Creating Misspecified Models in Moment Structure Analysis

Psychometrika - To understand how SEM methods perform in practice where models always have misfit, simulation studies often involve incorrect models. To create a wrong model, traditionally one...     

Statistical Tests for Comparing Possibly Misspecified and Nonnested Models

Model selection criteria (MSC) involves selecting the model with the best estimated goodness-of-fit to the data generating process. Following the method of Vuong (1989), a large sample Model Selection Test (MST), is introduced that can be used in conjunction with most existing MSC procedures to decide if the estimated goodness-of-fit for one model is significantly different from the estimated goodness-of-fit for another model. The MST extends the classical generalized likelihood ratio test, is valid in the presence of model misspecification, and is applicable to situations involving nonnested probability models. Simulation studies designed to illustrate the concept of the MST and its conservative decision rule (relative to the MSC method) are also presented. Copyright 2000 Academic Press.     

Teacher Versus Peer Models

Female Ss (N=80) observed either a teacher or a peer model (M) demonstrate a motor task with either high or low performance ability. The effects’ of viewing these live performances on S’s subsequent motor performances were examined. After the observation of M, S’s performance in the presence or absence of the observed M was investigated. A Control Group consisting of 20 Ss, viewing no M, was compared to ail other groups for analysis. Results of the M Type × M’s Ability Level × M Presence design indicated that those in the Control Group had significantly lower performances than all other groups on all blocks of trials. In addition, across all trial blocks, S’s performances were enhanced subsequent to having viewed a skillful teacher M. Upon observing M’s demonstration, Ss exhibited increased matching responses and better performances when performing in the presence of M.     

Computational Versus Associative Models of Simple Conditioning

In associative models of simple conditioning, conductive connections (associations, Hebbian synapses) are strengthened by the repetitive temporal pairing of stimuli. The associations cause the animal to behave more adaptively, but they do not encode information about objectively specifiable properties of the conditioning experience. In information processing (computational) models, the temporal intervals in that experience are timed and the results recorded in memory for later use in computations that determine the decisions whether and when to respond to the conditioned stimulus. The predictions of these latter models depend on the ratios of remembered and currently experienced temporal intervals; hence, they are time-scale invariant. Two examples of empirical time-scale invariance are described: Neither the delay of reinforcement nor the ratio of reinforced to unreinforced presentations of the conditioned stimulus affects rates of acquisition and extinction. Time-scale invariance has far-reaching implications for models of the processes that underlie conditioning, for example, models of Hebbian synapses.     

Models of Success Versus the Success of Models: Reliability without Truth

In computer simulations of physical systems, the construction of models is guided, but not determined, by theory. At the same time simulations models are often constructed precisely because data are sparse. They are meant to replace experiments and observations as sources of data about the world; hence they cannot be evaluated simply by being compared to the world. So what can be the source of credibility for simulation models? I argue that the credibility of a simulation model comes not only from the credentials supplied to it by the governing theory, but also from the antecedently established credentials of the model building techniques employed by the simulationists. In other words, there are certain sorts of model building techniques which are taken, in and of themselves, to be reliable. Some of these model building techniques, moreover, incorporate what are sometimes called “falsifications.” These are contrary-to-fact principles that are included in a simulation model and whose inclusion is taken to increase the reliability of the results. The example of a falsification that I consider, called artificial viscosity, is in widespread use in computational fluid dynamics. Artificial viscosity, I argue, is a principle that is successfully and reliably used across a wide domain of fluid dynamical applications, but it does not offer even an approximately “realistic” or true account of fluids. Artificial viscosity, therefore, is a counter-example to the principle that success implies truth – a principle at the foundation of scientific realism. It is an example of reliability without truth.     

Edgar Schein's Process Versus Content Consultation Models

Edgar Schein (1969, 1978, 1989, 1990) has proposed three models of consultation based on assumptions inherent in different helping styles. The first two models, purchase-of-expertise and doctor-patient, focus on the content of organizational problems. The client gives the problem to the consultant to find and implement solutions. The process consultation model focuses on how organizational problems are solved. The client and consultant collaborate throughout the problem-solving effort to find workable solutions. Because of the nature of organizations' cultures and the underlying assumptions that determine how they operate, Schein has suggested that it is most efficacious for consultants to begin in the process consultation mode and involve the client in the investigation of the problem.     

Career Decision-Making Models: Espoused Theory Versus Theory-in-Use

The author examines the limitations of career decision-making models from a theory-in-action perspective (Argyris, 1976). Research implications are discussed.     

Bayesian Comparison of Latent Variable Models: Conditional Versus Marginal Likelihoods

Psychometrika - Typical Bayesian methods for models with latent variables (or random effects) involve directly sampling the latent variables along with the model parameters. In high-level software...     

A Comparison of Health Communication Models: Risk Learning Versus Stereotype Priming

Health communication research and practice have been strongly influenced by the protection motivation theory (Rogers, 1975, 1983), the health belief model (Becker, Haefner, Kasl, et al., 1977; Becker, Haefner, & Maiman, 1977; Rosenstock, 1974), and similar conceptualizations. I refer to these as risk learning models because the goal is to teach new information about health risks and the behaviors that will minimize those risks. These models have garnered a substantial amount of empirical support and are apparently quite useful to practitioners (Conner & Norman, 1996). The goal of this article is to describe a less familiar, but complementary, approach to persuading people to avoid risky behaviors, which I will refer to as the stereotype priming model (Bargh, 1989; Bargh, Chen, & Burrows, 1996; Bargh, Raymond, Pryor, & Strack, 1995). The goal is to make salient preexisting social stereotypes about people who do or do not behave as advocated. The stereotype priming model posits that the stereotypes that we possess regarding the personality traits of groups of people (e.g., smokers, drunk drivers, marijuana users, people with suntans) to a large extent govern our behavior. In this view, it is sometimes useful to capitalize on and reinforce preexisting stereotypes in order to encourage healthy, and/or discourage unhealthy, behaviors. This article reviews the traditional risk learning models and evidence that supports them, provides corresponding information for the stereotype priming model, compares and contrasts the models, and discusses how the stereotype priming model might be used to design health communications campaigns.     

On the multivariate asymptotic distribution of sequential Chi-square statistics

The multivariate asymptotic distribution of sequential Chi-square test statistics is investigated. It is shown that: (a) when sequential Chi-square statistics are calculated for nested models on the same data, the statistics have an asymptotic intercorrelation which may be expressed in closed form, and which is, in many cases, quite high; and (b) sequential Chi-squaredifference tests are asymptotically independent. Some Monte Carlo evidence on the applicability of the theory is provided.This research was carried out while the first author was Visiting Professor in the Department of Statistics in the University of South Africa, and was supported in part by a research grant (NSERC #67-4640) from the National Sciences and Engineering Council of Canada to the first author. The support of both of these organizations is acknowledged with gratitude.     

Ensuring Positiveness of the Scaled Difference Chi-square Test Statistic

A scaled difference test statistic [(T)\tilde]_d\tilde{T}{}_{d} that can be computed from standard software of structural equation models (SEM) by hand calculations was proposed in Satorra and Bentler (Psychometrika 66:507–514, 2001). The statistic [(T)\tilde]_d\tilde{T}_{d} is asymptotically equivalent to the scaled difference test statistic [`(T)]<sub>d</sub>\bar{T}_{d} introduced in Satorra (Innovations in Multivariate Statistical Analysis: A Festschrift for Heinz Neudecker, pp. 233–247, 2000), which requires more involved computations beyond standard output of SEM software. The test statistic [(T)\tilde]<sub>d</sub>\tilde{T}_{d} has been widely used in practice, but in some applications it is negative due to negativity of its associated scaling correction. Using the implicit function theorem, this note develops an improved scaling correction leading to a new scaled difference statistic [`(T)]_d\bar{T}_{d} that avoids negative chi-square values.     

Bifactor Models of Personality and College Student Performance: A Broad Versus Narrow View

Research in the area of personality traits and academic performance has been supported by consistent meta‐analytic evidence demonstrating positive relationships between Conscientiousness and grade point average (GPA). However, academic performance is not solely a function of GPA but also a number of other important intellectual, interpersonal and intrapersonal behaviours. This wider criterion space opens up the possibility for many personality factors and their underlying facets to relate to academic performance. Using bifactor latent variable modelling, the current study investigates the six‐factor HEXACO model of personality, along with their 24 underlying facets, for predicting students' academic performance. Model results reveal interpretable and meaningful relationships between both broad factors and narrow personality facets in predicting college student outcomes. Implications for measurement, modelling and prediction are discussed. Copyright © 2014 European Association of Personality Psychology     

Autobiographical Memory Specificity in Patients with Tinnitus Versus Patients with Depression and Normal Controls

Several studies show that patients with depression and post-traumatic stress disorder respond with fewer specific autobiographical memories in a cued memory task (i.e. the autobiographical memory test; AMT) compared to healthy controls. One previous study found this phenomenon among tinnitus patients as well (Andersson, Ingerholt, & Jansson, 2003). The aim of this study was to replicate the previous study with an additional control group of depressed patients and memory errors as measured with the AMT as an additional outcome. We included 20 normal hearing tinnitus patients, 20 healthy controls and 20 persons diagnosed with clinical depression. The AMT was administered together with self-report measures of depression, anxiety and tinnitus distress. Both the tinnitus and depression groups differed from the healthy control group in that they reported fewer specific autobiographical memories. There were, however, differences between the tinnitus and depression groups in terms of the errors made on the AMT. The depression group had more overgeneral memories than the normal control group, whereas the tinnitus group did not differ from the control group on this memory error. The tinnitus group had more semantic associations and non-memories than the other two groups, suggesting that executive functioning may play a role for the tinnitus group when completing the AMT. Clinical and theoretical implications of the findings are discussed.     

Angry Versus Furious: A Comparison Between Valence and Arousal in Dimensional Models of Emotions

The present study compared the roles of valence and arousal, proposed by the dimensional models of emotions, in recognizing emotional expressions. It was hypothesized that the dimension of valence, due to its functional significance, would be more salient in the recognition of emotional expressions than the dimension of arousal would be. The results of the current study supported this hypothesis. The participants in all age groups were more accurate and quicker in recognizing an emotion when the expression was paired up with another emotional expression that was different in the polarity of the valence dimension than with this similar polarity difference in the arousal dimension. The insignificant difference in recognizing the positive and negative emotional expressions in the group of elders also rejected the Socio-Emotional Selectivity Theory.     

A normal approximation for testing the equality of two independent Chi-square variables

A simple normal approximation is supplied to test the equality of two independent Chi-square variables. The approximation appears to be valid for degrees of freedom as small as two. Thus no new table of critical values is needed. The standard unit normal distribution tables are sufficient.     

The Similarity-Attraction Link: Sequential Versus Parallel Multiple-Mediator Models Involving Inferred Attraction,Respect, and Positive Affect

We tested the hypothesis that the attitude similarity effect spreads sequentially through positive affect, respect, and inferred attraction to attraction. In Experiment 1, participants received information about a partner's similar or dissimilar attitudes and reported one of the three mediators before attraction. The similarity-attraction link was mediated by positive affect or respect but more strongly by inferred attraction. In Experiment 2, the three mediators were measured in different orders. Results falsified parallel and combined parallel-sequential multiple-mediator models and two of the fully sequential multiple-mediator models but supported four fully sequential models that were consistent with the affect-centered and affect primacy hypotheses.     

A Test of Positive Illusions Versus Shared Reality Models of Relationship Satisfaction Among Gay, Lesbian, and Heterosexual Couples

According to the positive illusions model ( Murray, Holmes, & Griffin, 1996a ), people in romantic relationships are more satisfied when they view their partners more favorably than the partners see themselves. By contrast, shared reality theory ( Hardin & Conley, 2001 ) emphasizes the benefits of perceiving a partner as the partner sees himself or herself. We analyzed archived data from the American Couples Study ( Blumstein & Schwartz, 1983 ) to test the applicability of the positive illusions model to gay and lesbian relationships. Structural equation models demonstrated that the positive illusions model effectively explains relationships among lesbian, gay, heterosexual cohabitating, and married couples.     

移植物抗宿主病与移植物抗肿瘤的处置

移植物抗宿主病与移植物抗肿瘤是当前异基因造血干细胞移植中一对不可调和的矛盾。在解决这一矛盾方面,科学家们已取得了很大的进展。随着移植学、免疫学的发展和认识的深入,最终能够最大限度的削弱移植物抗宿主病,增强移植物抗肿瘤效应,提高移植后患者的生存质量。