Kenneth Koslowski
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Kenneth Koslowski has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
- Jana Holtmann
Wilhelm Wundt Institute for Psychology, Leipzig University, Leipzig, Germany
5 shared publications - Tobias Koch
Friedrich-Schiller-University Jena
5 shared publications Jana Holtmann
4 shared publications- Fabian Felix Münch
Department of Psychology, Psychological Methods Division, Friedrich-Schiller-Universität Jena, Am Steiger 1, Haus 3, 07743, Jena, Germany.
3 shared publications - Veit Kubik
Department of Psychology, Bielefeld University, Bielefeld, Germany
3 shared publications - Alp Aslan
Department of Psychology, Martin Luther University Halle-Wittenberg, 06108 Halle (Saale), Germany
1 shared publication
Latest Publications
- Open Access
- 19 August 2026
- Published by Springer Nature in Communications Psychology
- Vol. 4 (1) , 121
Improving the comparability of momentary affect assessments and accounting for item-specific effects in within-person affect dynamics
A primary objective of intensive longitudinal studies is to investigate within-person dynamics. In this context, item heterogeneity plays a critical role, as within-person processes may vary across items within a scale. A common example is the assessment of momentary affect using adjective lists (e.g., sad, angry, anxious, stressed), where each item captures different facets of positive or negative affect, providing unique and non-interchangeable information. However, standard practices often overlook item heterogeneity by aggregating item scores or assuming a single within-person factor in dynamic structural equation models. This simplification does not permit a fine-grained analysis of within-person dynamics and compromises cross-study comparability when item pools differ across studies. In this article, we reanalyze five large-scale intensive longitudinal datasets assessing momentary affect to illustrate how item heterogeneity can be explicitly modeled. We introduce a flexible modeling approach that accommodates item-specific and person-specific dynamics while improving psychometric comparability across studies, based on residual dynamic structural equation modeling with reference items. We compare this method to conventional modeling strategies and provide practical guidance for addressing item heterogeneity in the analysis of intensive longitudinal data.
- 19 August 2026
- Published by Springer Nature in Communications Psychology
- Vol. 4 (1) , 121
Open AccessImproving the comparability of momentary affect assessments and accounting for item-specific effects in within-person affect dynamics
A primary objective of intensive longitudinal studies is to investigate within-person dynamics. In this context, item heterogeneity plays a critical role, as within-person processes may vary across items within a scale. A common example is the assessment of momentary affect using adjective lists (e.g., sad, angry, anxious, stressed), where each item captures different facets of positive or negative affect, providing unique and non-interchangeable information. However, standard practices often overlook item heterogeneity by aggregating item scores or assuming a single within-person factor in dynamic structural equation models. This simplification does not permit a fine-grained analysis of within-person dynamics and compromises cross-study comparability when item pools differ across studies. In this article, we reanalyze five large-scale intensive longitudinal datasets assessing momentary affect to illustrate how item heterogeneity can be explicitly modeled. We introduce a flexible modeling approach that accommodates item-specific and person-specific dynamics while improving psychometric comparability across studies, based on residual dynamic structural equation modeling with reference items. We compare this method to conventional modeling strategies and provide practical guidance for addressing item heterogeneity in the analysis of intensive longitudinal data.
- Open Access
- 8 August 2026
- Published by Center for Open Science in PsyArXiv
Improving the Comparability of Momentary Affect Assessments: The Role of Item-Specific Effects in Within-Person Affect Dynamics
A primary objective of intensive longitudinal studies is to investigate within-person dynamics. In this context, item heterogeneity plays a critical role, aswithin-person processes may vary across items within a scale. A common exam-ple is the assessment of momentary affect using adjective lists (e.g., sad, angry,anxious, stressed), where each item captures different facets of positive or neg-ative affect, providing unique and non-interchangeable information. However,standard practices often overlook item heterogeneity by aggregating item scores or assuming a single within-person factor in dynamic structural equation mod-els. This simplification does not permit a fine-grained analysis of within-persondynamics and compromises cross-study comparability when item pools differacross studies. In this article, we reanalyze five large-scale intensive longitudinaldatasets assessing momentary affect to illustrate how item heterogeneity can beexplicitly modeled. We introduce a flexible modeling approach that accommo-dates item-specific and person-specific dynamics while improving comparabilityacross studies, based on residualized dynamic structural equation modeling withreference items. We compare this method to conventional modeling strategiesand provide practical guidance for addressing item heterogeneity in the analysisof intensive longitudinal data.
- 8 August 2026
- Published by Center for Open Science in PsyArXiv
Open AccessImproving the Comparability of Momentary Affect Assessments: The Role of Item-Specific Effects in Within-Person Affect Dynamics
A primary objective of intensive longitudinal studies is to investigate within-person dynamics. In this context, item heterogeneity plays a critical role, aswithin-person processes may vary across items within a scale. A common exam-ple is the assessment of momentary affect using adjective lists (e.g., sad, angry,anxious, stressed), where each item captures different facets of positive or neg-ative affect, providing unique and non-interchangeable information. However,standard practices often overlook item heterogeneity by aggregating item scores or assuming a single within-person factor in dynamic structural equation mod-els. This simplification does not permit a fine-grained analysis of within-persondynamics and compromises cross-study comparability when item pools differacross studies. In this article, we reanalyze five large-scale intensive longitudinaldatasets assessing momentary affect to illustrate how item heterogeneity can beexplicitly modeled. We introduce a flexible modeling approach that accommo-dates item-specific and person-specific dynamics while improving comparabilityacross studies, based on residualized dynamic structural equation modeling withreference items. We compare this method to conventional modeling strategiesand provide practical guidance for addressing item heterogeneity in the analysisof intensive longitudinal data.
- Open Access
- 15 July 2026
- Published by Center for Open Science in PsyArXiv
Beneath the Floor: Censored DSEM Models for Analyzing ESM Data with Floor Effects
The distribution of responses to negative affect items collected via experience sampling methods (ESM) often exhibit floor effects. When analyzed via multilevel models, such floor effects may induce dependencies between individual means and variances. Measurement-related censoring processes offer one explanation for the occurrence of floor effects. In this study, we investigated the consequences of ignoring floor effects in the estimation of Dynamic Structural Equation Models (DSEM) with censored variables. We examined three common modeling scenarios: (1) a two-level AR(1) model with an external criterion, (2) a two-level bivariate VAR(1) model, and (3) a three-indicator AR(1) factor model. Using simulations as well as an empirical application, standard DSEM models were compared to censored DSEM variants, as implemented in the `mlts` R-package. Ignoring censoring led to substantial bias in random-effect correlations, particularly between individual mean levels and innovation variances. These correlations were consistently overestimated with low coverage and inflated Type I error rates, even when floor effects were as small as 5%. In scenario (3), the presence of a single censored indicator was sufficient to bias random-effect correlations when censoring was ignored. When random effects are used jointly to predict time-invariant outcomes, this bias introduces (spurious) redundancy between the random-effect predictors. Notably, censoring also adversely affected standard AR(1) models fitted to composite scores (i.e., averages across indicators). Overall, our findings highlight the importance of explicitly modeling floor effects in DSEM. We strongly recommend selecting modeling approaches that align with substantive assumptions about the origin of floor effects and briefly discuss alternative strategies.
- 15 July 2026
- Published by Center for Open Science in PsyArXiv
Open AccessBeneath the Floor: Censored DSEM Models for Analyzing ESM Data with Floor Effects
The distribution of responses to negative affect items collected via experience sampling methods (ESM) often exhibit floor effects. When analyzed via multilevel models, such floor effects may induce dependencies between individual means and variances. Measurement-related censoring processes offer one explanation for the occurrence of floor effects. In this study, we investigated the consequences of ignoring floor effects in the estimation of Dynamic Structural Equation Models (DSEM) with censored variables. We examined three common modeling scenarios: (1) a two-level AR(1) model with an external criterion, (2) a two-level bivariate VAR(1) model, and (3) a three-indicator AR(1) factor model. Using simulations as well as an empirical application, standard DSEM models were compared to censored DSEM variants, as implemented in the `mlts` R-package. Ignoring censoring led to substantial bias in random-effect correlations, particularly between individual mean levels and innovation variances. These correlations were consistently overestimated with low coverage and inflated Type I error rates, even when floor effects were as small as 5%. In scenario (3), the presence of a single censored indicator was sufficient to bias random-effect correlations when censoring was ignored. When random effects are used jointly to predict time-invariant outcomes, this bias introduces (spurious) redundancy between the random-effect predictors. Notably, censoring also adversely affected standard AR(1) models fitted to composite scores (i.e., averages across indicators). Overall, our findings highlight the importance of explicitly modeling floor effects in DSEM. We strongly recommend selecting modeling approaches that align with substantive assumptions about the origin of floor effects and briefly discuss alternative strategies.
- Open Access
- 12 May 2026
- Published by Center for Open Science in PsyArXiv
Attitudes Toward Open Science Practices Among German Psychologists
The wide adoption of Open Science (OS) practices is influencing psychological science positively by increasing openness, transparency, reproducibility, and replicability. Researchers are encouraged to openly share their data, materials, and reproducible code, and to pre-register their hypotheses, design, sample size, and analysis plans whenever possible. On the one hand, these OS practices incur objective and/or perceived benefits, thereby raising hopes among researchers. On the other hand, these practices may also incur objective and/or perceived costs, thereby potentially inducing worries, especially among early career researchers. To investigate researchers’ attitudes toward the application of OS practices in their research and teaching, we conducted a survey among members of the German Psychological Society (N = 596) in 2021. Across status groups, researchers agreed that psychological science generally profits from the application of OS practices and that OS practices will become commonplace in the future. Furthermore, researchers indicated increased trust in studies adopting OS practices. However, researchers of all status groups reported to mainly have experience with pre-registration but less or even no experience with other OS practices. Researchers, especially those early in their career, believed to profit from OS, but also reported worries of making errors, especially when it comes to openly sharing analysis scripts. We discuss the current state of OS implementation and propose interventions to facilitate the broader adoption of OS practices, thereby supporting the career progression of early career researchers.
- 12 May 2026
- Published by Center for Open Science in PsyArXiv
Open AccessAttitudes Toward Open Science Practices Among German Psychologists
The wide adoption of Open Science (OS) practices is influencing psychological science positively by increasing openness, transparency, reproducibility, and replicability. Researchers are encouraged to openly share their data, materials, and reproducible code, and to pre-register their hypotheses, design, sample size, and analysis plans whenever possible. On the one hand, these OS practices incur objective and/or perceived benefits, thereby raising hopes among researchers. On the other hand, these practices may also incur objective and/or perceived costs, thereby potentially inducing worries, especially among early career researchers. To investigate researchers’ attitudes toward the application of OS practices in their research and teaching, we conducted a survey among members of the German Psychological Society (N = 596) in 2021. Across status groups, researchers agreed that psychological science generally profits from the application of OS practices and that OS practices will become commonplace in the future. Furthermore, researchers indicated increased trust in studies adopting OS practices. However, researchers of all status groups reported to mainly have experience with pre-registration but less or even no experience with other OS practices. Researchers, especially those early in their career, believed to profit from OS, but also reported worries of making errors, especially when it comes to openly sharing analysis scripts. We discuss the current state of OS implementation and propose interventions to facilitate the broader adoption of OS practices, thereby supporting the career progression of early career researchers.
- Open Access
- 9 December 2025
- Published by Center for Open Science in PsyArXiv
A Tutorial on Bayesian Multilevel Latent Time Series Models using Stan with the mlts R-package
This tutorial introduces and illustrates the estimation of Bayesian multilevel latent time series models with the R package `mlts`. We provide a conceptual overview of Dynamic Structural Equation Modeling (DSEM) for intensive longitudinal data, highlighting how it combines time series analysis, multilevel modeling, and latent variable approaches. Step-by-step guidance is given on specifying, estimating, and interpreting two-level vector autoregressive models using `mlts`. We illustrate modeling extensions provided in `mlts`, including multiple-indicator measurement models, between-person covariates, latent interaction effects, the handling of censored variables, and multiple-group models. Practical considerations such as data preparation, model diagnostics, and handling of varying measurement intervals are discussed. Assessment of model fit is provided via posterior predictive checks. By use of working examples with empricial data, we showcase how `mlts` enables flexible Bayesian DSEM analyses while avoiding the complexities of custom Stan programming. This tutorial aims to make advanced dynamic modeling of intensive longitudinal data more accessible to applied researchers across the behavioral sciences.
- 9 December 2025
- Published by Center for Open Science in PsyArXiv
Open AccessA Tutorial on Bayesian Multilevel Latent Time Series Models using Stan with the mlts R-package
This tutorial introduces and illustrates the estimation of Bayesian multilevel latent time series models with the R package `mlts`. We provide a conceptual overview of Dynamic Structural Equation Modeling (DSEM) for intensive longitudinal data, highlighting how it combines time series analysis, multilevel modeling, and latent variable approaches. Step-by-step guidance is given on specifying, estimating, and interpreting two-level vector autoregressive models using `mlts`. We illustrate modeling extensions provided in `mlts`, including multiple-indicator measurement models, between-person covariates, latent interaction effects, the handling of censored variables, and multiple-group models. Practical considerations such as data preparation, model diagnostics, and handling of varying measurement intervals are discussed. Assessment of model fit is provided via posterior predictive checks. By use of working examples with empricial data, we showcase how `mlts` enables flexible Bayesian DSEM analyses while avoiding the complexities of custom Stan programming. This tutorial aims to make advanced dynamic modeling of intensive longitudinal data more accessible to applied researchers across the behavioral sciences.
- Open Access
- 11 November 2025
- Published by Center for Open Science in PsyArXiv
RETRIEVAL-POTENTIATED NEW LEARNING IN DIGITAL HIGHER EDUCATION
Background: Retrieving from memory directly benefits long-term retention and also enhances subsequent learning of new content, known as retrieval-potentiated new learning or the forward testing effect. Aims: The present study assessed the forward testing effect with complex learning materials and was the first to directly test motivational mechanisms, using cognitive effort, enjoyment, and self-paced practice time as motivational indicators.Sample: 308 undergraduates in an online psychology course participated.Methods: Students read three texts introducing new course concepts relevant to their final course exam. After studying Texts 1 and 2, students recalled or restudied the previous text for a self-paced or fixed amount of time. All students took a free-recall test after studying Text 3. Results: Recall of Text 3 was greater for retrieval practice than restudy of Texts 1 and 2. This forward testing effect was independent of learners’ prior knowledge and text interest, as well as whether practice times were self-regulated. Consistent with the encoding effort hypothesis, cognitive effort and practice times mediated the forward testing effect: Compared to restudying, retrieval practice reduced the declines in effort and practice time observed across blocks, thereby enhancing learning of Text 3. Enjoyment did not decline across blocks but there was no evidence it mediated the forward testing effect. Conclusions: The present research affirms the potential instructional value of retrieval practice for promoting new learning, including in online education. There was some support for motivational accounts of the forward testing effect but motivation likely cannot fully explain why retrieval practice facilitates new learning.
- 11 November 2025
- Published by Center for Open Science in PsyArXiv
Open AccessRETRIEVAL-POTENTIATED NEW LEARNING IN DIGITAL HIGHER EDUCATION
Background: Retrieving from memory directly benefits long-term retention and also enhances subsequent learning of new content, known as retrieval-potentiated new learning or the forward testing effect. Aims: The present study assessed the forward testing effect with complex learning materials and was the first to directly test motivational mechanisms, using cognitive effort, enjoyment, and self-paced practice time as motivational indicators.Sample: 308 undergraduates in an online psychology course participated.Methods: Students read three texts introducing new course concepts relevant to their final course exam. After studying Texts 1 and 2, students recalled or restudied the previous text for a self-paced or fixed amount of time. All students took a free-recall test after studying Text 3. Results: Recall of Text 3 was greater for retrieval practice than restudy of Texts 1 and 2. This forward testing effect was independent of learners’ prior knowledge and text interest, as well as whether practice times were self-regulated. Consistent with the encoding effort hypothesis, cognitive effort and practice times mediated the forward testing effect: Compared to restudying, retrieval practice reduced the declines in effort and practice time observed across blocks, thereby enhancing learning of Text 3. Enjoyment did not decline across blocks but there was no evidence it mediated the forward testing effect. Conclusions: The present research affirms the potential instructional value of retrieval practice for promoting new learning, including in online education. There was some support for motivational accounts of the forward testing effect but motivation likely cannot fully explain why retrieval practice facilitates new learning.
- Open Access
- 5 September 2025
- Published by Springer Nature in Behavior Research Methods
- Vol. 57 (10) , 1-24
Modeling within-level latent interaction effects in multilevel vector-autoregressive models
The study of time-dependent within-person dynamics has gained popularity in recent years through the use of multilevel (latent) time-series models. However, due to the complexity of the models, model applications are usually limited with respect to the inclusion of time-varying moderating factors on the longitudinal within-person relations between variables. That is, in common applications of multilevel time-series models, the within-person dynamics of constructs over time are regarded as being insensitive to changes in other time-varying factors or changes in contexts. We illustrate an extension of multilevel latent time-series models that incorporate latent interaction effects at the dynamic within-person level. We build on previous work that incorporated time-varying observed or latent moderator variables for the dynamic parameters in vector autoregressive models and provide a tutorial for the application of the models, implemented and estimated using Bayesian estimation via Markov chain Monte Carlo techniques. The models are illustrated by two empirical applications that investigate the temporal dynamics of negative affect, rumination, and mindful attention. The performance of different models with varying complexity is investigated via several simulation studies to provide recommendations regarding the models’ applicability for applied researchers. Required sample sizes range between at least 25 time points for around 50 persons in the simplest fixed-effects models and at least 100 time points for at least 100 persons in random-effects factor models, depending on the expected effect sizes of the dynamic parameters.
- 5 September 2025
- Published by Springer Nature in Behavior Research Methods
- Vol. 57 (10) , 1-24
Open AccessModeling within-level latent interaction effects in multilevel vector-autoregressive models
The study of time-dependent within-person dynamics has gained popularity in recent years through the use of multilevel (latent) time-series models. However, due to the complexity of the models, model applications are usually limited with respect to the inclusion of time-varying moderating factors on the longitudinal within-person relations between variables. That is, in common applications of multilevel time-series models, the within-person dynamics of constructs over time are regarded as being insensitive to changes in other time-varying factors or changes in contexts. We illustrate an extension of multilevel latent time-series models that incorporate latent interaction effects at the dynamic within-person level. We build on previous work that incorporated time-varying observed or latent moderator variables for the dynamic parameters in vector autoregressive models and provide a tutorial for the application of the models, implemented and estimated using Bayesian estimation via Markov chain Monte Carlo techniques. The models are illustrated by two empirical applications that investigate the temporal dynamics of negative affect, rumination, and mindful attention. The performance of different models with varying complexity is investigated via several simulation studies to provide recommendations regarding the models’ applicability for applied researchers. Required sample sizes range between at least 25 time points for around 50 persons in the simplest fixed-effects models and at least 100 time points for at least 100 persons in random-effects factor models, depending on the expected effect sizes of the dynamic parameters.
- Open Access
- 2 September 2025
- Published by Taylor & Francis in Multivariate Behavioral Research
- Vol. 60 (6) , 1199-1220
Unique Contributions of Dynamic Affect Indicators – Beyond Static Variability
Indicators of affect dynamics (IADs) capture temporal dependencies and instability in affective trajectories over time. However, the relevance of IADs for the prediction of time-invariant outcomes (e.g., depressive symptoms) was recently challenged due to results suggesting low predictive utility beyond intraindividual means and variances. We argue that these results may in part be explained by mathematical redundancies between IADs and static variability as well as the chosen modeling strategy. In three extensive simulation studies we investigate the accuracy and power for detecting non-null relations between IADs and an outcome variable in different relevant settings, illustrating the effect of the length of a time series, the presence of missing values or measurement error, as well as of erroneously fixing innovation variances to be equal across persons. We show that, if uncertainty in individual IAD estimates is not accounted for, relations between IADs (i.e., autoregressive effects) and a time-invariant outcome are underestimated even in large samples and propose the use of a latent multilevel one-step approach. In an empirical application we illustrate that the different modeling approaches can lead to different substantive conclusions regarding the role of negative affect inertia in the prediction of depressive symptoms.
- 2 September 2025
- Published by Taylor & Francis in Multivariate Behavioral Research
- Vol. 60 (6) , 1199-1220
Open AccessUnique Contributions of Dynamic Affect Indicators – Beyond Static Variability
Indicators of affect dynamics (IADs) capture temporal dependencies and instability in affective trajectories over time. However, the relevance of IADs for the prediction of time-invariant outcomes (e.g., depressive symptoms) was recently challenged due to results suggesting low predictive utility beyond intraindividual means and variances. We argue that these results may in part be explained by mathematical redundancies between IADs and static variability as well as the chosen modeling strategy. In three extensive simulation studies we investigate the accuracy and power for detecting non-null relations between IADs and an outcome variable in different relevant settings, illustrating the effect of the length of a time series, the presence of missing values or measurement error, as well as of erroneously fixing innovation variances to be equal across persons. We show that, if uncertainty in individual IAD estimates is not accounted for, relations between IADs (i.e., autoregressive effects) and a time-invariant outcome are underestimated even in large samples and propose the use of a latent multilevel one-step approach. In an empirical application we illustrate that the different modeling approaches can lead to different substantive conclusions regarding the role of negative affect inertia in the prediction of depressive symptoms.
- Open Access
- 5 August 2025
- Published by Center for Open Science in PsyArXiv
Unique Contributions of Dynamic Affect Indicators – Beyond Static Variability
Indicators of affect dynamics (IADs) capture temporal dependencies and instability in affective trajectories over time. However, the relevance of IADs for the prediction of time?invariant outcomes (e.g., depressive symptoms) was recently challenged due to results suggesting low predictive utility beyond intraindividual means and variances. We argue that these results may in part be explained by mathematical redundancies between IADs and static variability. In two extensive simulation studies we investigate the accuracy and power for detecting non-null relations between IADs and an outcome variable in different relevant settings, illustrating the effect of the length of a time series as well as of the presence of missing values or measurement error. We show that, if uncertainty in individual IAD estimates is not accounted for, relations between IADs (i.e., autoregressive effects) and a time-invariant outcome are underestimated even in large samples and propose the use of a latent multilevel one-step approach. In an empirical application we illustrate that the different modeling approaches can lead to different substantive conclusions regarding the role of negative affect inertia in the prediction of depressive symptoms.
- 5 August 2025
- Published by Center for Open Science in PsyArXiv
Open AccessUnique Contributions of Dynamic Affect Indicators – Beyond Static Variability
Indicators of affect dynamics (IADs) capture temporal dependencies and instability in affective trajectories over time. However, the relevance of IADs for the prediction of time?invariant outcomes (e.g., depressive symptoms) was recently challenged due to results suggesting low predictive utility beyond intraindividual means and variances. We argue that these results may in part be explained by mathematical redundancies between IADs and static variability. In two extensive simulation studies we investigate the accuracy and power for detecting non-null relations between IADs and an outcome variable in different relevant settings, illustrating the effect of the length of a time series as well as of the presence of missing values or measurement error. We show that, if uncertainty in individual IAD estimates is not accounted for, relations between IADs (i.e., autoregressive effects) and a time-invariant outcome are underestimated even in large samples and propose the use of a latent multilevel one-step approach. In an empirical application we illustrate that the different modeling approaches can lead to different substantive conclusions regarding the role of negative affect inertia in the prediction of depressive symptoms.