Comparing type 1 and type 2 error rates of different tests for heterogeneous treatment effects.

Psychologists are increasingly interested in whether treatment effects vary in randomized controlled trials. A number of tests have been proposed in the causal inference literature to test for such heterogeneity, which differ in the sample statistic they use (either using the variance terms of the experimental and control group, their empirical distribution functions, or specific quantiles), and in whether they make distributional assumptions or are based on a Fisher randomization procedure. In this manuscript, we present the results of a simulation study in which we examine the performance of the different tests while varying the amount of treatment effect heterogeneity, the type of underlying distribution, the sample size, and whether an additional covariate is considered. Altogether, our results suggest that researchers should use a randomization test to optimally control for type 1 errors. Furthermore, all tests studied are associated with low power in case of small and moderate samples even when the heterogeneity of the treatment effect is substantial. This suggests that current tests for treatment effect heterogeneity require much larger samples than those collected in current research.

Narcissists' affective well-being: Associations of grandiose narcissism with state affect level and variability.

Whereas grandiose narcissism has generally been found to be related to adaptive affective experiences (i.e., positive affective states), many theoretical conceptualizations have emphasized its associations with characteristics of low affective well-being (i.e., unstable, highly variable affective states). Empirical research on the association of grandiose narcissism with the mean level of and variability in affective states has been inconclusive, as studies have differed considerably in their conceptualizations and measurement of narcissism and affect dynamics and have suffered from methodological limitations. Here, we offer conceptual explanations for previously inconsistent findings, derive diverging hypotheses about different aspects of narcissism and affective well-being, and investigate these hypotheses in two daily diary and three experience-sampling data sets (overall N = 2,125; total measurements = 116,336). As hypothesized, we found diverging associations between agentic and antagonistic aspects of narcissism with affect levels: Whereas narcissistic admiration was related to more pleasant affective states, narcissistic rivalry was related to less pleasant ones. We also obtained some support for diverging effects of admiration and rivalry on affect variability. However, these associations were largely reduced when we corrected for (squared) mean levels of affective valence and arousal. In combination, these findings suggest that only the agentic aspect of grandiose narcissism is conducive to affective well-being, whereas its antagonistic aspect negatively influences affective well-being. Moreover, the assumed associations of grandiose narcissism with volatile affectivity seem to rely heavily on mean-level effects and primarily manifest in experiences of more diverse affective states rather than stronger or more frequent affective fluctuations in general. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

A screen-time-based mixture model for identifying and monitoring careless and insufficient effort responding in ecological momentary assessment data.

Ecological momentary assessment (EMA) involves repeated real-time sampling of respondents' current behaviors and experiences. The intensive repeated assessment imposes an increased burden on respondents, rendering EMAs vulnerable to respondent noncompliance and/or careless and insufficient effort responding (C/IER). We developed a mixture modeling approach that equips researchers with a tool for (a) gauging the degree of C/IER contamination of their EMA data and (b) studying the trajectory of C/IER across the study. For separating attentive from C/IER behavior, the approach leverages collateral information from screen times, which are routinely recorded in electronically administered EMAs, and translates theoretical considerations on respondents' behavior into component models for attentive and careless screen times as well as for the functional form of C/IER trajectories. We show how a sensible choice of component models (a) allows disentangling short screen times due to C/IER from familiarity effects due to repeated exposure to the same measures, (b) aids in gaining a fine-grained understanding of C/IER trajectories by distinguishing within-day from between-day effects, and (c) allows investigating interindividual differences in attentiveness. The approach shows good parameter recovery when attentive and C/IER screen time distributions exhibit sufficient separation and yields valid conclusions even in scenarios of uncontaminated data. The approach is illustrated on EMA data from the German Socio-Economic Panel innovation sample. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Three facets of emotion regulation in old and very old age: Strategy use, effectiveness, and variability.

The ability to regulate emotions in stressful situations is an important building block for high well-being across the lifespan. Yet, very little is known about how old and very old adults regulate their emotions. In this study, 123 young old adults (Mage = 67.18, SD = 0.94) and 47 very old adults (Mage = 86.70, SD = 1.46) were prompted 6 times a day for 7 consecutive days to report both their stressors and 10 emotion regulation strategies. Overall, there was little indication of age differences in the use of emotion regulation strategies during exposure to stressors, but very old, as compared with young old, individuals used three of the 10 strategies considered here more intensively. The 10 emotion regulation strategies were similarly effective across age groups based on their association with perceived overall emotion regulation success. We also did not find age group differences in within-strategy variability, defined as the variation in using a given strategy across stressor situations. By contrast, between-strategy variability, defined as the selective use of fewer rather than many strategies across stressor situations, was lower for very old participants. Only between-strategy, and not within-strategy, variability contributed to overall emotion regulation success. There was no age group difference in this regard. Taken together, the evidence suggests small age differences in emotion regulation if at all. This is noteworthy given the advanced age of the very old subsample in this study and the deficits in multiple domains of functioning reported in the literature for this advanced age. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

A Mixed-Effects Model in Which the Parameters of the Autocorrelated Error Structure Can Differ between Individuals.

Research in psychology has seen a rapid increase in the usage of experience sampling methods and daily diary methods. The data that result from using these methods are typically analyzed with a mixed-effects or a multilevel model because it allows testing hypotheses about the time course of the longitudinally assessed variable or the influence of time-varying predictors in a simple way. Here, we describe an extension of this model that does not only allow to include random effects for the mean structure but also for the residual variance, for the parameter of an autoregressive process of order 1 and/or the parameter of a moving average process of order 1. After we have introduced this extension, we show how to estimate the parameters with maximum likelihood. Because the likelihood function contains complex integrals, we suggest using adaptive Gauss-Hermite quadrature and Quasi-Monte Carlo integration to approximate it. We illustrate the models using a real data example and also report the results of a small simulation study in which the two integral approximation methods are compared.

Vascular risk factors, white matter microstructure, and depressive symptoms: a longitudinal analysis in the UK Biobank.

BACKGROUND: Cumulative burden from vascular risk factors (VRFs) has been associated with an increased risk of depressive symptoms in mid- and later life. It has been hypothesised that this association arises because VRFs disconnect fronto-subcortical white matter tracts involved in mood regulation, which puts older adults at higher risk of developing depressive symptoms. However, evidence for the hypothesis that disconnection of white matter tracts underlies the association between VRF burden and depressive symptoms from longitudinal studies is scarce. METHODS: This preregistered study analysed longitudinal data from 6,964 middle-aged and older adults from the UK Biobank who participated in consecutive assessments of VRFs, brain imaging, and depressive symptoms. Using mediation modelling, we directly tested to what extend white matter microstructure mediates the longitudinal association between VRF burden and depressive symptoms. RESULTS: VRF burden showed a small association with depressive symptoms at follow-up. However, there was no evidence that fractional anisotropy (FA) of white matter tracts mediated this association. Additional analyses also yielded no mediating effects using alternative operationalisations of VRF burden, mean diffusivity (MD) of single tracts, or overall average of tract-based white matter microstructure (global FA, global MD, white matter hyperintensity volume). CONCLUSIONS: Our results lend no support to the hypothesis that disconnection of white matter tracts underlies the association between VRF burden and depressive symptoms, while highlighting the relevance of using longitudinal data to directly test pathways linking vascular and mental health.

A latent variable mixed-effects location scale model that also considers between-person differences in the autocorrelation.

In public health research an increasing number of studies is conducted in which intensive longitudinal data is collected in an experience sampling or a daily diary design. Typically, the resulting data is analyzed with a mixed-effects model or mixed-effects location scale model because they allow one to examine a host of interesting longitudinal research questions. Here, we introduce an extension of the mixed-effects location scale model in which measurement error of the observed variables is considered by a latent factor model and in which-in addition to the mean-or location-related effects-the residual variance of the latent factor and the parameters of the autoregressive process of this latent factor can differ between persons. We show how to estimate the parameters of the model with a maximum likelihood approach, whose performance is also compared with a Bayesian approach in a small simulation study. We illustrate the models using a real data example and end with a discussion in which we suggest questions for future research.

A Tutorial Introduction to Heterogeneous Treatment Effect Estimation with Meta-learners.

Psychotherapy has been proven to be effective on average, though patients respond very differently to treatment. Understanding which characteristics are associated with treatment effect heterogeneity can help to customize therapy to the individual patient. In this tutorial, we describe different meta-learners, which are flexible algorithms that can be used to estimate personalized treatment effects. More specifically, meta-learners decompose treatment effect estimation into multiple prediction tasks, each of which can be solved by any machine learning model. We begin by reviewing necessary assumptions for interpreting the estimated treatment effects as causal, and then give an overview over key concepts of machine learning. Throughout the article, we use an illustrative data example to show how the different meta-learners can be implemented in R. We also point out how current popular practices in psychotherapy research fit into the meta-learning framework. Finally, we show how heterogeneous treatment effects can be analyzed, and point out some challenges in the implementation of meta-learners.

Mediation of the Association Between Vascular Risk Factors and Depressive Symptoms by C-Reactive Protein.

Background: This study examined whether C-reactive protein (CRP), a marker of low-grade systemic inflammation, mediates the association between vascular risk factor (VRF) burden and depressive symptoms. Methods: We drew on the prospective design of the UK Biobank to include participants with longitudinal data on VRF burden, CRP, and depressive symptoms. Total, direct, and indirect effects were estimated using regression-based mediation models while controlling for confounding by sociodemographic factors, baseline CRP, and baseline depression. Sensitivity analyses probed the robustness of results to unmeasured confounding. Results: We analyzed data from 10,470 participants from the UK Biobank (mean age = 56.75 years at baseline). Net of covariates, VRFs at baseline were associated with higher depressive symptoms at follow-up (total effect = 0.099; 95% CI, 0.002-0.163). CRP mediated this association (indirect effect = 0.010; 95% CI, 0.004-0.017), accounting for 10.0% (95% CI, 0.3%-30.0%) of the total effect of VRF burden on depressive symptoms. Exploratory analyses suggested that the total and indirect effects pertained to somatic depressive symptoms (tiredness and appetite). Conclusions: These results suggest that inflammation-promoting effects of VRFs may contribute to depressive symptoms in mid- and later life. However, the mediating pathway via CRP explains only a small part of the association between VRFs and depression after accounting for important covariates and may pertain to specific depressive symptoms. Future studies leveraging similar longitudinal designs are needed to further disentangle the time-varying effects between VRFs, inflammation, and certain depressive symptoms while addressing important confounders.

Routes to momentary self-esteem in adolescence: Links with interpersonal perceptions of liking and personality metaperceptions within social interactions.

OBJECTIVE: This study provides insights into the perceptual processes contributing to positive social experiences and momentary self-esteem in adolescents' initial social interactions. BACKGROUND: A person's self-esteem is shaped by their social experiences. However, little is known about which interpersonal perceptions are linked to momentary self-esteem within social interactions. Identifying two key interpersonal perceptions, we examined differential associations between an individual's meta-liking and other-perceptions of liking by interaction partners with momentary self-esteem. Further, we investigated how preceding personality metaperceptions (i.e., meta-accuracy and meta-positivity) extended these sociometer processes. METHOD: The current study used round-robin data from N = 296 adolescents participating in a virtual group interaction. Social accuracy modeling was used to estimate meta-accuracy and meta-positivity of personality metaperceptions, and path models were used to test associations with interpersonal perceptions of liking and momentary self-esteem. RESULTS: Two main findings emerged: First, meta-liking was consistently related to higher momentary self-esteem, whereas other-perceptions of liking were not. Second, meta-positivity of personality metaperceptions was linked to higher meta-liking and indirectly contributed to higher momentary self-esteem through meta-liking. CONCLUSION: These findings highlight the importance of considering different interpersonal perceptions to understand social interaction experiences.

Parametric and nonparametric propensity score estimation in multilevel observational studies.

There has been growing interest in using nonparametric machine learning approaches for propensity score estimation in order to foster robustness against misspecification of the propensity score model. However, the vast majority of studies focused on single-level data settings, and research on nonparametric propensity score estimation in clustered data settings is scarce. In this article, we extend existing research by describing a general algorithm for incorporating random effects into a machine learning model, which we implemented for generalized boosted modeling (GBM). In a simulation study, we investigated the performance of logistic regression, GBM, and Bayesian additive regression trees for inverse probability of treatment weighting (IPW) when the data are clustered, the treatment exposure mechanism is nonlinear, and unmeasured cluster-level confounding is present. For each approach, we compared fixed and random effects propensity score models to single-level models and evaluated their use in both marginal and clustered IPW. We additionally investigated the performance of the standard Super Learner and the balance Super Learner. The results showed that when there was no unmeasured confounding, logistic regression resulted in moderate bias in both marginal and clustered IPW, whereas the nonparametric approaches were unbiased. In presence of cluster-level confounding, fixed and random effects models greatly reduced bias compared to single-level models in marginal IPW, with fixed effects GBM and fixed effects logistic regression performing best. Finally, clustered IPW was overall preferable to marginal IPW and the balance Super Learner outperformed the standard Super Learner, though neither worked as well as their best candidate model.

Random Effects Multinomial Processing Tree Models: A Maximum Likelihood Approach.

The present article proposes and evaluates marginal maximum likelihood (ML) estimation methods for hierarchical multinomial processing tree (MPT) models with random and fixed effects. We assume that an identifiable MPT model with S parameters holds for each participant. Of these S parameters, R parameters are assumed to vary randomly between participants, and the remaining [Formula: see text] parameters are assumed to be fixed. We also propose an extended version of the model that includes effects of covariates on MPT model parameters. Because the likelihood functions of both versions of the model are too complex to be tractable, we propose three numerical methods to approximate the integrals that occur in the likelihood function, namely, the Laplace approximation (LA), adaptive Gauss-Hermite quadrature (AGHQ), and Quasi Monte Carlo (QMC) integration. We compare these three methods in a simulation study and show that AGHQ performs well in terms of both bias and coverage rate. QMC also performs well but the number of responses per participant must be sufficiently large. In contrast, LA fails quite often due to undefined standard errors. We also suggest ML-based methods to test the goodness of fit and to compare models taking model complexity into account. The article closes with an illustrative empirical application and an outlook on possible extensions and future applications of the proposed ML approach.

Gradient Tree Boosting for Hierarchical Data.

Gradient tree boosting is a powerful machine learning technique that has shown good performance in predicting a variety of outcomes. However, when applied to hierarchical (e.g., longitudinal or clustered) data, the predictive performance of gradient tree boosting may be harmed by ignoring the hierarchical structure, and may be improved by accounting for it. Tree-based methods such as regression trees and random forests have already been extended to hierarchical data settings by combining them with the linear mixed effects model (MEM). In the present article, we add to this literature by proposing two algorithms to estimate a combination of the MEM and gradient tree boosting. We report on two simulation studies that (i) investigate the predictive performance of the two MEM boosting algorithms and (ii) compare them to standard gradient tree boosting, standard random forest, and other existing methods for hierarchical data (MEM, MEM random forests, model-based boosting, Bayesian additive regression trees [BART]). We found substantial improvements in the predictive performance of our MEM boosting algorithms over standard boosting when the random effects were non-negligible. MEM boosting as well as BART showed a predictive performance similar to the correctly specified MEM (i.e., the benchmark model), and overall outperformed the model-based boosting and random forest approaches.

Do Agency and Communion Explain the Relationship Between Perceiver and Target Effects in Interpersonal Perception? A Meta-Analysis on Generalized Reciprocity.

This meta-analysis examines generalized reciprocity, that is, the relationship between how people perceive others and how they are perceived by others. It tests the hypothesis that generalized reciprocity varies as a function of the content domain under investigation. Generalized reciprocity for attributes with primarily communal content (e.g., friendliness) was hypothesized to be more positive than generalized reciprocity for attributes with primarily agentic content (e.g., assertiveness). Sixty-four primary studies reporting correlations between perceiver and target effects with a total number of 17,561 participants were included in the analysis. Results of a multilevel meta-analytical random effects model showed that reciprocity correlations were slightly negative, but around zero, for primarily agentic attributes (r = -.05) and became more positive with increasing communal content (up to r = .18 for primarily communal attributes). Generalized reciprocity thus varied depending on the extent to which the regarded attribute is agentic versus communal.

What determines hindsight bias in written work? One field and three experimental studies in the context of Wikipedia.

Hindsight bias not only occurs in individual perception but in written work (e.g., Wikipedia articles) as well. To avoid the possibility that biased written representations of events distort the views of broad audiences, one needs to understand the factors that determine hindsight bias in written work. Therefore, we tested the effect of three potential determinants: the extent to which an event evokes sense-making motivation, the availability of verifiable causal information regarding the event, and the provision of content policies. We conducted one field study examining real Wikipedia articles (N = 40) and three preregistered experimental studies in which participants wrote or edited articles based on different materials (total N = 720). In each experiment, we systematically varied one determinant. Findings provide further-and even more general-support that Wikipedia articles about various events contain hindsight bias. The magnitude of hindsight bias in written work was contingent on the sense-making motivation and the availability of causal information. We did not find support for the effect of content policies. Findings are in line with causal model theory and suggest that some types and topics of written work might be particularly biased by hindsight (e.g., coverage of disasters, research reports, written expert opinions). (PsycInfo Database Record (c) 2023 APA, all rights reserved).

The true role that suppressor effects play in condition-based regression analysis: None. A reply to Fiedler (2021).

Condition-based regression analysis (CRA) is a statistical method for testing self-enhancement effects. That is, CRA indicates whether, in a set of empirical data, people with higher values on the directed discrepancy self-view S minus reality criterion R (i.e., S-R) tend to have higher values on some outcome variable (e.g., happiness). In a critical comment, Fiedler (2021) claims that CRA yields inaccurate conclusions in data with a suppressor effect. Here, we show that Fiedler's critique is unwarranted. All data that are simulated in his comment show a positive association between S-R and H, which is accurately detected by CRA. By construction, CRA indicates an association between S-R and H only when it is present in the data. In contrast to Fiedler's claim, it also yields valid conclusions when the outcome variable is related only to the self-view or when there is a suppressor effect. Our clarifications provide guidance for evaluating Fiedler's comment, clear up with the common heuristic that suppressor effects are always problematic, and assist readers in fully understanding CRA. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Long-term Changes in Depressive Symptoms Before and After Stroke.

OBJECTIVES: To determine the trajectory of depressive symptoms several years before and after incident stroke. METHOD: We analysed data from 10,797 participants from the English Longitudinal Study of Ageing (ELSA) without a history of stroke at baseline (wave 1). We matched participants with incident stroke during the 12-year follow-up (waves 2-7) to stroke-free individuals using propensity scores accounting for age, gender, education, ethnicity, and vascular risk factors. Trajectories of depressive symptoms before and after stroke were analysed using multilevel models. RESULTS: Among the 10,797 participants (mean age 64.6 ± 9.9 years, 54.8 % women), we identified 425 individuals with incident stroke. At the assessment before stroke, these individuals demonstrated an increase in depressive symptoms compared to matched controls. There was a further increase in depressive symptoms in stroke survivors after the acute event, which persisted for several years. Symptom-level analyses revealed that differences in depressive symptoms between stroke survivors and stroke-free controls before and after stroke were most pronounced for mood- and fatigue-related symptoms. DISCUSSION: Incident stroke is associated with long-term increases in depressive symptoms. A small part of this increase occurs in the years before stroke, perhaps indicating the incipient pathological process. Particular attention should be paid to depressive symptoms in the long-term care of patients, and especially to fatigue-related symptoms.

The incremental validity of average states: A replication and extension of Finnigan and Vazire (2018).

States refer to our momentary thoughts, feelings, and behaviors. Average states (aggregates across multiple time points) are discussed as a more accurate and objective measure of personality compared to global self-reports since they do not only rely on people's general beliefs about themselves. Specifically, Finnigan and Vazire (2018) argued that, if average states better capture what a person is actually like, this should be reflected in their unique association with informant-reports of personality, and tested this idea based on two experience-sampling studies. Their results showed, however, that average self-reported states did not predict global informant-reported personality above and beyond global self-reports. In this research, we aimed at replicating and extending these results. We used data of five studies (total N = 806) that involved global self- and informant-reports and employed a variety of different experience-sampling methods (time-based with different sampling schedules, event-based). Across all studies, the original results (i.e., no incremental effects of average self-reported states) were replicated. Furthermore, as an extension to the original study, we found that average other-reported states (provided by peers, results based on one study) did indeed predict global informant-reports above and beyond global self-reports. These findings highlight the importance of differentiating between method effects (global reports vs. average states) from source of information effects (self vs. other). We discuss these results, focusing on the suitability of using informant-reports as a criterion variable and conceptual differences between assessment methods. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Positivity in peer perceptions over time: Personality explains variation at zero-acquaintance, popularity explains differential change.

People have characteristic ways of perceiving others' personalities. When judging others on several traits, some perceivers tend to form globally positive and others tend to form globally negative impressions. These differences, often termed perceiver effects, have mostly been conceptualized as a static construct that taps perceivers' personal stereotypes about the average other. Here, we assessed perceiver effects repeatedly in small groups of strangers who got to know each other over the course of 2-3 weeks and examined the degree to which positivity differences were stable versus developed systematically over time. Using second-order latent growth curve modeling, we tested whether initial positivity (i.e., random intercepts) could be explained by several personality variables and whether change (i.e., random slopes) could be explained by these personality variables and by perceivers' social experiences within the group. Across three studies (ns = 439, 257, and 311), personality variables characterized by specific beliefs about others, such as agreeableness and narcissistic rivalry, were found to explain initial positivity but personality was not reliably linked to changes in positivity over time. Instead, feeling liked and, to a lesser extent, being liked by one's peers partially explained changes in positivity. The results suggest that perceiver effects are best conceptualized as reflecting personal generalized stereotypes at an initial encounter but group-specific stereotypes that are fueled by social experiences as groups get acquainted. More generally, these findings suggest that perceiver effects might be a key variable to understanding reciprocal dynamics of small groups and interpersonal functioning. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

An extension of the mixed-effects growth model that considers between-person differences in the within-subject variance and the autocorrelation.

Experience sampling methods have led to a significant increase in the availability of intensive longitudinal data. Typically, this type of data is analyzed with a mixed-effects model that allows to examine hypotheses concerning between-person differences in the mean structure by including multiple random effects per individual (eg, random intercept and random slopes). Here, we describe an extension of this model that-in addition to the random effects for the mean structure-also includes a random effect for the within-subject variance and a random effect for the autocorrelation. After the description of the model, we show how its parameters can be efficiently estimated using a marginal maximum likelihood (ML) approach. We then illustrate the model using a real data example. We also present the results of a small simulation study in which we compare the ML approach with a Bayesian estimation approach.