|
1. |
The relationship of variance to interaction contrast in parallel systems factorial technology |
|
British Journal of Mathematical and Statistical Psychology,
Volume 49,
Issue 2,
1996,
Page 211-223
James T. Townsend,
Adele Diederich,
Preview
|
PDF (555KB)
|
|
摘要:
Mean interaction contrast has been shown to be a beneficial statistic for testing between serial and parallel processes (e.g. Schweickert, 1978; Sternberg, 1969; Townscnd, 1984). The present theoretical note investigates a conjecture of Egeth&Dagenbach (1991) that the mean interaction contrast is largest when the parallel processes are each deterministic, that is possessing zero variance. We show that their conjecture is true in certain situations but not in others. We further demonstrate that the magnitude of mean interaction contrast may be a non‐monotonic function of processing time variance. Finally, implications for experimentation are pointed ou
ISSN:0007-1102
DOI:10.1111/j.2044-8317.1996.tb01085.x
出版商:Blackwell Publishing Ltd
年代:1996
数据来源: WILEY
|
2. |
Information processing in similarity judgments |
|
British Journal of Mathematical and Statistical Psychology,
Volume 49,
Issue 2,
1996,
Page 225-240
J.‐P. Barthélemy,
E. Mullet,
Preview
|
PDF (802KB)
|
|
摘要:
To account for expert information processing of similarity judgments on objects characterized by several attributes, a flexible model, inspired by works of Montgomery (1983) and Barthélemy&Mullet (1986, 1992) is presented and illustrated on the basis of data collected from experienced subjects. This model coordinates three types of rules: threshold rules, conjunctive rules and disjunctive rules. It is based on the principle that a ‘bounded consensus rule’ is the major rule and that all others are merely employed to obtain a consensus structure quickly. Three basic principles are incorporated in the model: (i) a parsimony principle, (ii) a reliability principle, (iii) a decidability principle. The model can be represented as a process (flowchart) or as a formula (equat
ISSN:0007-1102
DOI:10.1111/j.2044-8317.1996.tb01086.x
出版商:Blackwell Publishing Ltd
年代:1996
数据来源: WILEY
|
3. |
Testing hypotheses about working memory capacity |
|
British Journal of Mathematical and Statistical Psychology,
Volume 49,
Issue 2,
1996,
Page 241-252
Raphael Gillett,
Preview
|
PDF (553KB)
|
|
摘要:
When an individual is engaged in sampling from a finite population, an assessment of their working memory for previously sampled objects is often desired. An extension of the classical occupancy distribution provides a framework for testing hypotheses about working memory capacity, based on a general model of working memory in which the individual retains a record ofmprevious inspections (and hence does not make the mistake of resampling these objects) and samples with replacement objects outside the current memory set. The buffer sizemprovides a common metric for assessing the effect of variables on performance in search tasks involving different numbers of items, or locations.
ISSN:0007-1102
DOI:10.1111/j.2044-8317.1996.tb01087.x
出版商:Blackwell Publishing Ltd
年代:1996
数据来源: WILEY
|
4. |
A review of some recent developments in robust regression |
|
British Journal of Mathematical and Statistical Psychology,
Volume 49,
Issue 2,
1996,
Page 253-274
Rand R. Wilcox,
Preview
|
PDF (1378KB)
|
|
摘要:
In situations where the goal is to understand how a random variableyis related to a set ofppredictor variables, modern robust regression methods can be invaluable. One reason is that even one unusual value in the design space, or one outlier among theyvalues, can have a large impact on the ordinary least squares estimate of the parameters of the usual linear model. That is, a single unusual value or outlier can give a highly distorted view of how two or more random variables are related. Another reason is that modern robust methods can be much more efficient than ordinary least squares yet maintain good efficiency under the ideal conditions of normality and a homoscedastic error term. Even when sampling is from light‐tailed distributions, there are situations where certain robust methods are highly efficient compared to least squares, as is indicated in this paper. Most applied researchers in psychology simply ignore these problems. In the hope of improving current practice, this paper reviews some of the robust methods currently available with an emphasis on recent developments. Of particular interest are methods for computing confidence intervals and dealing with heteroscedasticity in the error ter
ISSN:0007-1102
DOI:10.1111/j.2044-8317.1996.tb01088.x
出版商:Blackwell Publishing Ltd
年代:1996
数据来源: WILEY
|
5. |
The analysis of repeated measurements: A quantitative research synthesis |
|
British Journal of Mathematical and Statistical Psychology,
Volume 49,
Issue 2,
1996,
Page 275-298
Joanne C. Keselman,
Lisa M. Lix,
H. J. Keselman,
Preview
|
PDF (1486KB)
|
|
摘要:
Meta‐analytic methods were used to summarize the results of Monte Carlo studies investigating the Type I error and power properties of various univariate and multivariate procedures for testing within‐subjects effects in split‐plot repeated measures designs. Results indicated that all test procedures were generally robust to violations of the multivariate normality assumption, but varied in terms of their Type I error control when the sphericity assumption was not satisfied. For balanced designs, the usualFand ê adjustedFtests (Greenhouse&Geisser, 1959) were generally robust to moderate degrees of covariance heterogeneity, whereas the multivariate procedures were slightly more affected by departures from this assumption. When the design was unbalanced, however, all procedures were sensitive to the presence of heterogeneous covariance matrices, particularly when testing the within‐subjects interaction effect. Power rates varied little as a function of assumption violations. However, this finding may be due to the restricted range of many of the variables included in the meta‐analysis of the power data as well as the strong and overshadowing relationship between the degree of non‐centrality and power rates. For balanced designs, the use of either an ê‐adjusted univariate or a multivariate approach is recommended; for unbalanced designs, researchers should consider adopting one of several robust alternatives that have recently been suggested in
ISSN:0007-1102
DOI:10.1111/j.2044-8317.1996.tb01089.x
出版商:Blackwell Publishing Ltd
年代:1996
数据来源: WILEY
|
6. |
Test of linear trend in eigenvalues of a covariance matrix with application to data analysis |
|
British Journal of Mathematical and Statistical Psychology,
Volume 49,
Issue 2,
1996,
Page 299-312
P. M. Bentler,
Ke‐Hai Yuan,
Preview
|
PDF (632KB)
|
|
摘要:
Principal component analysis and factor analysis are the most widely used tools for dimension reduction in data analysis. Both methods require some good criterion to judge the number of dimensions to be kept. The classical method focuses on testing the equality of eigenvalues. As real data hardly have this property, practitioners turn to somead hoccriterion in judging the dimensionality of their data. One such popular method, the ‘scree test’ or ‘scree plot’ as described in many texts and statistical programs, is based on the trend in eigenvalues of sample covariance (correlation) matrix. The principal components or common factors corresponding to eigenvalues which exhibit a slow linear decrease arc discarded in further data analysis. This paper develops a formal statistical test for the ‘scree plot’. A special case of this test is the classical test for equality of eigenvalues which has been suggested in several texts as the criterion to decide the number of principal components to retain. Comparisons between equality of eigenvalues and the slow linear decrease in eigenvalues on some classical examples support the hypothesis of slow linear decrease. A physical background to such a phenomenon is als
ISSN:0007-1102
DOI:10.1111/j.2044-8317.1996.tb01090.x
出版商:Blackwell Publishing Ltd
年代:1996
数据来源: WILEY
|
7. |
A latent trait and a latent class model for mixed observed variables |
|
British Journal of Mathematical and Statistical Psychology,
Volume 49,
Issue 2,
1996,
Page 313-334
Irini Moustaki,
Preview
|
PDF (1128KB)
|
|
摘要:
Latent variable models are widely used in social sciences in which interest is centred on entities such as attitudes, beliefs or abilities for which there exist no direct measuring instruments. Latent modelling tries to extract these entities, here described as latent (unobserved) variables, from measurements on related manifest (observed) variables. Methodology already exists for fitting a latent variable model to manifest data that is either categorical (latent trait and latent class analysis) or continuous (factor analysis and latent profile analysis).In this paper a latent trait and a latent class model are presented for analysing the relationships among a set of mixed manifest variables using one or more latent variables. The set of manifest variables contains metric (continuous or discrete) and binary items. For the latent trait model the latent variables are assumed to follow a multivariate standard normal distribution. Our method gives maximum likelihood estimates of the model parameters and standard errors of the estimates by analysing the data as they are without using any underlying variables. The mixed latent trait and latent class models are fitted using an EM algorithm.To illustrate the use of the mixed model three data sets have been analysed. Two of the data sets contain five memory questions, the first on Thatcher's resignation and the second on the Hillsborough football disaster; these five questions were included in British Market Research Bureau International August 1993 face‐to‐face omnibus survey. The third data set is from the 1991 British Social Attitudes Survey; the questions which have been analysed are from the environment sect
ISSN:0007-1102
DOI:10.1111/j.2044-8317.1996.tb01091.x
出版商:Blackwell Publishing Ltd
年代:1996
数据来源: WILEY
|
8. |
Optimality criteria for principal component analysis and generalizations |
|
British Journal of Mathematical and Statistical Psychology,
Volume 49,
Issue 2,
1996,
Page 335-345
Jos M. F. Berge,
Henk A. L. Kiers,
Preview
|
PDF (669KB)
|
|
摘要:
Principal components analysis can be derived from various criteria. Because these give essentially the same results, the question of which criterion should be used has not received much attention. Nevertheless, it can be argued that the approach of Pearson and Eckart&Young, based on variance explained by components, is more elegant and flexible than the (more popular) approach of Hotelling, which is concerned with variance that components have rather than explain. When two or more correlation or covariance matrices, based on the same variables, are to be analyzed in generalized component analysis, the question of which criterion is used becomes of utmost importance. A taxonomy of generalized principal component methods is given. It appears that generalized component analysis based on the Hotelling criterion coincides with one particular generalization based on the criterion of Pearson and Eckart&Young.
ISSN:0007-1102
DOI:10.1111/j.2044-8317.1996.tb01092.x
出版商:Blackwell Publishing Ltd
年代:1996
数据来源: WILEY
|
9. |
A hyperbolic cosine latent trait model for unfolding polytomous responses: Reconciling Thurstone and Likert methodologies |
|
British Journal of Mathematical and Statistical Psychology,
Volume 49,
Issue 2,
1996,
Page 347-365
David Andrich,
Preview
|
PDF (1005KB)
|
|
摘要:
In the 1920s, Thurstone articulated a theory for the measurement of social variables that involved two distinct steps: first constructing and operationalizing a linear continuum by locating statements according to their affective values; second measuring persons. The first step involved thepair‐comparison designand thecumulative response mechanism, the second step involved thedirect‐response design(of the form Agree or Disagree) and theunfolding response mechanism.In the 1930s, Likert proposed a procedure that obviated the need for the first step and apparently permitted measuring persons from their responses to statements that were similar to those used by Thurstone, but which required a response that indicated degrees of intensity of the form Strongly Agree, Agree, Undecided, Disagree or Strongly Disagree. Furthermore, and in contrast to Thurstone, he implicitly used the cumulative mechanism, scoring the successive categories with successive integers and simply summing them to obtain a measurement for each person. The two procedures were considered to be distinct and alternative, and in general, this is still the perception with a number of matters still not reconciled between the two procedures.By resolving the Disagree response in the unfolding mechanism into its two constituent components, this paper presents an unfolding model for direct responses from first principles, and then generalises it to provide an unfolding model for polytomous responses of the Likert style. This model permits an understanding of those matters still not reconciled between the Thurstone and Likert approaches: First, Likert's success in using Thurstone‐like statements with the cumulative mechanism rather than the unfolding mechanism; second, the gap found between two clusters of locations when statements from a Likert‐style questionnaire are scaled using the Thurstone procedure; and third, the consistent problem with the middle category of Undecided in Likert‐style respons
ISSN:0007-1102
DOI:10.1111/j.2044-8317.1996.tb01093.x
出版商:Blackwell Publishing Ltd
年代:1996
数据来源: WILEY
|
10. |
Saturated models for repeated measures |
|
British Journal of Mathematical and Statistical Psychology,
Volume 49,
Issue 2,
1996,
Page 367-379
Gordon G. Bechtel,
André I. Khuri,
Preview
|
PDF (573KB)
|
|
摘要:
Psychological and physiological data often consist of correlated responses repeatedly taken from the same sample of individuals. The unknown covariances among these responses present a problem for statistical inference to parameters of the population from which the samples are drawn. The present paper addresses this problem for saturated mean response models, i.e. models in which parameters and mean responses are equal in number. In these models the unknown covariances among repeated measures, which are nuisance parameters, are shown to cancel out. This cancellation provides exact tests and estimates for parameters within and between groups by means of Wishart distribution theory. These tests and estimates are illustrated for three different types of repeated measures, each involving a saturated mean response model. The unknown covariances in saturated models are also discussed in relation to the structured covariances in random coefficient models.
ISSN:0007-1102
DOI:10.1111/j.2044-8317.1996.tb01094.x
出版商:Blackwell Publishing Ltd
年代:1996
数据来源: WILEY
|
|