Dein Suchergebnis zum Thema: Olive

Sequence mining techniques provide powerful tools to uncover patterns from time-stamped action sequences. The information extracted with a chosen method, however, may not necessarily be meaningful or generalize beyond the specific data and task. In this case study, we examined the generalizability of behavioral patterns identified across four tasks from the PISA 2025 Learning in a Digital World domain. We compared two subgroup discovery approaches: clustering based on atomistic features (e.g., sequence length, time on task) and sequence clustering using pairwise similarities. Both approaches identified consistent behavioral subgroups across tasks, revealed similar transition patterns, exhibited comparable cluster-covariate relationships with task scores, prior knowledge, and effort, and showed strong agreement with each other. These findings suggest that both approaches captured similar and generalizable differences in how examinees approached the tasks. From our results, we derive implications for mining action sequence data, strongly advocating for incorporating theoretical considerations even in exploratory analyses.
Esther Ulitzsch, Leonard Tetzlaff, Frank Goldhammer, Carolin Hahnel, Ulf Kroehne, Oliver

This study investigates longitudinal linking of dichotomous item responses across three time points, focusing on Haberman, Haebara, Stocking-Lord linking, and concurrent calibration under item parameter drift (IPD). Joint and chain linking are examined. Three simulation studies compared these methods under IPD and different linking designs. The first study assessed linking methods under sampling error and uniform IPD, showing Haberman joint linking minimized bias and root mean square error (RMSE), especially at larger sample sizes, while concurrent calibration was most accurate without IPD. The second study introduced two designs: using anchor items across all three time points or only adjacent items between consecutive time points. Without IPD, concurrent calibration remained most precise, while under uniform IPD, Haberman methods balanced bias and precision. The third study analyzed nonuniform IPD, finding Haberman chain linking achieved the best bias and RMSE performance. Overall, chain linking outperformed joint linking under IPD. Implications for analyzing longitudinal linking designs are discussed.
Fachzeitschrift › Forschung › begutachtet Publikationsdaten Von Oskar Engels, Oliver

Previous longitudinal studies on changes in narcissism mostly investigated change over long periods of time in samples consisting of young adults at the first wave. Thus, little is known about how narcissism changes in the short term and in adults of all ages. This study addresses these questions by tracking changes in narcissism over multiple assessments covering a period of approximately 2 years in a sample of adults aged 18 to 80 at the first wave (N = 3,599). Furthermore, we investigated whether narcissism predicted life events and life experiences from different domains and how changes in narcissism were related to these events and experiences. On average, participants decreased slightly on overall narcissism and three narcissism facets over 2 years. Narcissism predicted several life events and experiences, especially from the work (e.g., receiving a promotion) and relationship domains (e.g., breaking up with one’s partner). Several events were related to changes in narcissism; such as starting to date someone new and breaking up with one’s partner. Future confirmatory research testing these associations in a large sample with more frequent measurements is needed.
Fachzeitschrift › Forschung › begutachtet Publikationsdaten Von Eunike Wetzel, Oliver

Previous longitudinal studies on changes in narcissism mostly investigated change over long periods of time in samples consisting of young adults at the first wave. Thus, little is known about how narcissism changes in the short term and in adults of all ages. This study addresses these questions by tracking changes in narcissism over multiple assessments covering a period of approximately 2 years in a sample of adults aged 18 to 80 at the first wave (N = 3,599). Furthermore, we investigated whether narcissism predicted life events and life experiences from different domains and how changes in narcissism were related to these events and experiences. On average, participants decreased slightly on overall narcissism and three narcissism facets over 2 years. Narcissism predicted several life events and experiences, especially from the work (e.g., receiving a promotion) and relationship domains (e.g., breaking up with one’s partner). Several events were related to changes in narcissism; such as starting to date someone new and breaking up with one’s partner. Future confirmatory research testing these associations in a large sample with more frequent measurements is needed.
Fachzeitschrift › Forschung › begutachtet Publikationsdaten Von Eunike Wetzel, Oliver

This study investigates longitudinal linking of dichotomous item responses across three time points, focusing on Haberman, Haebara, Stocking-Lord linking, and concurrent calibration under item parameter drift (IPD). Joint and chain linking are examined. Three simulation studies compared these methods under IPD and different linking designs. The first study assessed linking methods under sampling error and uniform IPD, showing Haberman joint linking minimized bias and root mean square error (RMSE), especially at larger sample sizes, while concurrent calibration was most accurate without IPD. The second study introduced two designs: using anchor items across all three time points or only adjacent items between consecutive time points. Without IPD, concurrent calibration remained most precise, while under uniform IPD, Haberman methods balanced bias and precision. The third study analyzed nonuniform IPD, finding Haberman chain linking achieved the best bias and RMSE performance. Overall, chain linking outperformed joint linking under IPD. Implications for analyzing longitudinal linking designs are discussed.
Fachzeitschrift › Forschung › begutachtet Publikationsdaten Von Oskar Engels, Oliver

Existing estimators of parameters of item response theory (IRT) models exploit the likelihood function. In small samples, however, the IRT likelihood oftentimes contains little informative value, potentially resulting in biased and/or unstable parameter estimates and large standard errors. To facilitate small-sample IRT estimation, we introduce a novel approach that does not rely on the likelihood. Our estimation approach derives features from response data and then maps the features to item parameters using a neural network (NN). We describe and evaluate our approach for the three-parameter logistic model; however, it is applicable to any model with an item characteristic curve. Three types of NNs are developed, supporting the obtainment of both point estimates and confidence intervals for IRT model parameters. The results of a simulation study demonstrate that these NNs perform better than Bayesian estimation using Markov chain Monte Carlo methods in terms of the quality of the point estimates and confidence intervals while also being much faster. These properties facilitate (1) pretesting items in a real-time testing environment, (2) pretesting more items and (3) pretesting items only in a secured environment to eradicate possible compromise of new items in online testing.
Belov, Oliver Lüdtke, Esther Ulitzsch Originalsprache Englisch Erschienen in British

In nonexperimental studies, obtaining an unbiased estimate of the average treatment effect (ATE) typically requires two key assumptions: that all relevant covariates are measured (i.e., no unmeasured confounding) and that the statistical model used for covariate adjustment is correctly specified. Two common approaches for adjustment are specifying an outcome model and propensity score weighting. To mitigate bias from model misspecification, doubly robust methods combine both approaches, ensuring unbiased ATE estimates if either the outcome model or the propensity score model is correctly specified. In this study, we review four doubly robust methods that have received considerable attention in the methodological literature but remain underutilized in psychological research: augmented inverse probability weighting, regression weighted by the inverse propensity score, regression incorporating the inverse propensity score as a covariate, and calibrated propensity score weighting. Using two simulation studies, we compare these methods with regression estimation and inverse probability weighting estimators. Our results suggest that doubly robust methods—particularly regression weighted by the inverse propensity score—offer greater protection against bias from model misspecification across various data-generating scenarios. We also discuss practical considerations for implementing doubly robust methods, including weight normalization, propensity score truncation, and potential efficiency losses due to overfitting. The different methods for estimating the ATE are illustrated in a data example.
Fachzeitschrift › Forschung › begutachtet Publikationsdaten Von Jingyu Zhang, Oliver

Existing estimators of parameters of item response theory (IRT) models exploit the likelihood function. In small samples, however, the IRT likelihood oftentimes contains little informative value, potentially resulting in biased and/or unstable parameter estimates and large standard errors. To facilitate small-sample IRT estimation, we introduce a novel approach that does not rely on the likelihood. Our estimation approach derives features from response data and then maps the features to item parameters using a neural network (NN). We describe and evaluate our approach for the three-parameter logistic model; however, it is applicable to any model with an item characteristic curve. Three types of NNs are developed, supporting the obtainment of both point estimates and confidence intervals for IRT model parameters. The results of a simulation study demonstrate that these NNs perform better than Bayesian estimation using Markov chain Monte Carlo methods in terms of the quality of the point estimates and confidence intervals while also being much faster. These properties facilitate (1) pretesting items in a real-time testing environment, (2) pretesting more items and (3) pretesting items only in a secured environment to eradicate possible compromise of new items in online testing.
Belov, Oliver Lüdtke, Esther Ulitzsch Originalsprache Englisch Erschienen in British

In nonexperimental studies, obtaining an unbiased estimate of the average treatment effect (ATE) typically requires two key assumptions: that all relevant covariates are measured (i.e., no unmeasured confounding) and that the statistical model used for covariate adjustment is correctly specified. Two common approaches for adjustment are specifying an outcome model and propensity score weighting. To mitigate bias from model misspecification, doubly robust methods combine both approaches, ensuring unbiased ATE estimates if either the outcome model or the propensity score model is correctly specified. In this study, we review four doubly robust methods that have received considerable attention in the methodological literature but remain underutilized in psychological research: augmented inverse probability weighting, regression weighted by the inverse propensity score, regression incorporating the inverse propensity score as a covariate, and calibrated propensity score weighting. Using two simulation studies, we compare these methods with regression estimation and inverse probability weighting estimators. Our results suggest that doubly robust methods—particularly regression weighted by the inverse propensity score—offer greater protection against bias from model misspecification across various data-generating scenarios. We also discuss practical considerations for implementing doubly robust methods, including weight normalization, propensity score truncation, and potential efficiency losses due to overfitting. The different methods for estimating the ATE are illustrated in a data example.
Fachzeitschrift › Forschung › begutachtet Publikationsdaten Von Jingyu Zhang, Oliver

Missing data are a common challenge in multilevel designs, and multiple imputation (MI) is often used for handling them. Past research has shown that multilevel MI provides an effective treatment of missing data, so long as the imputation model takes the multilevel structure and the intended analyses into account, and modern methods have been developed that can accommodate even complex types of analyses. However, multilevel MI can be difficult to apply in practice, where the multilevel structure is often not very pronounced or not of immediate interest in the analysis. In these applications, existing methods can become unstable and often struggle to provide reliable results. In this article, we introduce a fully conditional specification (FCS) approach to multilevel MI that combines single-level imputation methods with group means (GM) or adjusted group means (AGM) to accommodate the multilevel structure. Based on theoretical investigations and multiple simulation studies, we evaluated the performance of these methods across balanced and unbalanced designs and with larger numbers of variables. Our findings suggest that the AGM approach – though not the GM approach – performs well across most scenarios we investigated and can even outperform conventional multilevel MI approaches in challenging applications. We also provide an illustrative example of implementing these methods in a simulated setting and discuss the implications of our findings for practice.
Fachzeitschrift › Forschung › begutachtet Publikationsdaten Von Simon Grund, Oliver