A general model for the analysis of multilevel data

A general model is developed for the analysis of multivariate multilevel data structures. Special cases of the model include repeated measures designs, multiple matrix samples, multilevel latent variable models, multiple time series, and variance and covariance component models.We would like to acknowledge the helpful comments of Ruth Silver. We also wish to thank the referees for helping to clarify the paper. This work was partly carried out with research funds provided by the Economic and Social Research Council (U.K.).     

Promoting imitative play through generalized observational learning in autisticlike children

Three autisticlike boys with deficits in imitative and appropriate play behaviors were taught to play with an unfamiliar toy by observing a peer model demonstrate appropriate play with the toy. Generalization was assessed across toys and models, and temporal maintenance was assessed at 1 and 3 weeks following generalization. Following training each child demonstrated increases in imitative play concomitant with decreases in the frequency of nonmanipulated selfstimulatory behavior. Results are discussed in terms of their implications for the amelioration of social isolation and self-stimulatory behavior.The authors wish to acknowledge the help and cooperation of the Henry Wiseman Kendall Center and the Highpoint Preschool Enrichment Program.     

Network models for social influence processes

This paper generalizes thep* class of models for social network data to predict individual-level attributes from network ties. Thep* model for social networks permits the modeling of social relationships in terms of particular local relational or network configurations. In this paper we present methods for modeling attribute measures in terms of network ties, and so constructp* models for the patterns of social influence within a network. Attribute variables are included in a directed dependence graph and the Hammersley-Clifford theorem is employed to derive probability models whose parameters can be estimated using maximum pseudo-likelihood. The models are compared to existing network effects models. They can be interpreted in terms of public or private social influence phenomena within groups. The models are illustrated by an empirical example involving a training course, with trainees' reactions to aspects of the course found to relate to those of their network partners.This research was supported by grants from the Australian Research Council. The authors would like to acknowledge the help of Stanley Wasserman, Janice Langan-Fox and Larry Hubert, and would like to thank four anonymous reviewers for their helpful comments. Earlier versions of this article were presented at the North American Conference of the Psychometric Society, Lawrence, Kansas, June, 1999, and at the Australasian Mathematical Psychology Conference, Brisbane, Australia, December 1999.     

The need for teaching philosophy in medical education

The dearth of philosophical contributions to medicine has recently been discussed in a series of articles in this journal. The present article focuses on physicians' lack of training in philosophy as a part of the explanation of the scarcity of works in philosophy of medicine. In section I I outline two philosophy courses which would be reasonable additions to the medical school curriculum required of all medical students. In section II I suggest two other philosophy courses as electives in a medical education. All four courses are in the fields of epistemology and metaphysics, and so will help others see the relevance to medicine of philosophical fields other than ethics.     

Might “unique” factors be “common”? On the possibility of indeterminate common–unique covariances

The present paper shows that the usual factor analytic structured data dispersion matrix Λ Ψ Λ′ + Δ can readily arise from a set of scores y = Λ η + ε, where the “common” (η) and “unique” (ε) factors have nonzero covariance: Γ = Covε,η) ≠ 0. Implications of this finding are discussed for the indeterminacy of factor scores, and for the issue of invariance of factor analytic covariance models. The size of the problem is explored with numerical examples. I would like to acknowledge the large amount of effort and stimulating input supplied on the previous drafts of this paper from the reviewers, Associate Editor, and Editors of Psychometrika. Particular thanks go to William Meredith for his assistance with the final draft. Requests for reprints should be sent to Dave Grayson, 14 Poplar Grove, Lawson, NSW 2783, Australia     

Taï chimpanzees anticipate revisiting high-valued fruit trees from further distances

The use of spatio-temporal memory has been argued to increase food-finding efficiency in rainforest primates. However, the exact content of this memory is poorly known to date. This study investigated what specific information from previous feeding visits chimpanzees (Pan troglodytes verus), in Taï National Park, Côte d’Ivoire, take into account when they revisit the same feeding trees. By following five adult females for many consecutive days, we tested from what distance the females directed their travels towards previously visited feeding trees and how previous feeding experiences and fruit tree properties influenced this distance. To exclude the influence of sensory cues, the females’ approach distance was measured from their last significant change in travel direction until the moment they entered the tree’s maximum detection field. We found that chimpanzees travelled longer distances to trees at which they had previously made food grunts and had rejected fewer fruits compared to other trees. In addition, the results suggest that the chimpanzees were able to anticipate the amount of fruit that they would find in the trees. Overall, our findings are consistent with the hypothesis that chimpanzees act upon a retrieved memory of their last feeding experiences long before they revisit feeding trees, which would indicate a daily use of long-term prospective memory. Further, the results are consistent with the possibility that positive emotional experiences help to trigger prospective memory retrieval in forest areas that are further away and have fewer cues associated with revisited feeding trees.     

Do medical decision analyses' largest gains grow from the smallest trees?

Some decision analysts recommend reserving their models for careful, complete study of complex problems. Others believe simple problem analyses provide the greatest gains. To clarify these issues, I tested: (1) whether simple published analyses compared to complex ones, produced results as non-intuitive; and (2) whether non-intuitiveness represented analytic gain (or equivalently, intuitive loss). Very limited data also examined (3) the gain from adding model intricacy for a given problem vs. selecting more complex problems for analysis. To assess two non-intuitiveness measures, 75 subjects (33 physicians and 42) non-physicians stated their intuitive preferences in 40 (22 published and 18 unpublished) medical dilemmas. For both physicians and non-physicians, simpler models (those with lees than 30 terminal branches in the decision tree) had about one third the non-intuitiveness of more complex ones. Three tests also supported the premise that analyses outperform intuition, therefore that on average the formers' non-intuitiveness reflects less analytic gain. In addition, for different models of the same problem, tree complexity did not correlate with gain. Thus, simpler trees may not generally gain less because they inadequately describe problems. Instead, simpler analyses may represent simpler problems, with more similar intuitive and formal solutions. If so, these findings may help us avoid some costly but unnecessary simple analyses.     

Free trees and bidirectional trees as representations of psychological distance

Two types of graph-theoretic representations of psychological distances or dissimilarities are proposed: weighted free trees and weighted bidirectional trees. A weighted free tree is a generalization of the sort of graph representation used in hierarchical clustering. A weighted bidirectional tree is a further generalization which allows for asymmetric dissimilarities. The properties of these structures are discussed and numerical methods are presented that can be used to derive a representation for any given set of dissimilarities. The applicability of these structures is illustrated by using them to represent the data from experiments on the similarities among animal terms and on memory for sentences. An analysis based upon free trees is compared and contrasted with analyses using hierarchical clustering and multidimensional scaling.     

The Inherent Empirical Underdetermination of Mental Causation

It has become a popular view among non-reductive physicalists that it is possible to devise empirical tests generating evidence for the causal efficacy of the mental, whereby the exclusion worries that have haunted the position of non-reductive physicalism for decades can be dissolved once and for all. This paper aims to show that these evidentialist hopes are vain. I argue that, if the mental is taken to supervene non-reductively on the physical, there cannot exist empirical evidence for its causal efficacy. While causal structures without non-reductive supervenience relations can be conclusively identified in ideal discovery circumstances, it is impossible, in principle, to generate evidence that would favour models with mental causation over models without. Ascribing causal efficacy to the mental, for the non-reductive physicalist, is a modelling choice that must be made on the basis of metaphysical background theories or pragmatic maxims guiding the selection among empirically indistinguishable models.     

Elusive Counterfactuals

I offer a novel solution to the problem of counterfactual skepticism: the worry that all contingent counterfactuals without explicit probabilities in the consequent are false. I argue that a specific kind of contextualist semantics and pragmatics for would‐ and might‐counterfactuals can block both central routes to counterfactual skepticism. One, it can explain the clash between would‐ and might‐counterfactuals as in: (1) If you had dropped that vase, it would have broken. and (2) If you had dropped that vase, it might have safely quantum tunneled to China. Two, it can explain why counterfactuals like (1) can be true despite the fact that quantum tunneling worlds are among the most similar worlds. I further argue that this brand of contextualism accounts for the data better than other existing solutions to the problem.     

Logic-based XML data integration: a semi-materializing approach

We first describe the approach to XML data integration in the XPathLog/LoPiX project that uses a warehouse strategy. We show that the DOM model and the XML Query Data Model are not suitable for this task since the integrated database is not necessarily a tree, but a set of overlapping (original and integrated) trees. The problem is solved by using a node-labeled graph-based data model, called XTreeGraph, for the internal XML database that represents multiple, overlapping XML trees, or tree views.In the second part, we return to the standard XML data model—by still keeping the overlapping tree idea by “simulating” it: The data is internally represented by XML where the “overlayed” resulting tree is represented by XLink elements that refer to the original sources. By using a logical, transparent data model for XLinks as investigated in [May, in: 11th WWW Conference, 2002], all queries behave as stated against the XTreeGraph. The use of links for partial materialization also turns the approach from a warehouse approach into a mixed approach that combines the advantages of the warehouse approach and of the virtual approach. The approach is again illustrated by using XPathLog as data integration language.     

Reward dominance: Associations with anxiety,conduct problems,and psychopathy in children

The associations between children's behavior and their performance on a task with a steadily increasing ratio of punished to rewarded responses was investigated in a group of clinic-referred (n= 92) and normal control (n= 40) children between the ages of 6 and 13. Clinic-referred children with an anxiety disorder played significantly fewer trials than clinic-referred children without an anxiety disorder but the response style of the anxious children did not differ from that of a normal control group. Children with severe conduct problems who had no anxiety disorder played more trials than (a) children with severe conduct problems and a comorbid anxiety disorder, (b) nonanxious children with attention-deficit hyperactivity disorder, and (c) children in the normal control group. The strongest evidence for the reward dominant response style was for nonanxious subjects with elevations on a measure of psychopathic features, irrespective of whether they also had conduct problems and irrespective of whether they were clinic-referred. We would like to thank the staff of the Alabama School-Aged Assessment Service and the Tuscaloosa County School System for their help in data collection. We would like to especially acknowledge the help of John Gurley in data collection.     

Parsing to Learn

Learning a language by parameter setting is almost certainly less onerous than composing a grammar from scratch. But recent computational modeling of how parameters are set has shown that it is not at all the simple mechanical process sometimes imagined. Sentences must be parsed to discover the properties that select between parameter values. But the sentences that drive learning cannot be parsed with the learner's current grammar. And there is not much point in parsing them with just one new grammar. They must apparently be parsed with all possible grammars, in order to find out which one is most successful at licensing the language. The research task is to reconcile this with the fact that the human sentence parsing mechanism, even in adults, has only very limited parallel parsing capacity. I have proposed that all possible grammars can be folded into one, if parameter values are fragments of sentential tree structures that the parser can make use of where necessary to assign a structure to an input sentence. However, the problem of capacity limitations remains. The combined grammar will afford multiple analyses for some sentences, too many to be computed on-line. I propose that the parser computes only one analysis per sentence but can detect ambiguity, and that the learner makes use of unambiguous input only. This provides secure information but relatively little of it, particularly at early stages of learning where few grammars have been excluded and ambiguity is rife. I consider three solutions: improving the parser's ability to extract unambiguous information from partially ambiguous sentences, assuming default parameter values to temporarily eliminate ambiguity, reconfiguring the parameters so that some are subordinate to others and do not present themselves to the learner until the others have been set. A more radical alternative is to give up the quest for error-free learning and permit parameters to be set without regard for whether the parser may have overlooked an alternative analysis of the sentence. If it can be assumed that the human parser keeps a running tally of the parameter values it has accessed, then the learner would do nothing other than parse sentences for comprehension, as adults do. The most useful parameter values would become more and more easily accessed; the noncontributors would drop out of the running. There would be no learning mechanism at all, over and above the parser. But how accurate this system would be remains to be established.     

Replies to Deng,Lee, and Skow

This paper is a contribution to a book symposium on my book Experiencing Time. I reply to comments on the book by Natalja Deng, Geoffrey Lee and Bradford Skow. Although several chapters of the book are discussed, the main focus of my reply is on Chapters 2 and 6. In Chapter 2 I argue that the putative mind-independent passage of time could not be experienced, and from this I develop an argument against the A-theory of time. In Chapter 6 I offer one part of an explanation of why we are disposed to think that time passes, relating to the supposedly ‘dynamic’ quality of experienced change. Deng, Lee, and Skow’s comments help me to clarify several issues, add some new thoughts, and make a new distinction that was needed, and I acknowledge, as I did in the book, that certain arguments in Chapter 6 are not conclusive; but I otherwise concede very little regarding the main claims and arguments defended in the book.     

Tree inference with factors selectively influencing processes in a processing tree

We show how to systematically construct a processing tree using data from an experiment in which two experimental factors are intended to selectively influence two different processes. Only two arrangements are possible for the two selectively influenced processes: ordered and unordered. We show that there are only two standard trees that need be considered, one for each arrangement. Necessary and sufficient conditions are given for each tree to be applicable. If the conditions do not hold, no tree is possible. If the selectively influenced processes are ordered, their order is sometimes determined by the data. We consider two factors that selectively influence two processes in every tree in a set of trees, such that on a given trial, a tree is selected from the set at random. We show that if the influenced processes are ordered in every tree or unordered in every tree, then the mixture of trees is equivalent to one of the two standard trees. To illustrate the method, a tree is constructed from simulated data from a model of Batchelder and Riefer [Batchelder, W.H., Riefer, D.M. (1980). Separation of storage and retrieval factors in free recall of clusterable pairs. Psychological Review, 87, 375–397].     

Argument practices

The move to Postmodernism in argumentation is often predicated on the rejection of the formal basis of argument in logic. While this rejection may be justified, and is widely discussed in the literature, the loss of logic creates problems that a Postmodern theory of argument must address without recourse to logic and its attendant modernist assumptions. This essay argues that conceiving of argument in terms ofpractices will address the key problematics of Postmodernism without abandoning those features of argumentation that make it an important social, cultural and political practice. Implications for both theory and pedagogy are discussed.The author would like to acknowledge the invaluable commentary of, and discussions with, John Lyne, Charles Willard, Steven Fuller, and Ed Schiappa.     

A comparison of 13-year-old boys with attention deficit and/or reading disorder on neuropsychological measures

This study compared 13 year-old-boys with attention deficit disorder (ADD) and/or reading disability (RD), and controls with neither disorder on a battery of verbal and nonverbal neuropsychological measures. The aim was to examine whether ADD was associated with a qualitatively distinct pattern of deficits compared with RD. None of the measures differentiated the boys with ADD-only from the controls; the only deficit associated with ADD was slightly lower IQ. RD, on the other hand, was associated with deficits in memory and verbal skills.The Dunedin Multidisciplinary Health and Development Research Unit is supported by the Medical Research Council of New Zealand and involves several Departments of the University of Otago. Much of the data are gathered by voluntary workers from the Dunedin community.The authors are indebted to the many people who contributions make this ongoing study possible. We wish to acknowledge the advice and help of Dr. P. A. Silva, director of the Dunedin Unit, in preparing this report.     

Tall is typical: Central tendency, ideal dimensions, and graded category structure among tree experts and novices

Many accounts of categorization equate goodness-of-example with central tendency for common taxonomic categories; the best examples of a category are average members--those that are most similar to most other category members. In the present study, we asked 24 tree experts and 20 novices to rate goodness-of-example for a sample of 48 trees and found (1) that the internal structure of the category tree differed between novices and experts and (2) that central tendency did not determine goodness-of-example ratings for either group. For novices, familiarity determined goodness-of-example ratings. For experts, the "ideal" dimensions of height and weediness, rather than average similarity to other trees, were the primary predictors of goodness-of-example ratings for experts. The best examples of tree were not species of average height, but of extreme height. The worst examples were the weediest trees. We also found systematic differences in predictors of goodness-of-example as a function of type of expertise. We argue that the internal structure of taxonomic categories can be shaped by goal-related experience and is not necessarily a reflection of the attributional structure of the environment. Implications for models of category structure and category learning are discussed.