People are variables too: multilevel structural equations modeling

Estimating nonlinear effects of random slopes: A comparison of multilevel structural equation modeling with a two-step, a single-indicator, and a plausible values approach

Multilevel structural equation modeling (MSEM) is a statistical framework of major relevance for research concerned with people's intrapersonal dynamics. An application domain that is rapidly gaining relevance is the study of individual differences in the within-person association (WPA) of variables that fluctuate over time. For instance, an individual's social reactivity - their emotional response to social situations - can be represented as the association between repeated measurements of the individual's social interaction quantity and momentary well-being. MSEM allows researchers to investigate the associations between WPAs and person-level outcome variables (e.g., life satisfaction) by specifying the WPAs as random slopes in the structural equation on level 1 and using the latent representations of the slopes to predict outcomes on level 2. Here, we are concerned with the case in which a researcher is interested in nonlinear effects of WPAs on person-level outcomes - a U-shaped effect of a WPA, a moderation effect of two WPAs, or an effect of congruence between two WPAs - such that the corresponding MSEM includes latent interactions between random slopes. We evaluate the nonlinear MSEM approach for the three classes of nonlinear effects (U-shaped, moderation, congruence) and compare it with three simpler approaches: a simple two-step approach, a single-indicator approach, and a plausible values approach. We use a simulation study to compare the approaches on accuracy of parameter estimates and inference. We derive recommendations for practice and provide code templates and an illustrative example to help researchers implement the approaches.

Multilevel Semiparametric Latent Variable Modeling in R with "galamm"

We present the R package galamm, whose goal is to provide common ground between structural equation modeling and mixed effect models. It supports estimation of models with an arbitrary number of crossed or nested random effects, smoothing splines, mixed response types, factor structures, heteroscedastic residuals, and data missing at random. Implementation using sparse matrix methods and automatic differentiation ensures computational efficiency. We here briefly present the implemented methodology, give an overview of the package and an example demonstrating its use.

Epidemiology with psychometric spirit: MoBa leads autism's interdisciplinary future-a commentary on Havdahl et al. (2023)

Havdahl et al.'s (2023) Norwegian Mother, Father and Child Cohort Study (MoBa) skill loss study stands out for their creative consideration of scale items to gain a better understanding of skill loss/regression. This commentary outlines how the MoBa team continues to challenge the field by conducting "basic" measurement analyses with their public health longitudinal population data. Their creative use of items, validity-oriented analyses, and transparent reporting of item correlations emulates early-stage scale development in psychometric research, and sets the stage for considering how psychometricians and epidemiologists might more directly work with each other to improve early autism identification research.

Grandparents' educational attainment is associated with grandchildren's epigenetic-based age acceleration in the National Growth and Health Study

We examined three generations (grandparents, mothers, and grandchildren) to assess the association between grandparents' educational attainment and their grandchildren's epigenetic-based age acceleration and whether the association was mediated by parental educational attainment and mothers' life course health-related factors. Mothers were recruited to the NHLBI Growth and Health Study at 9-10 years and followed for 10 years (1987-1998). Mothers were then re-contacted three decades later (ages 37-42) to participate in the National Growth and Health Study (NGHS), and health information from their youngest child (i.e., grandchildren; N = 241, ages 2-17) was collected, including their saliva samples to calculate epigenetic age. Five epigenetic-based age acceleration measures were included in this analysis, including four epigenetic clock age accelerations (Horvath, Hannum, GrimAge, and PhenoAge) and DunedinPACE. Grandparents reported their highest education during the initial enrollment interviews. Parental educational attainment and mothers' life course health-related factors (childhood BMI trajectories, adult cardiovascular health behavioral risk score, and adult c-reactive protein) are included as mediators. Grandparents' education was significantly associated with Horvath age acceleration (b = -0.32, SE = 0.14, p = 0.021). Grandchildren with college-degree grandparents showed significantly slower Horvath age accelerations than those without college degrees. This association was partially mediated by parental education and mothers' health-related factors, especially adult cardiovascular health behavioral risk score and CRP, but not mothers' childhood BMI trajectory. This ability to conserve the speed of biological aging may have considerable consequences in shaping health trajectories across the lifespan.

A Multilevel Person-Centered Perspective on the Role of Job Demands and Resources for Employees' Job Engagement and Burnout Profiles

The present study examined the configurations, or profiles, taken by distinct global and specific facets of job engagement and burnout (by relying on a bifactor operationalization of these constructs) among a nationally representative sample of Canadian Defence employees (n = 13,088; nested within 65 work units). The present study also adopted a multilevel perspective to investigate the role of job demands (work overload and role ambiguity), as well as individual (psychological empowerment), workgroup (interpersonal justice), supervisor (transformational leadership), and organizational (organizational support) resources in the prediction of profile membership. Latent profile analyses revealed five profiles of employees: Burned-Out/Disengaged (7.13%), Burned-Out/Involved (12.13%), Engaged (18.14%), Engaged/Exhausted (15.50%), and Normative (47.10%). The highest turnover intentions were observed in the Burned-Out/Disengaged profile, and the lowest in the Engaged profile. Employees' perceptions of job demands and resources were also associated with profile membership across both levels, although the effects of psychological empowerment were more pronounced than the effects of job demands and resources related to the workgroup, supervisor, and organization. Individual-level effects were also more pronounced than effects occurring at the work unit level, where shared perceptions of work overload and organizational support proved to be the key shared drivers of profile membership.

Further exploration of the effects of time-varying covariate in growth mixture models with nonlinear trajectories

Growth mixture modeling (GMM) is an analytical tool for identifying multiple unobserved sub-populations in longitudinal processes. In particular, it describes change patterns within each latent sub-population and investigates between-individual differences in within-individual change for each sub-group. A key research interest in using GMMs is examining how covariates influence the heterogeneity in change patterns. Liu & Perera (2022b) extended mixture-of-experts (MoE) models, which primarily focus on time-invariant covariates, to allow covariates to account for both within-group and between-group differences and investigate the heterogeneity in nonlinear trajectories. The present study further extends Liu & Perera, 2022b by examining the effects of time-varying covariates (TVCs) on trajectory heterogeneity. Specifically, we propose methods to decompose a TVC into an initial trait (the baseline value of the TVC) and a set of temporal states (interval-specific slopes or changes of the TVC). The initial trait is allowed to account for within-group differences in growth factors of trajectories (i.e., baseline effect), while the temporal states are allowed to impact observed values of a longitudinal process (i.e., temporal effects). We evaluate the proposed models using a simulation study and real-world data analysis. The simulation study demonstrates that the proposed models are capable of separating trajectories into several clusters and generally producing unbiased and accurate estimates with target coverage probabilities. The proposed models reveal the heterogeneity in initial trait and temporal states of reading ability across latent classes of students' mathematics performance. Additionally, the baseline and temporal effects on mathematics development of reading ability are also heterogeneous across the clusters of students.

The impact of different data handling strategies in exploratory and confirmatory factor analysis of diary measures: an evaluation using simulated and real-world asthma nighttime symptoms diary data

BACKGROUND

Daily diaries are an important modality for patient-reported outcome assessment. They typically comprise multiple questions, so understanding their underlying structure is key to appropriate analysis and interpretation. Structural evaluation of such measures poses challenges due to the high volume of repeated measurements. Potential strategies include selecting a single day, averaging item-level observations over time, or using all data while accounting for its multilevel structure.

METHOD

The above strategies were evaluated in a simulated dataset via exploratory and confirmatory factor modelling by comparing their impact on various estimates (i.e., inter-item correlations, factor loadings, model fit). Each strategy was additionally explored using real-world data from an observational study (the Asthma Nighttime Symptoms Diary).

RESULTS

Both single day and item average strategies resulted in biased factor loadings. The former displayed lower overall bias (single day: 0.064; item average: 0.121) and mean square error (single day: 0.007; item average: 0.016) but greater frequency of incorrect factor number identification compared with the latter (single day: 46.4%; item average: 0%). Increased estimated inter-item correlations were apparent in the item-average method. Non-trivial between- and within-person variance highlighted the utility of a multilevel approach. However, convergence issues and Heywood cases were more common under the multilevel approach (90.2% and 100.0%, respectively).

CONCLUSIONS

Our findings suggest that a multilevel approach can enhance our insight when evaluating the structural properties of daily diary data; however, implementation challenges still remain. Our work offers guidance on the impact of data handling decisions in diary assessment.

A multilevel factor analysis of the short form of the Centrality of Event Scale

Introduction

The Centrality of Event Scale (CES) has frequently been used to measure the degree to which positive and negative life events are perceived central to a person's identity and life story; and previous research suggests that individuals rate their most positive memory as more central compared to their most negative one. When comparing the centrality of two (or more) memories within individuals, one needs to ensure that the CES (or its short form) is equally valid for different types of events (i.e., positive and negative) as well as on different levels of analyses (i.e., on the between-person and the within-person level), pointing to the issue of measurement invariance.

Methods

Three-hundred sixty-five adults (18-89 years of age) reported up to ten positive and up to ten negative autobiographical memories. For each memory reported, participants completed the seven-item short form of the CES, which measures three different components of centrality: Events can form a central component of identity (two items), a turning point in the life story (three items), and a reference point for everyday inferences (two items).

Results

Based on exploratory and confirmatory factor analyses, we found a two-factor structure (Self-Perception and Life-Course) to fit the data best at both levels of analyses and for both positive and negative events. Strict measurement invariance could be applied for positive and negative events at between-person level and at within-person level. The two factors, which measure the impact of an event on either a person's self-perception or their (future) life course, were rated higher for positive compared to negative memories. This difference, however, was stronger for the self-perception factor.

Discussion

The present study provides a first examination of the factorial structure of the CES short form on two levels (within and between persons) as well as for two types of life events (positive and negative). Whereas, a unidimensional scale might be sufficient to measure the centrality of stressful or traumatic life events, a more fine-graded measure seems better suited to understand the different roles of positive and negative life events for a person's identity and life story.

Using Dynamic Structural Equation Modeling to Examine Between- and Within-Persons Factor Structure of the DASS-21

The recent integration of traditional time series analysis and confirmatory factor analysis techniques allows researchers to evaluate the psychometric properties of measurement instruments at between- and within-persons levels while accounting for autoregressive dependencies. The current study applies a dynamic structural equation modeling (SEM) latent factor analysis (i.e., DSEM-CFA) to a sample of 333 individuals who completed the DASS-21 at their regular therapy sessions. The results of the DSEM-CFA illuminate the reliability, invariance, and structural features of each DASS-21 subscale both between and within persons. The results suggest that the DASS-21 reliably measures depression, anxiety, and stress symptoms when evaluating differences between persons, but does not reliably assess within-persons fluctuations in symptoms over time. The results also suggest that currently accepted methods of modeling sensitivity to change within an instrument are likely lacking and the DSEM-CFA provides insight into reliability within and between persons, which is extremely important for instruments used across time.

The Hitchhiker's guide to longitudinal models: A primer on model selection for repeated-measures methods

Longitudinal data are becoming increasingly available in developmental neuroimaging. To maximize the promise of this wealth of information on how biology, behavior, and cognition change over time, there is a need to incorporate broad and rigorous training in longitudinal methods into the repertoire of developmental neuroscientists. Fortunately, these models have an incredibly rich tradition in the broader developmental sciences that we can draw from. Here, we provide a primer on longitudinal models, written in a beginner-friendly (and slightly irreverent) manner, with a particular focus on selecting among different modeling frameworks (e.g., multilevel versus latent curve models) to build the theoretical model of development a researcher wishes to test. Our aims are three-fold: (1) lay out a heuristic framework for longitudinal model selection, (2) build a repository of references that ground each model in its tradition of methodological development and practical implementation with a focus on connecting researchers to resources outside traditional neuroimaging journals, and (3) provide practical resources in the form of a codebook companion demonstrating how to fit these models. These resources together aim to enhance training for the next generation of developmental neuroscientists by providing a solid foundation for future forays into advanced modeling applications.

Assessing mediational processes using piecewise linear growth curve models with individual measurement occasions

Longitudinal processes often unfold concurrently where the growth patterns of two or more longitudinal outcomes are associated. Additionally, if the study under investigation is long, the growth curves may exhibit nonconstant change with respect to time. Multiple existing studies have developed multivariate growth models with nonlinear functional forms to explore joint development where two longitudinal records are correlated over time. However, the relationship between multiple longitudinal outcomes may also be unidirectional. Accordingly, it is of interest to estimate regression coefficients of such unidirectional paths. One statistical tool for such analyses is longitudinal mediation models. In this study, we develop two models to evaluate mediational processes where the linear-linear piecewise functional form is utilized to capture the change patterns. We define the mediational process as either the baseline covariate or the change in covariate influencing the change in the mediator, which, in turn, affects the change in the outcome. We present the proposed models through simulation studies and real-world data analyses. Our simulation studies demonstrate that the proposed mediational models can provide unbiased and accurate point estimates with target coverage probabilities with a 95% confidence interval. The empirical analyses demonstrate that the proposed models can estimate covariates' direct and indirect effects on the change in the outcome. We also provide the corresponding code for the proposed models.

The Family Lifestyles, Actions and Risk Education (FLARE) study: Protocol for a randomized controlled trial of a sun protection intervention for children of melanoma survivors

BACKGROUND

Children of parents who had melanoma are more likely to develop skin cancer themselves owing to shared familial risks. The prevention of sunburns and promotion of sun-protective behaviors are essential to control cancer among these children. The Family Lifestyles, Actions and Risk Education (FLARE) intervention will be delivered as part of a randomized controlled trial to support parent-child collaboration to improve sun safety outcomes among children of melanoma survivors.

METHODS

FLARE is a two-arm randomized controlled trial design that will recruit dyads comprised of a parent who is a melanoma survivor and their child (aged 8-17 years). Dyads will be randomized to receive FLARE or standard skin cancer prevention education, which both entail 3 telehealth sessions with an interventionist. FLARE is guided by Social-Cognitive and Protection Motivation theories to target child sun protection behaviors through parent and child perceived risk for melanoma, problem-solving skills, and development of a family skin protection action plan to promote positive modeling of sun protection behaviors. At multiple assessments through one-year post-baseline, parents and children complete surveys to assess frequency of reported child sunburns, child sun protection behaviors and melanin-induced surface skin color change, and potential mediators of intervention effects (e.g., parent-child modeling).

CONCLUSION

The FLARE trial addresses the need for melanoma preventive interventions for children with familial risk for the disease. If efficacious, FLARE could help to mitigate familial risk for melanoma among these children by teaching practices which, if enacted, decrease sunburn occurrence and improve children's use of well-established sun protection strategies.

A comparison of the multilevel MIMIC model to the multilevel regression and mixed ANOVA model for the estimation and testing of a cross-level interaction effect: A simulation study

When observing data on a patient-reported outcome measure in, for example, clinical trials, the variables observed are often correlated and intended to measure a latent variable. In addition, such data are also often characterized by a hierarchical structure, meaning that the outcome is repeatedly measured within patients. To analyze such data, it is important to use an appropriate statistical model, such as structural equation modeling (SEM). However, researchers may rely on simpler statistical models that are applied to an aggregated data structure. For example, correlated variables are combined into one sum score that approximates a latent variable. This may have implications when, for example, the sum score consists of indicators that relate differently to the latent variable being measured. This study compares three models that can be applied to analyze such data: the multilevel multiple indicators multiple causes (ML-MIMIC) model, a univariate multilevel model, and a mixed analysis of variance (ANOVA) model. The focus is on the estimation of a cross-level interaction effect that presents the difference over time on the patient-reported outcome between two treatment groups. The ML-MIMIC model is an SEM-type model that considers the relationship between the indicators and the latent variable in a multilevel setting, whereas the univariate multilevel and mixed ANOVA model rely on sum scores to approximate the latent variable. In addition, the mixed ANOVA model uses aggregated second-level means as outcome. This study showed that the ML-MIMIC model produced unbiased cross-level interaction effect estimates when the relationships between the indicators and the latent variable being measured varied across indicators. In contrast, under similar conditions, the univariate multilevel and mixed ANOVA model underestimated the cross-level interaction effect.

Longitudinal Modeling of Age-Dependent Latent Traits with Generalized Additive Latent and Mixed Models

We present generalized additive latent and mixed models (GALAMMs) for analysis of clustered data with responses and latent variables depending smoothly on observed variables. A scalable maximum likelihood estimation algorithm is proposed, utilizing the Laplace approximation, sparse matrix computation, and automatic differentiation. Mixed response types, heteroscedasticity, and crossed random effects are naturally incorporated into the framework. The models developed were motivated by applications in cognitive neuroscience, and two case studies are presented. First, we show how GALAMMs can jointly model the complex lifespan trajectories of episodic memory, working memory, and speed/executive function, measured by the California Verbal Learning Test (CVLT), digit span tests, and Stroop tests, respectively. Next, we study the effect of socioeconomic status on brain structure, using data on education and income together with hippocampal volumes estimated by magnetic resonance imaging. By combining semiparametric estimation with latent variable modeling, GALAMMs allow a more realistic representation of how brain and cognition vary across the lifespan, while simultaneously estimating latent traits from measured items. Simulation experiments suggest that model estimates are accurate even with moderate sample sizes.

Consuming oral cannabidiol prior to a standard alcohol dose has minimal effect on breath alcohol level and subjective effects of alcohol

RATIONALE

Cannabidiol (CBD) is found in the cannabis plant and has garnered attention as a potential treatment for alcohol use disorder (AUD). CBD reduces alcohol consumption and other markers of alcohol dependence in rodents, but human research on CBD and alcohol is limited. It is unknown whether CBD reduces drinking in humans, and mechanisms through which CBD could impact behavioral AUD phenotypes are unknown.

OBJECTIVES

This study explores effects of oral CBD on breath alcohol level (BrAC), and subjective effects of alcohol in human participants who report heavy drinking.

METHODS

In this placebo-controlled, crossover study, participants consumed 30 mg CBD, 200 mg CBD, or placebo CBD before receiving a standardized alcohol dose. Participants were blind to which CBD dose they received at each session and completed sessions in random order. Thirty-six individuals completed at least one session and were included in analyses.

RESULTS

Differences in outcomes across the three conditions and by sex were explored using multilevel structural equation models. BrAC fell fastest in the placebo condition, followed by 30 mg and 200 mg CBD. Stimulation decreased more slowly in the 200 mg CBD condition than in placebo (b =  - 2.38, BCI [- 4.46, - .03]). Sedation decreased more slowly in the 30 mg CBD condition than in placebo (b =  - 2.41, BCI [- 4.61, - .09]). However, the magnitude of condition differences in BrAC and subjective effects was trivial.

CONCLUSIONS

CBD has minimal influence on BrAC and subjective effects of alcohol. Further research is needed to test whether CBD impacts alcohol consumption in humans, and if so, what mechanism(s) may explain this effect.

Multilevel SEM with random slopes in discrete data using the pairwise maximum likelihood

Pairwise maximum likelihood (PML) estimation is a promising method for multilevel models with discrete responses. Multilevel models take into account that units within a cluster tend to be more alike than units from different clusters. The pairwise likelihood is then obtained as the product of bivariate likelihoods for all within-cluster pairs of units and items. In this study, we investigate the PML estimation method with computationally intensive multilevel random intercept and random slope structural equation models (SEM) in discrete data. In pursuing this, we first reconsidered the general 'wide format' (WF) approach for SEM models and then extend the WF approach with random slopes. In a small simulation study we the determine accuracy and efficiency of the PML estimation method by varying the sample size (250, 500, 1000, 2000), response scales (two-point, four-point), and data-generating model (mediation model with three random slopes, factor model with one and two random slopes). Overall, results show that the PML estimation method is capable of estimating computationally intensive random intercept and random slopes multilevel models in the SEM framework with discrete data and many (six or more) latent variables with satisfactory accuracy and efficiency. However, the condition with 250 clusters combined with a two-point response scale shows more bias.

Addressing Adolescents' Prejudice toward Immigrants: The Role of the Classroom Context

According to social learning theory, classrooms are essential socialization contexts for intergroup attitudes, but analyses of contextual factors net of the impact of individual variables affecting prejudice toward immigrants are very limited. This study was conducted on a large sample of Italian adolescents (N = 2904; M_age = 13.70; females = 48.5%; 168 classrooms). It examined the role of classroom contextual factors affecting adolescents' prejudice toward immigrants, relying on the combination of groups' warmth and competence, and their antecedents (i.e., competition and status). Multilevel structural equation analyses revealed that classroom contextual factors (i.e., classroom socio-economic status-SES; classroom open to discussion climate; classroom educational achievements) indirectly affected, at the class level, adolescents' perceived warmth and competence of immigrants through the mediating role of perceived competition (and status) of immigrants. These findings suggest that interventions targeting the classroom context can help to hinder prejudice in adolescence at the class level.

Products of Variables in Structural Equation Models

A general method is introduced in which variables that are products of other variables in the context of a structural equation model (SEM) can be decomposed into the sources of variance due to the multiplicands. The result is a new category of SEM which we call a Products of Variables Model (PoV). Some useful and practical features of PoV models include estimation of interactions between latent variables, latent variable moderators, manifest moderators with missing values, and manifest or latent squared terms. Expected means and covariances are analytically derived for a simple product of two variables and it is shown that the method reproduces previously published results for this special case. It is shown algebraically that using centered multiplicands results in an unidentified model, but if the multiplicands have non-zero means, the result is identified. The method has been implemented in OpenMx and Ωnyx and is applied in five extensive simulations.

A Cautionary Note Regarding Multilevel Factor Score Estimates from Lavaan

To compute factor score estimates, lavaan version 0.6–12 offers the function lavPredict( ) that can not only be applied in single-level modeling but also in multilevel modeling, where characteristics of higher-level units such as working environments or team leaders are often assessed by ratings of employees. Surprisingly, the function provides results that deviate from the expected ones. Specifically, whereas the function yields correct EAP estimates of higher-level factors, the ML estimates are counterintuitive and possibly incorrect. Moreover, the function does not provide the expected standard errors. I illustrate these issues using an example from organizational research where team leaders are evaluated by their employees, and I discuss these issues from a measurement perspective.

Why Do People (Not) Take Breaks? An Investigation of Individuals' Reasons for Taking and for Not Taking Breaks at Work

Although breaks can help employees stay energized and maintain high levels of performance throughout the day, employees sometimes refrain from taking a break despite wanting to do so. Unfortunately, few studies have investigated individuals' reasons for taking and for not taking a break at work. To address this gap, we developed a model for predicting employees' break-taking behaviors. We developed hypotheses by integrating theories of work stress, self-regulation, and the results of a qualitative survey conducted as part of the current research (Study 1). Specifically, we predicted that high workloads would be positively related to the desire to detach from work, but that at the same time, high workloads would also deter employees from actually taking breaks. Furthermore, we predicted that employees would be less likely to act upon their desire to take a break within an environment where breaks are frowned upon by supervisors and coworkers, relative to an environment where breaks are allowed and encouraged. The results of a daily diary study of full-time employees (Study 2) provided general support for these predictions. Altogether, this research provides insights into the manner in which employees' psychological experiences and characteristics of the work environment combine to predict break-taking.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s10869-022-09866-4.

The Reliability Factor: Modeling Individual Reliability with Multiple Items from a Single Assessment

Reliability is a crucial concept in psychometrics. Although it is typically estimated as a single fixed quantity, previous work suggests that reliability can vary across persons, groups, and covariates. We propose a novel method for estimating and modeling case-specific reliability without repeated measurements or parallel tests. The proposed method employs a "Reliability Factor" that models the error variance of each case across multiple indicators, thereby producing case-specific reliability estimates. Additionally, we use Gaussian process modeling to estimate a nonlinear, non-monotonic function between the latent factor itself and the reliability of the measure, providing an analogue to test information functions in item response theory. The reliability factor model is a new tool for examining latent regions with poor conditional reliability, and correlates thereof, in a classical test theory framework.

Protocol for a mobile laboratory study of co-administration of cannabis concentrates with a standard alcohol dose in humans

Cannabis is commonly used among people who drink alcohol, yet evidence on acute effects of co-use is conflicting. Two important variables that may influence the effects of cannabis and alcohol are cannabinoid content (i.e., the ratio of cannabidiol [CBD] and 9-tetrahydrocannabinol [THC]) as well as the order of use (i.e., cannabis before alcohol vs. alcohol before cannabis). Research is mixed regarding the acute imapct of cannabis on alcohol consumption and intoxication, with some studies suggesting additive effects of alcohol and cannabis, and others demonstrating negligible effects of combining these substances. Further complicating this, high-THC-content cannabis concentrates are increasingly popular on the legal-market, but to our knowledge, no studies have explored concentrate and alcohol co-use. In addition to cannabinoid content, order of use may influence intoxication and other acute effects, but is also understudied. Co-use studies typically administer a fixed dose of alcohol before cannabis, and there is a lack of data on the acute effects of cannabis before alcohol. Thus, there is a need for experimental co-use studies exploring the impact of cannabinoid content (particularly of highly potent cannabis concentrates) and order effects on intoxication. This study uses a federally-compliant mobile laboratory procedure to explore the effects of co-administration of legal-market cannabis concentrates with a moderate alcohol dose (.8g/kg) in a sample of community participants who regularly use alcohol and cannabis. The study will also explore alcohol and cannabis order effects (cannabis before alcohol vs. alcohol before cannabis). Outcomes are objective intoxication (measured using blood cannabinoid level, heart rate, psychomotor performance and breath alcohol level [BrAC]) and subjective intoxication (assessed via self-report measures). Overall, this study may influence harm-reduction recommendations for individuals who drink alcohol and use cannabis.

Emotional Intelligence as a Personality State

Abstract. Contemporary research has begun to explore the notion that emotional intelligence (EI) has an important state component in addition to the trait component, as represented in the whole trait theory. This implies that state EI (or enacted EI) has similar cognitive, affective, and motivational contents as its corresponding trait. The question, however, of whether a trait EI construct means the same across the individual (trait) and state levels of analysis has not been empirically investigated. To address this gap, the present study examines the assessment of enacted EI, using the full version of the Wong and Law Emotional Intelligence Scale (WLEIS) on both between-person and within-person levels of analysis. Participants were 493 Greek employees who completed the WLEIS for 5 consecutive workdays. Multilevel confirmatory factor analyses confirmed that the original four-factor multilevel model appeared to best fit the data. Multilevel measurement invariance analysis supported the equivalence of the measure across different levels of analysis. In conclusion, the WLEIS is a configural cluster construct, believed to be a valuable and reliable tool for assessing enacted EI within the workplace. Implications for future research on enacted EI are discussed.

China’s R&D Investment’s Impact on Environmental Pollution: An Integrated Approach Based on Panel Moderated Mediation and Regression Discontinuity

How to reduce environmental pollution is fundamental for current civilization. Increasing in R&D investment may reduce the environmental pollution, yet whether and how R&D investment influence the environmental pollution needs further discussion and verification. Considering that the R&D investment can directly and also indirectly influence the environmental pollution by affecting the economic growth, and the fact that there is an obvious discontinuity for economic growth during the observed period, this paper firstly proposes an integrated approach based on panel moderated mediation analysis and regression discontinuity. It examines the R&D impact on environmental pollution on the basis of province-level data in China and uses the integrated approach to test its direct and indirect effect. Finally conclusion is made according to the findings.

The Sampling Ratio in Multilevel Structural Equation Models: Considerations to Inform Study Design

Multilevel structural equation modeling (MSEM) allows researchers to model latent factor structures at multiple levels simultaneously by decomposing within- and between-group variation. Yet the extent to which the sampling ratio (i.e., proportion of cases sampled from each group) influences the results of MSEM models remains unknown. This article explores how variation in the sampling ratio in MSEM affects the measurement of Level 2 (L2) latent constructs. Specifically, we investigated whether the sampling ratio is related to bias and variability in aggregated L2 construct measurement and estimation in the context of doubly latent MSEM models utilizing a two-step Monte Carlo simulation study. Findings suggest that while lower sampling ratios were related to increased bias, standard errors, and root mean square error, the overall size of these errors was negligible, making the doubly latent model an appealing choice for researchers. An applied example using empirical survey data is further provided to illustrate the application and interpretation of the model. We conclude by considering the implications of various sampling ratios on the design of MSEM studies, with a particular focus on educational research.

The Impact of Genes and Environment on Brain Ageing in Males Aged 51 to 72 Years

Magnetic resonance imaging data are being used in statistical models to predicted brain ageing (PBA) and as biomarkers for neurodegenerative diseases such as Alzheimer's Disease. Despite their increasing application, the genetic and environmental etiology of global PBA indices is unknown. Likewise, the degree to which genetic influences in PBA are longitudinally stable and how PBA changes over time are also unknown. We analyzed data from 734 men from the Vietnam Era Twin Study of Aging with repeated MRI assessments between the ages 51-72 years. Biometrical genetic analyses "twin models" revealed significant and highly correlated estimates of additive genetic heritability ranging from 59 to 75%. Multivariate longitudinal modeling revealed that covariation between PBA at different timepoints could be explained by a single latent factor with 73% heritability. Our results suggest that genetic influences on PBA are detectable in midlife or earlier, are longitudinally very stable, and are largely explained by common genetic influences.

Public Negative Labeling Effects on Team Interaction and Performance

Across four studies, we examine how public negative labeling, which is when a group member is publicly identified as bad, affects team performance. Across three experiments and one field study, we test and find support for our model, that public negative labeling undermines team performance via reduced perceptions of team interaction quality. Our study contributes to the expansive conversation on team effectiveness which highlights that “fighting fire with fire” in terms of public negative labeling is ineffective for dealing with uncivil workplace behavior.

A Bayesian EAP-Based Nonlinear Extension of Croon and Van Veldhoven’s Model for Analyzing Data from Micro–Macro Multilevel Designs

Croon and van Veldhoven discussed a model for analyzing micro–macro multilevel designs in which a variable measured at the upper level is predicted by an explanatory variable that is measured at the lower level. Additionally, the authors proposed an approach for estimating this model. In their approach, estimation is carried out by running a regression analysis on Bayesian Expected a Posterior (EAP) estimates. In this article, we present an extension of this approach to interaction and quadratic effects of explanatory variables. Specifically, we define the Bayesian EAPs, discuss a way for estimating them, and we show how their estimates can be used to obtain the interaction and the quadratic effects. We present the results of a “proof of concept” via Monte Carlo simulation, which we conducted to validate our approach and to compare two resampling procedures for obtaining standard errors. Finally, we discuss limitations of our proposed extended Bayesian EAP-based approach.

Catching Up on Multilevel Modeling

This review focuses on the use of multilevel models in psychology and other social sciences. We target readers who are catching up on current best practices and sources of controversy in the specification of multilevel models. We first describe common use cases for clustered, longitudinal, and cross-classified designs, as well as their combinations. Using examples from both clustered and longitudinal designs, we then address issues of centering for observed predictor variables: its use in creating interpretable fixed and random effects of predictors, its relationship to endogeneity problems (correlations between predictors and model error terms), and its translation into multivariate multilevel models (using latent-centering within multilevel structural equation models). Finally, we describe novel extensions—mixed-effects location–scale models—designed for predicting differential amounts of variability.

A multilevel structural equation model for assessing a drug effect on a patient-reported outcome measure in on-demand medication data

We analyze data from a clinical trial investigating the effect of an on-demand drug for women with low sexual desire. These data consist of a varying number of measurements/events across patients of when the drug was taken, including data on a patient-reported outcome consisting of five items measuring an unobserved construct (latent variable). Traditionally, these data are aggregated prior to analysis by composing one sum score per event and averaging this sum score over all observed events. In this paper, we explain the drawbacks of this aggregating approach. One drawback is that these averages have different standard errors because the variance of the underlying events differs between patients and because the number of events per patient differs. Another drawback is the implicit assumption that all items have equal weight in relation to the latent variable being measured. We propose a multilevel structural equation model, treating the events (level 1) as nested observations within patients (level 2), as alternative analysis method to overcome these drawbacks. The model we apply includes a factor model measuring a latent variable at the level of the event and at the level of the patient. Then, in the same model, the latent variables are regressed on covariates to assess the drug effect. We discuss the inferences obtained about the efficacy of the on-demand drug using our proposed model. We further illustrate how to test for measurement invariance across grouping covariates and levels using the same model.

Loyalty of Paratransit Users in the Era of Competition with Ride Sourcing

As a mode serving urban mobility in developing countries, paratransit is experiencing competition with the fast and massive growth of ride sourcing. This study aims to investigate the loyalty of paratransit users in Bandung, Indonesia, in the era of competition with ride sourcing. Data collected by questionnaires and distributed to 400 paratransit users in Bandung were analysed using hierarchical structural equation modelling. The analysis shows that satisfaction is positively influenced by service quality variables. Though there is less satisfaction compared to the previous decade, satisfaction with the service still has a positive influence on total satisfaction with the mode. It can also be concluded that loyalty is positively influenced by satisfaction. Satisfaction with the image and the unique characteristics of paratransit retains passengers’ intentions toward loyalty in the future, including current personal characteristics (i.e., age, income, occupation) and travel characteristics (i.e., fare, vehicle ownership) of passengers. This study warns of the high probability of mode-changing behaviour from paratransit to another mode.

Longitudinal Relationships Between Subjective Cognitive Decline and Objective Memory: Depressive Symptoms Mediate Between-Person Associations

BACKGROUND

Subjective cognitive decline (SCD) may be an early indicator of cognitive impairment, but depressive symptoms can confound this relationship. Associations may be influenced by differences between individuals (i.e., between-persons) or how each individual changes in their experiences over time (i.e., within-persons).

OBJECTIVE

We examined depressive symptoms as a mediator of the between- and within-person associations of SCD and objective memory in older adults.

METHODS

Coordinated analyses were conducted across four datasets drawn from large longitudinal studies. Samples (range: n = 1,889 to n = 15,841) included participants 65 years of age or older with no dementia at baseline. We used multilevel structural equation modeling to examine the mediation of SCD and objective memory through depressive symptoms, as well as direct relationships among SCD, objective memory, and depressive symptoms.

RESULTS

Older adults who were more likely to report SCD had lower objective memory on average (between-person associations), and depressive symptoms partially mediated this relationship in three of four datasets. However, changes in depressive symptoms did not mediate the relationship between reports of SCD and declines in objective memory in three of four datasets (within-person associations).

CONCLUSION

Individual differences in depressive symptoms, and not changes in an individual's depressive symptoms over time, partially explain the link between SCD and objective memory. Older adults with SCD and depressive symptoms may be at greater risk for poor cognitive outcomes. Future research should explore how perceived changes in memory affect other aspects of psychological well-being, and how these relationships influence cognitive decline and Alzheimer's disease risk.

Multi-Level Multi-Growth Models: New opportunities for addressing developmental theory using advanced longitudinal designs with planned missingness

Longitudinal models have become increasingly popular in recent years because of their power to test theoretically derived hypotheses by modeling within-person processes with repeated measures. Growth models constitute a flexible framework for modeling a range of complex trajectories across time in outcomes of interest, including non-linearities and time-varying covariates. However, these models can be expanded to include the effects of multiple growth processes at once on a single outcome. Here, I outline such an extension, showing how multiple growth processes can be modeled as a specific case of the general ability to include time-varying covariates in growth models. I show that this extension of growth models cannot be accomplished by statistical models alone, and that study design plays a crucial role in allowing for proper parameter recovery. I demonstrate these principles through simulations to mimic important theoretical conditions where modeling the effects of multiple growth processes can address developmental theory including, disaggregating the effects of age and practice or treatment in repeated assessments and modeling age- and puberty-related effects during adolescence. I compare how these models behave in two common longitudinal designs, cohort and accelerated, and how planned missingness in observations is key to parameter recovery. I conclude with directions for future substantive research using the method outlined here.

The impact of information integration on purchase order finance and new product launch: a case study

Purpose

In this study, we examine the implementation of purchase order finance (POF) which is an innovative supply chain finance (SCF) solution by an innovative SCF lender (i.e. supply chain service provider (SCSP)). The effect of information integration between the SCSP (lender) and product designers (borrowers) on the lender's POF decisions and the borrowers' new product launch is investigated.

Design/methodology/approach

We conduct a case study in the Chinese smartphone industry. A mixed methods design is used, and data are collected from both the supply chain service provider (SCSP) and product designers. We first conduct a qualitative study. Hypotheses are developed concerning the relationships between information integration, in terms of social interaction and information system integration, POF and new product launch. We then conduct a quantitative study. The multilevel structural equation modelling method is used to test the hypotheses.

Findings

We find that information system integration is positively associated with POF but has no significant effect on new product launch. Social interaction is negatively associated with POF but positively associated with new product launch. POF is positively associated with new product launch.

Originality/value

This study contributes to the literature by empirically examining the implementation of POF from both the lender's and borrower's perspectives. We find that information system integration and social interaction have different effects on POF and new product launch. The results thus provide insights into how a lender makes POF decisions and reveal the benefits of POF for borrowers.

Evaluating Cluster-Level Factor Models with lavaan and Mplus

Background: Researchers frequently use the responses of individuals in clusters to measure cluster-level constructs. Examples are the use of student evaluations to measure teaching quality, or the use of employee ratings of organizational climate. In earlier research, Stapleton and Johnson (2019) provided advice for measuring cluster-level constructs based on a simulation study with inadvertently confounded design factors. We extended their simulation study using both Mplus and lavaan to reveal how their conclusions were dependent on their study conditions. Methods: We generated data sets from the so-called configural model and the simultaneous shared-and-configural model, both with and without nonzero residual variances at the cluster level. We fitted models to these data sets using different maximum likelihood estimation algorithms. Results: Stapleton and Johnson’s results were highly contingent on their confounded design factors. Convergence rates could be very different across algorithms, depending on whether between-level residual variances were zero in the population or in the fitted model. We discovered a worrying convergence issue with the default settings in Mplus, resulting in seemingly converged solutions that are actually not. Rejection rates of the normal-theory test statistic were as expected, while rejection rates of the scaled test statistic were seriously inflated in several conditions. Conclusions: The defaults in Mplus carry specific risks that are easily checked but not well advertised. Our results also shine a different light on earlier advice on the use of measurement models for shared factors.

Using Dyadic Modeling in Nursing Research: Introduction of Theory and Application

Use of dyadic modeling in nursing has theoretical and practical importance, as the interpersonal processes related to health behaviors can be captured. Theoretical models focusing on dyadic coping with chronic illness and illness management are established in family nursing. However, few studies utilized dyadic designs in empirical research, as most studies are patient-centric or care partner-centric. With theoretical elaborations and examples, we first review how conventional health models have been extended using a dyadic perspective and then briefly review the major dyadic frameworks in nursing. Five frequently used dyadic models are described with examples from health and nursing research fields. Statistical applications and cultural considerations are reviewed. We conclude that dyadic modeling provides a useful lens for nursing research but continues to be underutilized.

Stability and change in vocational interest profiles and interest congruence over the course of vocational education and training

The current study is concerned with the stability of and changes in vocational interest profiles and interest congruence in vocational education and training (VET). Specifically, we examined (1) the stability of vocational interest profiles, (2) the existence of occupational socialization effects that manifest themselves as increases in person-environment (P-E) congruence, and (3) the question of whether or not changes in P-E congruence are psychologically relevant because they are related to trainees’ attitudes towards their VET course. We used data from a three-wave longitudinal sample comprising N = 2611 trainees from five different VET courses in Germany. Through the use of meta-analytical aggregation techniques, we were able to analyze interindividual differences in intraindividual interest stability and P-E congruence and to relate these differences to trainees’ satisfaction with VET. On average, interest profiles turned out to be highly stable over the entire course of VET. However, we found substantial interindividual and intergroup differences in interest stability. Average P-E congruence increased slightly in two groups, providing only little evidence for the presumed socialization effects. Nevertheless, interindividual differences in P-E congruence and changes in P-E congruence were psychologically relevant because they were linked to trainees’ satisfaction with their VET course and changes therein.

None the wiser: Year-long longitudinal study on effects of adversity on wisdom

Research on consequences of adversity appears inconclusive. Adversity can be detriment to mental health, promoting maladaptive patterns of thoughts. At the same time, posttraumatic growth studies suggest that overcoming major adversity facilitates growth in wisdom-related patterns of thoughts. We address this puzzle by examining how distinct types of adversity impact wisdom over time and how individual differences in self-distanced (rather than self-immersed) reflection on adversity relate to different wisdom trajectories. In a four-wave prospective year-long study, participants ( N = 499) recalled and reflected every three months on the most significant recent adverse event in their life. They reported how much they engaged in wise reasoning—intellectual humility, open-mindedness to diverse perspectives and change, search for compromises and resolution—as well as self-distancing during reflections. Independent raters identified seven distinct adversity types (e.g. social conflict, economic hardship, major trauma) in open-ended descriptions. Growth curve analyses revealed little evidence of positive change in wise-reasoning over the course of a year, and some evidence of negative change for health-related adversity. Although self-distancing was associated with stability in wisdom, self-immersing was associated with negative change in wisdom in reflections on social conflicts over time. We discuss implications these results have for adversity, change vs. resilience in character strengths, and self-distancing.

Comparing the Effects of Reading Intervention Versus Reading and Mindset Intervention for Upper Elementary Students With Reading Difficulties

The primary purpose of this study was to examine the effects of providing mindset intervention in addition to reading intervention compared with only reading intervention for fourth graders with reading difficulties. Reading intervention was provided daily in 45 min sessions throughout the school year. Mindset intervention occurred in small groups for 24-30 min lessons. Multilevel structural equation modeling (SEM) via n-level SEM was used to account for the latent variable representation of constructs, and the complex nesting and cross-classification structure of the data. Students in the reading intervention plus mindset condition significantly outperformed the business as usual condition on nonword reading (d = 0.35) as did students in the reading intervention condition (d = 0.20), who also outperformed the business as usual condition on phonological processing (d = 0.28). There were no significant differences among students in the three conditions on nonword reading, word reading, phonological processing, reading comprehension, or growth mindset. Initial reading achievement, mindset, and problem behavior did not generally moderate these findings.

Self-Critical and Self-Punishment Cognitions Differentiate Those With and Without a History of Nonsuicidal Self-Injury: An Ecological Momentary Assessment Study

The aim of this study was to examine trait, state, and temporal instability measures of self-critical and self-punishment cognitions to evaluate their respective roles in nonsuicidal self-injury (NSSI). Participants were university students with a history of NSSI (n = 64) and those with no history of NSSI (n = 59). At baseline, participants completed measures assessing history of NSSI behavior, as well as trait measures of self-criticism and self-punishment. After completion of baseline procedures, participants subsequently participated in a 10-day ecological momentary assessment protocol in which self-critical and self-punishment cognitions were assessed in real time three times daily. Employing bivariate and multivariate frameworks, our results demonstrate that both trait and state levels of self-critical and self-punishment cognitions robustly differentiate between young adults with and without a lifetime history of NSSI. The present results also confirm that the temporal instability of these cognitive states also meaningfully differentiate between groups, such that those who exhibit greater fluctuations in these cognitive states are more likely to have a history of NSSI. The current findings suggest that trait, state, and temporal instability of negative self-focused cognitions may be vulnerability factors for engagement in NSSI.

Multivariate Generalized Linear Mixed-Effects Models for the Analysis of Clinical Trial-Based Cost-Effectiveness Data

Economic evaluations conducted alongside randomized controlled trials are a popular vehicle for generating high-quality evidence on the incremental cost-effectiveness of competing health care interventions. Typically, in these studies, resource use (and by extension, economic costs) and clinical (or preference-based health) outcomes data are collected prospectively for trial participants to estimate the joint distribution of incremental costs and incremental benefits associated with the intervention. In this article, we extend the generalized linear mixed-model framework to enable simultaneous modeling of multiple outcomes of mixed data types, such as those typically encountered in trial-based economic evaluations, taking into account correlation of outcomes due to repeated measurements on the same individual and other clustering effects. We provide new wrapper functions to estimate the models in Stata and R by maximum and restricted maximum quasi-likelihood and compare the performance of the new routines with alternative implementations across a range of statistical programming packages. Empirical applications using observed and simulated data from clinical trials suggest that the new methods produce broadly similar results as compared with Stata’s merlin and gsem commands and a Bayesian implementation in WinBUGS. We highlight that, although these empirical applications primarily focus on trial-based economic evaluations, the new methods presented can be generalized to other health economic investigations characterized by multivariate hierarchical data structures.

The long shadow of childhood trauma for depression in midlife: examining daily psychological stress processes as a persistent risk pathway

BACKGROUND

Childhood trauma (CT) increases the risk of adult depression. Buffering effects require an understanding of the underlying persistent risk pathways. This study examined whether daily psychological stress processes - how an individual interprets and affectively responds to minor everyday events - mediate the effect of CT on adult depressive symptoms.

METHODS

Middle-aged women (N = 183) reported CT at baseline and completed daily diaries of threat appraisals and negative evening affect for 7 days at baseline, 9, and 18 months. Depressive symptoms were measured across the 1.5-year period. Mediation was examined using multilevel structural equation modeling.

RESULTS

Reported CT predicted greater depressive symptoms over the 1.5-year time period (estimate = 0.27, s.e. = 0.07, 95% CI 0.15-0.38, p < 0.001). Daily threat appraisals and negative affect mediated the effect of reported CT on depressive symptoms (estimate = 0.34, s.e. = 0.08, 95% CI 0.22-0.46, p < 0.001). Daily threat appraisals explained more than half of this effect (estimate = 0.19, s.e. = 0.07, 95% CI 0.08-0.30, p = 0.004). Post hoc analyses in individuals who reported at least moderate severity of CT showed that lower threat appraisals buffered depressive symptoms. A similar pattern was found in individuals who reported no/low severity of CT.

CONCLUSIONS

A reported history of CT acts as a latent vulnerability, exaggerating threat appraisals of everyday events, which trigger greater negative evening affect - processes that have important mental health consequences and may provide malleable intervention targets.