A Bayesian Semiparametric Latent Variable Model for Mixed Responses

In this paper we introduce a latent variable model (LVM) for mixed ordinal and continuous responses, where covariate effects on the continuous latent variables are modelled through a flexible semiparametric Gaussian regression model. We extend existing LVMs with the usual linear covariate effects by including nonparametric components for nonlinear effects of continuous covariates and interactions with other covariates as well as spatial effects. Full Bayesian modelling is based on penalized spline and Markov random field priors and is performed by computationally efficient Markov chain Monte Carlo (MCMC) methods. We apply our approach to a German social science survey which motivated our methodological development. We thank the editor and the referees for their constructive and helpful comments, leading to substantial improvements of a first version, and Sven Steinert for computational assistance. Partial financial support from the SFB 386 “Statistical Analysis of Discrete Structures” is also acknowledged.     

The Effects of Thematic Importance on Story Recall among Children with Attention Deficit Hyperactivity Disorder and Comparison Children

This study examined the recall of televised stories for younger (4-6 years) and older (7-9 years) children with and without attention deficit hyperactivity disorder (ADHD) under two different viewing conditions (toys present/toys absent). Each child watched two Rugrats television programs, once with toys present and once with toys absent. Immediately after viewing a program, the child completed a free recall of the observed story. Comparison children's recall increased more than ADHD children's as importance level increased, and comparison children recalled more information overall than children with ADHD. When toys were present, children with ADHD retold less coherent stories than comparison children, as indexed by smaller correlations between the story units recalled and the order of these units in the story. In summary, children with ADHD demonstrated multiple difficulties in story comprehension. These findings add to our understanding of the differences in higher-order cognitive processing abilities between children with ADHD and comparison children, and suggest important areas of focus in designing more effective academic interventions for children with ADHD.     

澳门人的风险知觉与赌博行为

采用匿名问卷法,调查澳门人的风险知觉与赌博行为。结果显示：（1）赌场的劝世文既无劝勉也无劝阻人们赌博的作用;（2）社会关系网的钱财支援不影响实际赌博次数;（3）多次性博弈与一次性博弈所采用的是不同的机制,一次性博弈不是由期望值所决定的;（4）相互监视和制裁系统会影响赌博：受法律制裁约束的职业者比不受法律制裁约束的职业者更不好赌;自估在赌场遇见认识人的机会与赌博次数呈负相关;（5）性别及母语等“本质趋向”差异表明,所谓澳门人不好赌有其渊源并可能是在长期的文化沉淀中形成的;（6）受教育程度是影响赌博的有效预测变量。这些发现的理论含义一并进行了讨论。     

A variable-selection heuristic for K-means clustering

One of the most vexing problems in cluster analysis is the selection and/or weighting of variables in order to include those that truly define cluster structure, while eliminating those that might mask such structure. This paper presents a variable-selection heuristic for nonhierarchical (K-means) cluster analysis based on the adjusted Rand index for measuring cluster recovery. The heuristic was subjected to Monte Carlo testing across more than 2200 datasets with known cluster structure. The results indicate the heuristic is extremely effective at eliminating masking variables. A cluster analysis of real-world financial services data revealed that using the variable-selection heuristic prior to the K-means algorithm resulted in greater cluster stability.     

An overview of applied robust methods

Data in social sciences are typically non-normally distributed and characterized by heavy tails. However, most widely used methods in social sciences are still based on the analyses of sample means and sample covariances. While these conventional methods continue to be used to address new substantive issues, conclusions reached can be inaccurate or misleading. Although there is no ‘best method’ in practice, robust methods that consider the distribution of the data can perform substantially better than the conventional methods. This article gives an overview of robust procedures, emphasizing a few that have been repeatedly shown to work well for models that are widely used in social and behavioural sciences. Real data examples show how to use the robust methods for latent variable models and for moderated mediation analysis when a regression model contains categorical covariates and product terms. Results and logical analyses indicate that robust methods yield more efficient parameter estimates, more reliable model evaluation, more reliable model/data diagnostics, and more trustworthy conclusions when conducting replication studies. R and SAS programs are provided for routine applications of the recommended robust method.     

A bayesian estimation of factor score in confirmatory factor model with polytomous, censored or truncated data

The posterior analysis in estimating factor score in a confirmatory factor analysis model with polytomous, censored or truncated data is investigated in this paper. For the above three types of data, posterior distributions of the factor score are studied, and the estimators of the factor score are obtained to be the location parameters of the posterior distributions. The accuracy of Bayesian estimates is studied via simulation studies.This research was supported by a Hong Kong UGC grant.     

Stimulus-specific contrast effects during operant discrimination learning

In two experiments, pigeons' responding was equally reinforced in the presence of four line-orientation stimuli. Responding was then reinforced when only two of the four orientation stimuli were present; the remaining two orientations appeared during extinction. Response rates were often highest in the stimulus adjacent to the orientations presented during extinction and often lowest in that orientation adjacent to the orientations presented with reinforcement. These effects were stronger and more persistent when the stimuli were separated by a smaller angle, rendering the discrimination more difficult. These and other data suggest that discrimination training may not be accurately explained in terms of the simple effects of reinforcement and nonreinforcement associated with isolated stimuli, nor by accounts that depend upon stimulus generalization. Recent accounts of contrast that depend upon “emotionality” produced by nonreinforced responding or upon reinforcement-elicited responses are also difficult to apply to these data.     

Matching with a trio of concurrent variable-interval schedules of reinforcement

A trio of concurrent variable-interval schedules of reinforcement was arranged according to a changeover-key procedure, including a changeover delay of 1.5 sec. The three schedules provided a combined maximum reinforcement rate of 45 reinforcements per hour. With that restriction, the nine experimental conditions included several combinations of variable-interval schedules, sometimes including extinction. The pigeons matched relative response rate and relative time to relative reinforcement rate. Relative time appeared to match some-what better than relative response rate. Performance adjusted rapidly from one experimental condition to the next, whether the change involved two or all three schedules of the concurrent trio.     

RESPONSE REDUCTION THROUGH THE SUPERIMPOSITION OF CONTINUOUS REINFORCEMENT: A SYSTEMATIC REPLICATION

Prior clinical research suggests that superimposition and subsequent removal of a schedule of continuous reinforcement (CRF) may be a viable rate-decreasing procedure in that an extinction-like condition is arranged. The arrangement of similar conditions in the laboratory, however, resulted in the quick recovery of baseline rates. Lever-pressing patterns of eight male rats maintained by different schedules of variable-ratio and variable-interval food reinforcement were examined in an A-B-A experimental design of CRF food superimposition and removal. Responding was substantially reduced during the superimposition of CRF. Upon removal of the superimposed schedule, responding quickly approached presuperimposition baseline rates.