Search

BackgroundIn the face of the COVID-19 pandemic, the German government and the 16 German federal states implemented a variety of nonpharmaceutical interventions (NPIs) to decelerate the spread of the SARS-CoV-2 virus and thus prevent a collapse of the health care system. These measures comprised, among others, social distancing, the temporary closure of shops and schools, and a ban of large public gatherings and meetings with people not living in the same household. ObjectiveIt is fair to assume that the issued NPIs have heavily affected social life and psychological functioning. We therefore aimed to examine possible effects of this lockdown in conjunction with daily new infections and the state of the national economy on people’s interests, motives, and other psychological states. MethodsWe derived 249 keywords from the Google Trends database, tapping into 27 empirically and rationally selected psychological domains. To overcome issues with reliability and specificity of individual indicator variables, broad factors were derived by means of time series factor analysis. All domains were subjected to a change point analysis and time series regression analysis with infection rates, NPIs, and the state of the economy as predictors. All keywords and analyses were preregistered prior to analysis. ResultsWith the pandemic arriving in Germany, significant increases in people’s search interests were observed in virtually all domains. Although most of the changes were short-lasting, each had a distinguishable onset during the lockdown period. Regression analysis of the Google Trends data confirmed pronounced autoregressive effects for the investigated variables, while forecasting by means of the tested predictors (ie, daily new infections, NPIs, and the state of economy) was moderate at best. ConclusionsOur findings indicate that people’s interests, motives, and psychological states are heavily affected in times of crisis and lockdown. Specifically, disease- and virus-related domains (eg, pandemic disease, symptoms) peaked early, whereas personal health strategies (eg, masks, homeschooling) peaked later during the lockdown. Domains addressing social life and psychosocial functioning showed long-term increases in public interest. Renovation was the only domain to show a decrease in search interest with the onset of the lockdown. As changes in search behavior are consistent over multiple domains, a Google Trends analysis may provide information for policy makers on how to adapt and develop intervention, information, and prevention strategies, especially when NPIs are in effect.

Traditionally, researchers employ human raters for scoring responses to creative thinking tasks. Apart from the associated costs this approach entails two potential risks. First, human raters can be subjective in their scoring behavior (inter-rater-variance). Second, individual raters are prone to inconsistent scoring patterns (intra-rater-variance). In light of these issues, we present an approach for automated scoring of Divergent Thinking (DT) Tasks. We implemented a pipeline aiming to generate accurate rating predictions for DT responses using text mining and machine learning methods. Based on two existing data sets from two different laboratories, we constructed several prediction models incorporating features representing meta information of the response or features engineered from the response's word embeddings that were obtained using pre-trained GloVe and Word2Vec word vector spaces. Out of these features, word embeddings and features derived from them proved to be particularly effective. Overall, longer responses tended to achieve higher ratings as well as responses that were semantically distant from the stimulus object. In our comparison of three state-of-the-art machine learning algorithms, Random Forest and XGBoost tended to slightly outperform the Support Vector Regression.

This article is concerned with the large-sample parameter estimator behavior in applications of Bayesian confirmatory factor analysis in behavioral measurement. The property of strong convergence of the popular Bayesian posterior median estimator is discussed, which states numerical convergence with probability 1 of the resulting estimates to the population parameter value as sample size increases without bound. This property is stronger than the consistency and convergence in distribution of that estimator, which have been commonly referred to in the literature. A numerical example is utilized to illustrate this almost sure convergence of a Bayesian latent correlation estimator. The paper contributes to the body of research on optimal statistical features of Bayesian estimates and concludes with a discussion of the implications of this large-sample property of the Bayesian median estimator for empirical measurement studies.

As measures of general language proficiency, C-tests are ubiquitous in language testing. Speeded C-tests are quite recent developments in the field and are deemed to be more discriminatory and provide more accurate diagnostic information than power C-tests especially with high-ability participants. Item response theory modeling of speeded C-tests has not been discussed in the literature, and current approaches for power C-tests based on ordinal models either violate the model assumptions or are relatively complex to be reliably fitted with small samples. Count data models are viable alternatives with less restrictive assumptions and lower complexity. In the current study, we compare count data models with commonly applied ordinal models for modeling a speeded C-test. It was found that a flexible count data model fits equally well in absolute and relative terms as compared with ordinal models. Implications and feasibility of count data models for the psychometric modeling of C-tests are discussed.

Instructional effects in creative-thinking tasks are important to understand in order to promote creative performance of individuals. In divergent-thinking tasks, for example, instructional and strategic enhancement effects have been extensively studied for verbal tasks. However, while studies on instructional enhancement effects on creative drawing tasks exist, it is surprising that strategy enhancement in figural divergent thinking is still underresearched. In this study, we used a strategy manipulation approach to reassess the role of executive strategy implementation and the moderating role of an indicator of fluid intelligence, figural analogical reasoning, in two types of figural divergent-thinking tasks (abstract vs. concrete). The sample comprised N = 75 high-school students. Importantly, we found strategic enhancement effects by combining strategy instructions with a prompt to "be creative." This combined instruction was contrasted with a standard instruction, and main effects were found for overall, concrete, and abstract creativity. Moreover, we found in a regression analysis a main effect for figural analogical reasoning on overall creativity and creativity for only the concrete object tasks. An expected interaction effect of instruction and figural analogical reasoning was not found. As another addition, the role of current motivation in figural divergent thinking was explored.

Regression models are commonly applied in the analysis of transportation data. This research aims at broadening the range of methods used for this task by modeling the spatial distribution of bike‐sharing trips in Cologne, Germany, applying both parametric regression models and a modified machine learning approach while incorporating measures to account for spatial autocorrelation. Independent variables included in the models consist of land use types, elements of the transport system and sociodemographic characteristics. Out of several regression models with different underlying distributions, a Tweedie generalized additive model is chosen by its values for AIC, RMSE, and sMAPE to be compared to an XGBoost model. To consider spatial relationships, spatial splines are included in the Tweedie model, while the estimations of the XGBoost model are modified using a geographically weighted regression. Both methods entail certain advantages: while XGBoost leads to far better values regarding RMSE and sMAPE and therefore to a better model fit, the Tweedie model allows an easier interpretation of the influence of the independent variables including spatial effects. [ABSTRACT FROM AUTHOR]