Decomposing person and occasion-specific effects: an extension of latent state-trait (LSI) theory to hierarchical LST models

An extension of latent state-trait (LST) theory to hierarchical LST models is presented. In hierarchical LST models, the covariances between 2 or more latent traits are explained by a general 3rd-order factor, and the covariances between latent state residuals pertaining to different traits measured on the same measurement occasion are explained by 2nd-order latent occasion-specific factors. Analogous to recent developments in multitrait-multimethod methodology, all factors are interpreted in relation to factors taken as comparison standards. An empirical example from test anxiety research illustrates how estimates of additive variance components due to general trait, specific trait, occasion, state residual, method, and measurement error can be obtained using confirmatory factor analysis. Advantages and limitations of these models are discussed.     

Integrating personality, daily life events and emotion: Role of anxiety and positive affect in emotion regulation dynamics

We investigated the roles of anxiety and positive affect in emotion regulation, looking simultaneously at personality, daily life events, and affects. We hypothesized that individual differences in the temporal dynamics of affective experience related to trait anxiety would manifest themselves both in affective responsiveness to life events and in homeostatic regulatory forces. Data were collected from 49 adults, who rated their affective state three times a day over a 40-day period. Data were analyzed using a dynamical system model and graphical representations in the form of vector fields. Results showed that anxiety chiefly interacted with home base (attractor) positions as a function of life events. It also influenced the shape of positive affectivity trajectories in response to negative events.     

Modeling Nonstationary Emotion Dynamics in Dyads using a Time-Varying Vector-Autoregressive Model

Emotion dynamics are likely to arise in an interpersonal context. Standard methods to study emotions in interpersonal interaction are limited because stationarity is assumed. This means that the dynamics, for example, time-lagged relations, are invariant across time periods. However, this is generally an unrealistic assumption. Whether caused by an external (e.g., divorce) or an internal (e.g., rumination) event, emotion dynamics are prone to change. The semi-parametric time-varying vector-autoregressive (TV-VAR) model is based on well-studied generalized additive models, implemented in the software R. The TV-VAR can explicitly model changes in temporal dependency without pre-existing knowledge about the nature of change. A simulation study is presented, showing that the TV-VAR model is superior to the standard time-invariant VAR model when the dynamics change over time. The TV-VAR model is applied to empirical data on daily feelings of positive affect (PA) from a single couple. Our analyses indicate reliable changes in the male’s emotion dynamics over time, but not in the female’s—which were not predicted by her own affect or that of her partner. This application illustrates the usefulness of using a TV-VAR model to detect changes in the dynamics in a system.     

A Bayesian hierarchical diffusion model decomposition of performance in Approach–Avoidance Tasks

Common methods for analysing response time (RT) tasks, frequently used across different disciplines of psychology, suffer from a number of limitations such as the failure to directly measure the underlying latent processes of interest and the inability to take into account the uncertainty associated with each individual's point estimate of performance. Here, we discuss a Bayesian hierarchical diffusion model and apply it to RT data. This model allows researchers to decompose performance into meaningful psychological processes and to account optimally for individual differences and commonalities, even with relatively sparse data. We highlight the advantages of the Bayesian hierarchical diffusion model decomposition by applying it to performance on Approach–Avoidance Tasks, widely used in the emotion and psychopathology literature. Model fits for two experimental data-sets demonstrate that the model performs well. The Bayesian hierarchical diffusion model overcomes important limitations of current analysis procedures and provides deeper insight in latent psychological processes of interest.     

Latent Class Models for Diary Method Data: Parameter Estimation by Local Computations

The increasing use of diary methods calls for the development of appropriate statistical methods. For the resulting panel data, latent Markov models can be used to model both individual differences and temporal dynamics. The computational burden associated with these models can be overcome by exploiting the conditional independence relations implied by the model. This is done by associating a probabilistic model with a directed acyclic graph, and applying transformations to the graph. The structure of the transformed graph provides a factorization of the joint probability function of the manifest and latent variables, which is the basis of a modified and more efficient E-step of the EM algorithm. The usefulness of the approach is illustrated by estimating a latent Markov model involving a large number of measurement occasions and, subsequently, a hierarchical extension of the latent Markov model that allows for transitions at different levels. Furthermore, logistic regression techniques are used to incorporate restrictions on the conditional probabilities and to account for the effect of covariates. Throughout, models are illustrated with an experience sampling methodology study on the course of emotions among anorectic patients. Frank Rijmen was partly supported by the Fund for Scientific Research Flanders (FWO).     

新世纪20年国内心理统计方法研究回顾

新世纪头20年, 国内心理学11本专业期刊一共发表了213篇统计方法研究论文。研究范围主要包括以下10类(按论文篇数排序)：结构方程模型、测验信度、中介效应、效应量与检验力、纵向研究、调节效应、探索性因子分析、潜在类别模型、共同方法偏差和多层线性模型。对各类做了简单的回顾与梳理。结果发现, 国内心理统计方法研究的广度和深度都不断增加, 研究热点在相互融合中共同发展; 但综述类论文比例较大, 原创性研究论文比例有待提高, 研究力量也有待加强。     

Emotional inertia: A key to understanding psychotherapy process and outcome

The processes underlying psychotherapeutic change have increasingly been emphasized in both research and clinical practice. Nonlinear dynamical systems theory (NDS) offers a transdisciplinary scientific approach to the study of these processes. This paper introduces the NDS concept of “emotional inertia”, the property of human emotion by which it retains its course so long as it is not acted upon by an external force, as a key to understanding moment-by-moment and also longer-term change processes within psychotherapy. A testable mathematical model of emotional inertia is presented that represents specific impacts of psychotherapeutic processes on emotional dynamics over time. Emotional trajectories in phase space, treatment energy, and the interaction between them are the essential elements of the model, and a detailed explanation is provided. Procedures for testing this model are described, such as by tracking the movement of emotion in phase space within and across therapy sessions, along with clinical implications of the model, which can potentially help to make more clear the complementary roles of therapeutic force, timing, and leverage.     

The temporal factor of change in stressor-strain relationships: a growth curve model on a longitudinal study in east Germany

Several theoretical models describing how stressor-strain relationships unfold in time (e.g., M. Frese & D. Zapf, 1988) were tested with a longitudinal study, with 6 measurement waves, using multivariate latent growth curve models. The latent growth curve model made it possible to decompose trait and state components of strains and to show that both trait and state components are affected by work stressors. Because East Germany constitutes a high-change environment, it is an appropriate setting in which to study the relationship between work stressors and strains. The results showed that both the state and trait components of strains were affected by stressors. For example, individual trends in uncertainty (stressor) and worrying (strain) were related, whereas worrying also showed a short-term relationship with time pressure (another stressor). In particular, the decomposition into trait and state components was only possible with the growth curve method that was used.     

Development and Evolution of Cognitive Behavior Therapy for Depressed Adolescents

Since the 1980’s cognitive behavior therapy (CBT) has been developed and tested with adolescents experiencing depression. Early studies demonstrated that CBT was more effective than no intervention for young people with depressive symptoms. Beginning in the 1990’s CBT has been proven to be efficacious for adolescents with diagnosed depressive disorders, specifically Major Depressive Disorder (MDD). A series of increasingly challenging and methodologically rigorous trials showed that CBT was an effective intervention for mild to moderate MDD, and enhanced the benefits of anti-depressant medication, even for those adolescents who had already failed to respond to an initial medication. Simultaneously, some limitations of CBT were suggested by these studies. Results of CBT alone were not encouraging with severely depressed adolescents, and the combination of CBT plus medication was not uniformly supported. As CBT investigators worked with more challenging, complex clinical cases, such as depressed adolescents who were abusing substances or engaging in suicidal or self-harm behaviors, standard components of CBT were augmented by increased emphasis placed on emotion regulation, safety planning, and by more intensive outpatient treatment models. Newer developments include “third wave” models, which have shown initial promise.     

A hierarchical bayesian model of human decision-making on an optimal stopping problem

We consider human performance on an optimal stopping problem where people are presented with a list of numbers independently chosen from a uniform distribution. People are told how many numbers are in the list, and how they were chosen. People are then shown the numbers one at a time, and are instructed to choose the maximum, subject to the constraint that they must choose a number at the time it is presented, and any choice below the maximum is incorrect. We present empirical evidence that suggests people use threshold-based models to make decisions, choosing the first currently maximal number that exceeds a fixed threshold for that position in the list. We then develop a hierarchical generative account of this model family, and use Bayesian methods to learn about the parameters of the generative process, making inferences about the threshold decision models people use. We discuss the interesting aspects of human performance on the task, including the lack of learning, and the presence of large individual differences, and consider the possibility of extending the modeling framework to account for individual differences. We also use the modeling results to discuss the merits of hierarchical, generative and Bayesian models of cognitive processes more generally.     

Evidence for an emotion-cognition interaction in the statistical prediction of suicide attempts

Suicidal behavior is a prevalent problem among adolescents and young adults. Although most theoretical models of suicide suggest that this behavior results from the interaction of different risk factors, most prior studies have tested only bivariate associations between individual risk factors and suicidal behaviors. The current study was designed to address this limitation by testing the effect of an emotion–cognition interaction on suicide attempts among youth. Specifically, we hypothesized that the interaction of emotion reactivity and problem-solving skills would statistically predict the probability of a recent suicide attempt among 87 adolescents and young adults. Results revealed a significant interaction, such that emotion reactivity was strongly associated with the probability of a suicide attempt among those with poor problem-solving skills, moderately associated among those with average problem-solving skills, and not significantly associated among those with good problem-solving skills. The next generation of studies on suicidal behavior should continue to examine how risk factors interact to predict this dangerous outcome.     

Intellectual growth in children as a function of domain specific and domain general working memory subgroups

This study examined whether children's growth on measures of fluid (Raven Colored Progressive Matrices) and crystallized (reading and math achievement) intelligence was attributable to domain-specific or domain-general functions of working memory (WM). A sample of 290 elementary school children was tested on measures of intelligence across three testing waves. Two methods, a hierarchical factor model predicting latent growth and a double dissociation design comparing children divided into high and low performers, tested whether general and/or specific components accounted for performance on intelligence measures. A general domain model for the total sample provided the best fit for latent growth on all measures. Further, the lack of a significant WM subgroup × domain interaction suggested that a general WM system underlies the individual differences on measures of fluid and crystallized intelligence. Overall, the findings support a domain-general view of WM capacity on intelligence measures in children, but also suggest that domain specific storage systems may come into play on isolated measures.     

An Unscented Kalman Filter Approach to the Estimation of Nonlinear Dynamical Systems Models

In the past several decades, methodologies used to estimate nonlinear relationships among latent variables have been developed almost exclusively to fit cross-sectional models. We present a relatively new estimation approach, the unscented Kalman filter (UKF), and illustrate its potential as a tool for fitting nonlinear dynamic models in two ways: (1) as a building block for approximating the log–likelihood of nonlinear state–space models and (2) to fit time-varying dynamic models wherein parameters are represented and estimated online as other latent variables. Furthermore, the substantive utility of the UKF is demonstrated using simulated examples of (1) the classical predator-prey model with time series and multiple–subject data, (2) the chaotic Lorenz system and (3) an empirical example of dyadic interaction.     

The Joint Multivariate Modeling of Multiple Mixed Response Sources: Relating Student Performances with Feedback Behavior

The present study concerns a Dutch computer-based assessment, which includes an assessment process about information literacy and a feedback process for students. The assessment is concerned with the measurement of skills in information literacy and the feedback process with item-based support to improve student learning. To analyze students’ feedback behavior (i.e. feedback use and attention time), test performance, and speed of working, a multivariate hierarchical latent variable model is proposed. The model can handle multivariate mixed responses from multiple sources related to different processes and comprehends multiple measurement components for responses and response times. A flexible within-subject latent variable structure is defined to explore multiple individual latent characteristics related to students’ test performance and feedback behavior. Main results of the computer-based assessment showed that feedback-information pages were less visited by well-performing students when they relate to easy items. Students’ attention paid to feedback was positively related to working speed but not to the propensity to use feedback.     

The latent variable approach as applied to transitive reasoning

We discuss the limitations of hypothesis testing using (quasi-) experiments in the study of cognitive development and suggest latent variable modeling as a viable alternative to experimentation. Latent variable models allow testing a theory as a whole, incorporating individual differences with respect to developmental processes or abilities in the model. Experiments, in contrast, aim at testing hypotheses that refer to a specific part of a theory; also they ignore individual differences or model the individual differences using age group as a proxy for developmental stage. Drawing on a sample of 409 5–13-year olds, we demonstrate the advantages of latent variable models in the area of transitive reasoning. A comparison of three models showed that the latent variable model that represented fuzzy trace theory had a better fit than the models representing Piaget's theory or linear ordering theory.     

Adolescents’ Perceptions of Family Connectedness,Intrinsic Religiosity,and Depressed Mood

Using a sample of 248 ninth and tenth grade students at public high schools, we examined adolescents’ perceptions of family connectedness, intrinsic religiosity, and adolescents’ gender in relation to depressed mood and whether intrinsic religiosity and gender moderated the association of aspects of family connectedness to adolescent depressed mood. Using hierarchical multiple regression analyses we tested models separately for three forms of family connectedness (overall family cohesion, mothers’ support, and fathers’ support), intrinsic religiosity, and depressed mood. In each model, family connectedness was negatively associated with depressed mood. Intrinsic religiosity was not significantly associated with depressed mood. However, in the mothers’ support model, both a two-way interaction (mothers’ support × intrinsic religiosity) and a three-way interaction (adolescents’ gender × mothers’ support × intrinsic religiosity) were significantly related to depressed mood. In the two-way interaction, higher intrinsic religiosity was a moderator, strengthening the association between mothers’ support and depressed mood. In the three-way interaction, gender differences were found. For boys, high intrinsic religiosity strengthened the association between mothers’ support and depressed mood. Among girls, when mothers’ support was low, intrinsic religiosity provided an additional source of connectedness in protecting against depressed mood. Our findings show that connectedness in overall family systems, mother–adolescent subsystems, and father–adolescent subsystems are all important to emotional resilience in adolescents by protecting against depressed mood. Future studies of adolescent religiosity may benefit from including diverse forms of family connectedness in understanding the protective processes provided by aspects of religiosity in promoting adolescents’ emotional resilience.     

高阶项目反应模型的发展与应用

在测量具有层阶结构的潜质时, 标准项目反应模型对项目参数估计和能力参数估计都具有较低的效率, 多维项目反应模型虽然在估计第一阶潜质时具有高效性, 但没有考虑到潜质层阶的情况, 所以它不适合用来处理具有层阶结构的潜质; 而高阶项目反应模型在处理这种具有层阶结构的潜质时, 不仅能够高效准确地对项目参数和能力参数进行估计, 而且还能同时获得高阶潜质与低阶潜质。目前存在的高阶项目反应模型有高阶DINA模型、高阶双参数正态肩型层阶模型、高阶逻辑斯蒂模型、多级评分的高阶项目反应模型和高阶题组模型。未来对高阶项目反应模型的研究方向应注意多水平高阶项目反应模型、项目内多维情况下的高阶项目反应模型以及高阶认知诊断模型。     

Overgeneral autobiographical memory in depressed adolescents with, versus without, a reported history of trauma

Autobiographical memory retrieval is fundamental to the developing self-concept in adolescence, to emotion regulation, and to social problem solving and has been shown to be compromised in adolescents diagnosed with depression (R. J. Park, I. M. Goodyer, & J. D. Teasdale, 2002). The authors of the present study assessed autobiographical memory overgenerality in a sample of depressed adolescent participants with, versus without, a reported history of trauma and never-depressed control participants by using an emotion word-cueing paradigm. The authors' analyses showed for the first time that adolescents with major depression and with no reported history of trauma exhibited an overgeneral memory bias. They also revealed that depressed adolescents who reported a history of trauma retrieved fewer overgeneral memories than did depressed adolescents who reported no history of trauma. Among depressed adolescents who reported a history of trauma, more severe posttraumatic stress symptoms were associated with less overgenerality. Possible accounts of these findings are suggested.     

Structural Validation of the Hamilton Depression Rating Scale

Despite the prolific use of the Hamilton Depression Rating Scale (HDRS), there have been no prior studies using confirmatory factor analysis (CFA) to compare competing HDRS latent structures. The current study undertakes this task, examining CFAs on 397 older adults on the most stable latent models previously posited and a new rationally derived model by Cole and Motivala. CFAs were evaluated in their original first-order structure and in hierarchically related counterparts. All hierarchical models performed significantly better than their first-order counterparts. Moreover, the hierarchical Cole and Motivala model was significantly better on all comparative criteria than any of the other models. However, none of the tested models obtained sufficient fit with the data suggesting psychometric, scoring, or content limitations of the scale.