward to account for both these observations in infants and a considerable amount of work has been devoted to investigating the trajectory of face recognition abilities through early and middle childhood into adulthood. The influence of factors such as distinctiveness and inversion on the ability to discriminate and recognize faces at different ages has been examined. The results indicate that children’s face recognition is less disrupted by the inversion of faces than is adults’, that children may be less able to use distinctiveness information in face recognition tasks, and, not surprisingly, that they are less able to correctly identify photographs of a person taken 20 years apart and are more susceptible to being deceived by irrelevant paraphernalia when performing a face recognition task. To account for such observations, Carey (1992) suggested that some of the age-related differences in face processing may be due to the fact that children’s face processing may rely more on isolated features of a face and may tend to disregard the configurational information present in a face. Ellis and Flin (1990), on the other hand, propose that the developmental changes observed are due to the fact that older individuals are simply able to extract more information from a face within a given amount of time.

In view of the above presented evidence for the presence of developmental factors in various aspects of face processing, the comparative paucity of work dealing with the development of mechanisms underlying the perception of facial attractiveness in adults and children at different ages is surprising. While there is a considerable amount of work on infants’ basic abilities and extensive research on the social consequences of an individual’s facial attractiveness, very little is known about the extent to which children’s and adults’ preferences are the same and the extent to which such preferences are modified by the types of stimuli used.

From a developmental point of view it appears possible that, whereas children may prefer faces rated as very attractive by adults over faces rated as quite unattractive, the children’s preferences may be less pronounced than those of the adults, i.e., they may have difficulty perceiving as fine differences in facial attractiveness as adults do. Prior work did not address this point and rather focused on the question of whether children are basically able to make distinctions based on attractiveness, and whether these distinctions are in the same direction as the ones adults make. There is ample evidence that this is already the case at a very early age. But whether the preferences of children are already as pronounced as those of adults is largely unclear. On the basis of the above mentioned neurophysiological data one might expect more pronounced preferences in older subjects on purely brain maturational grounds which, of course, will also be correlated with the experiential factors emerging from cognitive studies. Furthermore, stimulus attributes may play a role in mediating children’s ability to discriminate between attractive and less attractive faces. For example, children may perceive a greater variation in children’s faces than in adults’ faces. This might either be the case because such faces are of greater social importance to them than adults’ faces as attractiveness represents a salient category for social comparison with peers, or because school-age children have comparatively more every-day experience with this type of face. Such an “own-age effect” might be analogous to the well-known “own-race bias”, where it has been demonstrated to be easier for people to recognize members of their own race over members of a different race. With regard to “own-race” effects in the perception of facial attractiveness, recent research has also shown that, while there are great overall consistencies across cultures, the within-culture agreement tends to be somewhat larger than the cross-cultural one. On the other hand, adults may, by virtue of their greater overall experience with all kinds of faces and because of more efficient encoding strategies, perceive a greater variability of attractiveness in both adults’ and children’s faces.

Cognitive psychologists have now studied the response performance of children with attention-deficit/hyperactivity disorder (ADHD) on a wide variety of information processing tasks (for a comprehensive review see Douglas, 1999). In conjunction with the fact that ADHD children often make more errors than control children, the most consistent finding in the ADHD cognitive literature is that the overall response times of ADHD children are typically both slower and more variable than those of control children. However, as noted by Douglas (1999), because much of the theoretical and empirical research involving ADHD children has focussed on more specific task manipulations, these two pervasive phenomena of slow and highly variable ADHD response times have still not been adequately addressed.

This study will directly address these phenomena through a detailed statistical examination of the actual distributions of response times obtained from ADHD and control children within a four-choice warned reaction time (4C-WRT) task conducted at McGill University. In this task, groups of ADHD and age-matched control boys provided spatially compatible responses indicating which one of four, highly discriminable, stimuli had been presented in a stimulus display. On each choice trial, the relevant stimulus appeared at the conclusion of either a 2-, 4-, or 8-s fixed foreperiod (also known as a preparatory interval [PI]) that was marked by the presentation of an initial warning signal. In addition, parallel data were collected from a group of control boys several years younger than the ADHD and age-matched control boys to also determine whether the response time distributional profiles of the ADHD boys simply reflect an immature pattern of responding typical of younger children.

We were motivated to undertake an investigation of ADHD response time distributional data for several reasons. First, it is our view that dysfunctional regulatory or control processes are responsible for the performance problems associated with ADHD. In her recent review of the empirical ADHD findings, Douglas (1999) points out that the high degree of variability in ADHD performance on many cognitive tasks seems to signify a pervasive manifestation of regulatory problems involving the inconsistent allocation of effort. Hence, the acquisition of a better understanding of the nature of this variability represents an essential scientific step towards a determination of the precise role of this aspect of regulatory or control processing in ADHD.

Second, upon a closer examination of these types of data, we are continually struck by the fact that the response time distributions of ADHD children can typically be distinguished from those of control children more by the presence of a substantially larger number of abnormally slow responses than by an overall pattern of slow responses, who reported experimental work in which suboptimal processing conditions mainly affected the right end of the response time distributions of hyperactive children). In other words, the standard positive skewing, or asymmetry, that is typically present in the response time distributions obtained within almost all psychological research paradigms is highly exaggerated in the response time distributional profiles of ADHD children. Statistically, positive skew leads to a number of extreme values that have a disproportionate influence on the calculation of the response time mean and, similarly, on the size of the variance measure. We believe that this skew is an important empirical marker that reflects the presence of periodic attentional ‘lapses’ in the responding of ADHD children, in contrast to a general inability to respond quickly. Given the obvious nature of the potential theoretical implications of such a phenomenon, it is important to consider ways in which it might be measured in a quantitative fashion that then allows it to be subjected to a rigorous statistical analysis.

Third, there is a growing recognition that a quantitative study of the shapes of empirical response time distributions can provide much more information from a set of response time data than that which is given by the more standard statistical summary measures of the mean and the variance. That is, response time distributional analyses can be used to describe psychological performance at a more fine-grained level than those standard measures and, thus, can also provide a fuller set of empirical constraints against which to evaluate any existing psychological theories for the underlying process(es) in question. Finally, the recent availability of a statistical package now allows cognitive researchers to easily obtain quantitative summary measures of the shapes of response time distributions (in terms of the three parameters of the ex-Gaussian distributional model).

Our foray into response time distributional analyses has proven fruitful. In this article, it will be demonstrated that these analyses lead to the identification of one specific aspect of the ADHD distributional data that we believe uniquely characterizes the responding of ADHD populations in these types of cognitive psychological tasks; so much so, that this aspect will be shown to be highly diagnostic of ADHD in this sample of boys. Moreover, it will also be established that the two phenomena of slow and highly variable ADHD response times are intimately coupled, in that both can be explained mainly in terms of this same aspect of the ADHD response time distributions. Finally, the additional information obtained from these analyses will also show that, unlike either the response time means or standard deviations, the overall distributional pattern of ADHD responding can be dissociated from that of the younger control responding.

The ex-Gaussian distributional model can be used to provide useful quantitative measures of the distributional properties of a set of response times. This model assumes that each response time can be represented as the sum of a normally distributed random variable and an independent exponentially distributed random variable, and therefore, that the full distribution of response times can be characterized as a convolution of the normal and exponential distribution functions. Parametrically, the ex-Gaussian distribution has three constituents: mu (?) and sigma (?), that, respectively, describe the mean and standard deviation of the normal component, and tau (?), that describes the mean of the exponential component. Fig. 1 shows the probability densities of two ex-Gaussian distributions, along with their normal and exponential distributional components. Each of the two ex-Gaussian distributions in Fig. 1 have identical normal components but differ with respect to their exponential components.

Fig. 1. Probability density (p.d.) functions for two normal distributions with ?=500 ms and ?=75 ms ((a), (b)), two exponential distributions with ?=150 ms (a) and ?=350 ms (b), and the two resultant ex-Gaussian response time (RT) distributions.

When manually intercepting moving targets, it has been shown that the characteristics of the target motion influence how the arm reaches toward the target. In computer simulations of reaching van Donkelaar, Lee and Gellman (1992) have shown that the duration of a reaching movement is shorter when the target is moving faster, both in a full vision condition, and when vision of the hand is restricted. This relationship between interception speed and target speed was also demonstrated by Smeets and Brenner (1995) in a study where participants were required to hit targets moving across a computer screen with a handheld rod, as quickly as possible. In spite of the instruction to hit all the targets as quickly as possible, it was found that fast targets were hit with a higher velocity than slow targets. In another study, when generating reaching and grasping movements to a three-dimensional target rolling down a ramp, the speed of the reach has also been shown to increase as target speed increases, and when the target was moving very slowly (<40 mm/s), participants reached even more slowly than they did for a stationary control target. Conversely, when participants self-selected the location where they would intercept a moving target, their aiming speed decreased as the target speed increased. However, this was probably due to the observation that participants grasped the faster moving targets closer to their body. The shorter movements and related lower peak velocities that were observed when intercepting the fast targets are consistent with the predictions of Fitts’ Law.

While this relationship between the speed of the interceptive motion and the speed of the moving target exists, it has not been clearly established which characteristics of the target motion are most important. As a target moves faster between two points, two movement characteristics can change. Not only is the velocity of the faster target greater, its movement time is shorter. It is not clear whether both of these two target motion characteristics can influence the interceptive movement. One of the most commonly cited models to explain the visuomotor control of target interception has been the Tau model originally proposed by Lee (1976). While there have been many variations or updates to the model over recent years, the basic premise of this model is that a time variable (time to contact) is used to trigger the appropriate interceptive action. For example the Tau model has been shown to model actions such as jumping up to intercept a falling volleyball, or the hitting of a table tennis ball. This model does not take into account the changes in the velocity characteristics of the target, only its temporal characteristics. However, in a review by Bootsma, Fayt, Zaal and Laurent (1997) the possibility is entertained that when intercepting accelerating targets a second-order variable may be used.

In a recent study in which participants intercepted targets on a computer screen, targets moved either at a constant acceleration, a constant deceleration, or at a constant velocity. It was proposed that the response time of the interceptive movement is composed of a constant processing time, and the time required for the stimulus to travel a threshold distance, thus response time is velocity dependent. It was also suggested that some participants use an estimate of time to contact to determine when to initiate their interceptive movements. It was further shown that the interceptive movements toward slow-moving targets were produced by generating a series of sub movements that were produced to keep the displacement of the hand proportional to the target position and that this process is independent of the acceleration of the target. Thus, these authors have shown that participants are not able to utilize a higher order variable (acceleration) to control their interceptive actions, but instead, base the initiation and production of their interception movements on the temporal characteristics of the moving target.

It might not be surprising that individuals in the Lee et al. (1997) study were not able to successfully utilize target acceleration information. The general consensus in the psychophysical literature is that accelerations cannot be directly perceived from target motion. Instead it has been widely argued that target accelerations are computed indirectly from the comparisons of two instantaneous velocities acquired across a specific length of time. However, Rosenbaum noted that “the perceptual system is primarily responsive to changes in stimulation, (and) it is noteworthy that acceleration involves more changes than does constant velocity”. Thus, our visuomotor system could be viewed as one that searches for the optimal set of information that will allow it to perform the required task. Most objects in real life move with non-constant velocity profiles. Therefore it would be beneficial for the visuomotor system to be able to retrieve and use visual information about changes in object velocity to produce successful interception movements. While the majority of the evidence collectively suggests that only temporal information about target motion, or information about instantaneous velocity, is used to guide interceptive actions, in the present study, changes in the velocity characteristics of the targets were manipulated.

The purpose of the present study was to examine if the changing velocity characteristics of a target have any influence on the temporal and kinematic characteristics of manual interceptive movements. Participants intercepted targets with common movement times and magnitudes of peak velocity. However, the temporal properties of the target’s trajectory were varied across conditions. If participants are able to use information about the changing velocity of the target’s trajectory then it is expected that the kinematics of the interception movement will vary as a function of target velocity condition. A related question is how fast can that information be retrieved from the trajectory of the moving target? As well, how much of the information is needed to precisely intercept the target and when is it needed? The next two experiments were designed to investigate these issues.

For tasks where the target disappears for more than one second before contact, the interception hand trajectory cannot be controlled using continuous time to contact information, but rather the visuomotor system must rely on an internal model or representation of the future target trajectory, built during the initial planning phase. This view was experimentally supported by Whiting’s (1968) study of baseball catching. With decreased amounts of visual information, the author concluded that constant visual monitoring of the ball flight was not necessary for precise interception. Whiting showed that performance did not decrease significantly with a decreased amount of visual information about the ball trajectory. Based on these findings, one can suppose that the visuomotor system is able to create an internal model of the target trajectory and use this model to produce successful manual interception. In such a case, the performance of the interception task should suffer significantly when the visual information about the target trajectory is removed from the planning period. This hypothesis was tested by Whiting and Sharp (1974) where participants were required to catch tennis balls in a dark room. For a brief period of time the ball was visible to the participant during its flight. It was found that when the visual information about the ball trajectory was blocked until very late in the flight, the catching performance suffered, emphasizing the impact of visual information early in a target’s trajectory.

During their career people go through a professional socialization process in which they learn the theories of their field and adopt its metaphysical assumptions. This process defines for the members of the profession the phenomena that are worthy of observation, and the type of information that is relevant to their professional judgments. While this process is an integral part of the development of professional skills, it may also lead professionals to base their judgments on perceptions derived from the dominant theories in their field, and ignore, to some extent, important relationships among variables in the environment. This phenomenon has been demonstrated in domains such as managerial decision making and scientific research.

For many years, an important characteristic of the professional socialization process of clinical psychologists has been an emphasis on the psycho-pathological aspects of the mind. The training of clinical psychologists concentrated on issues such as the origins of psycho-pathology, its development, and its remedy; it focused on the diagnosis of the pathological rather than on the identification of the benign; and it centered on the study of the deviant rather than the understanding of the normal. Furthermore, to a certain extent, even normal behavior were often understood by clinical psychologists to be the result of unconscious pathological aspects of the mind, such as murder impulses, incestuous fantasies, and death wishes.

Did a socialization process that emphasize pathology influence the professional judgment of clinicians? A number of studies which have dealt with this question concluded that the answer is positive. Rosenham’s (1972) paper entitled ‘On being sane in insane places’ is the most well-known. In this paper, Rosenham reports that behaviors, which would otherwise appear normal, were judged as pathological by the staff of psychiatric hospitals. However, Rosenham’s results could be attributed to initial false information which was supplied to the judges in his study, or to the high base-rate probability of pathology for inmates of psychiatric hospitals.

The current paper examines the hypothesis that in making clinical judgments, psychologists assigne excessively heavy weight to information regarding the presence or absence of severe pathology (which will be labelled pathological information), in comparison to information regarding the presence or absence of mild pathology (which, within the context of the current study, could be labelled non-pathological information). The paper presents evidence suggesting that such biased weighing influence clinical judgment, and explores the outcomes of this bias. Finally, the cognitive processes underlying this bias are discussed in terms of a confirmation bias in hypothesis testing – the tendency to overemphasize information confirming, rather than disconfirming, expectations; and it is suggested that clinicians’ confirmatory hypothesis are associated with the existence of severe, rather than mild, pathology.

The data used in the paper were collected in the mid-1950s by Meehl. The analysis of these data played a major role in the study of the validity of clinical judgments. In these studies, researchers were primarily interested in the actuarial validity of the judgment, and in particular, in whether clinical judgments have a higher correlation with the criterion than the predictions of a linear model. The approach of the current paper to the study of the validity of clinical judgment is different. The paper focuses on analyzing and explaining biases, or systematic deviations, from optimal actuarial validity. As a result, the method used in the paper is also different from the method used in the previous studies. While in previous studies the validity of clinical judgment was studied by correlating judgment with the criterion, in the current study it is studied by comparing models of the judgment to models of the criterion. In particular, the study compares the weight of information associated with severe pathology to the weight of information associated with mild pathology in these models.

Meehl’s data include 861 MMPI profiles of psychiatric patients – the patients’ scores on the 11 most commonly used scales of the MMPI. They also include the criterion – the diagnosis given to the patient in the clinic in which he/she received treatment. 47% of the patients were diagnosed as psychotics and 53% diagnosed as neurotics. These diagnoses were based primarily on information about the patient’s past and present behavior, and, to a certain extent, on the results of various psychological tests. For some of the patients, the information on which the diagnosis was based did not include the MMPI profiles, but for some, the MMPI profiles were available when the clinic’s diagnosis was made.

The data also include evaluations of the 861 profiles that were made by 29 clinicians, whose schooling represented a wide variety of approaches to Clinical Psychology at the time Meehl’s experiment took place (Meehl, personal communication). Each clinician judged the 861 MMPI profiles on an 11-step forced normal distribution scale from least psychotic (1) to most psychotic (11). The clinicians were instructed that the patients could be either psychotics or neurotics.

One aspect of the data which is particularly important to the current study is that the MMPI scales of the 861 profiles have a clear dimensional organization, for the results of a factor analysis of the scales.) One dimension is associated with the neurotic scales of the MMPI, another with the psychotic scales, and a third with scales that identify defensiveness in test taking. These dimensions, and in particular the neurotic and the psychotic dimensions, were likely to have played an important role in the process by which the clinicians used the MMPI profiles in their diagnostic judgments in Meehl’s experiment.

The criterion was modeled using a logistic regression, where Y is the log of the odds of having been diagnosed as psychotic by the clinic. The judgment was modeled in three ways. First, by modeling the mean judgment, which was created by averaging the judgments of all 29 clinicians to each profile. Second, by modeling the judgments of each of the 29 judges. And third, by modeling a binary variable, called the MMPI diagnosis, created by rank ordering the profiles according to their mean judgment, and labeling the top 47% as having diagnosis of psychosis and the bottom 53% as having a diagnosis of neurosis. The MMPI diagnosis represents the diagnosis that would have been assigned to the patients based on a consensus judgment of the MMPI profiles, and it could be compared to the criterion, or the clinic’s diagnosis, which was made primarily on the basis of actual behavior. Although some loss of information is involved in using the MMPI diagnosis rather than the untransformed judgment, I discuss the results primarily in terms of this binary variable, because it makes the judgment directly comparable to the criterion. However, the results of the models of the untransformed judgments are reported as well.