31325 research outputs found
Sort by
are there differences in dimensions of emotional intelligence when comparing different sports?
No full textThe article processing charge was funded by the Open Access Publication Fund of Humboldt-Universität zu Berlin.Emotional intelligence (EI) is a determinant of athletic performance. Since each sport has its own psychological requirements, individual EI dimensions could have different relevance for athletic performance. This study investigated EI variations among athletes participating in 48 different sports, considering the specific emotional challenges faced by athletes in competitive sports using the appraisal theory of emotions. A total of 605 athletes (44.3% female) actively engaged in competitive sports completed the Trait Emotional Intelligence Questionnaire via an online survey. Analyses of variance, controlling for age and gender, revealed no overall link between EI and different sports. However, the dimension of self-control within EI demonstrated significant relevance among athletes in contact sports, sports with a higher degree of control over environmental conditions, and sports with direct referee intervention. On the other hand, the dimension of well-being within EI exhibited significant relevance for athletes in individual sports. Thus, this study suggests that investigating appraisal processes in athletes’ emotional experiences, particularly focusing on the self-control dimension of EI, useful for future research in the field of EI.Peer Reviewe
Animacy-based nominal classification in northwestern Bantu and beyond
No full textThis paper builds on Tom Güldemann’s work on Central African nominal classification (Güldemann 2023), investigating the cut-off points of animacy-based agreement in 57 northwestern Bantu languages, stemming from Di Garbo and Verkerk (2022). In the majority of these languages, the gender system revolves around a human-nonhuman contrast, inherited from Proto-Bantu. However, in most languages with animacy-based semantic agreement or restructured gender systems, an animate-inanimate contrast is prevalent. In a few languages, semantic agreement and gender restructuring can revolve around a human-nonhuman contrast. The geographic distribution of these different animacy cut-offs suggests that the animate-inanimate contrast is particularly common in Bantu languages that are geographically close to non-Bantu Central African languages, where this contrast is widespread. Conversely, the human-nonhuman contrast is less common outside Bantu. The study corroborates the centrality of animacy in Central African nominal morphosyntax, where animacy contrasts of different types may have emerged independently multiple times
Machine Learning Applications in Digital Mental Health Interventions
No full textMachine Learning (ML) und Digital Mental Health (DMH) sind vielversprechende Technologien zur Bewältigung der Belastungen durch psychische Erkrankungen – besonders in Kombination. Methodisch rigorose, anwendungsfallspezifische Arbeiten mit großen naturalistischen Datensätzen sind jedoch zentral für die Umsetzung dieses Potentials. Zudem ist unklar, inwiefern ML in DMH bereits genutzt wird. Diese Arbeit vereint daher drei empirische Studien und einen systematischen Review. Zwei Studien untersuchten die Realisierbarkeit von ML-Anwendungen in chatbasierten DMH-Interventionen für junge Menschen. In der ersten Studie wurden ca. 20.000 Chatverläufe analysiert, um den erneuten Kontakt der Chatberatung als Zeichen weiterer Hilfebedürftigkeit vorherzusagen. Es wurde eine AUROC von 0.68 für ungenutzte Testdaten erreicht. In der zweiten Studie wurde eine vergleichbare Vorhersagegüte für die subjektive Beratungsqualität erzielt – basierend auf Daten von 2.500 Hilfesuchenden. Ein transformer-basiertes Modell schnitt dabei nicht besser ab als ein deutlich einfacherer Ansatz. In einer dritten Studie wurden HRV-Daten von ca. 2.600 Patienten in einer DMH-Intervention genutzt. Zwar lag die Vorhersagegüte über dem Zufallsniveau, jedoch klar unter früheren Arbeiten – mit begrenztem praktischem Nutzen. Im Review wurden 94 DMH-Interventionen für depressive Symptome betrachtet. ML wurde darin kaum zur Personalisierung genutzt. Gerade die vielversprechenden Ergebnisse der zwei Chatstudien zeigen Potential für ML in DMH – insbesondere für Stepped-Care und automatische Qualitätskontrolle. Zugleich wurden Herausforderungen deutlich: Die niedrigere Vorhersagegüte der dritten Studie im Vergleich zu kleineren Datensätzen mit weniger strengen Verfahren mahnt zur Vorsicht gegenüber zu optimistischen Erwartungen. Zudem zeigte der Review, dass ein prominenter Anwendungsfall kaum implementiert wurde. Insgesamt braucht es dringend weitere Arbeiten zu ML-basierten Interventionen, ihrer Integration in klinische Abläufe und randomisierte Studien zur Wirksamkeit.Machine Learning (ML) and Digital Mental Health (DMH) hold significant promise for addressing the burden of mental disorders. However, methodologically sound research with large naturalistic samples is essential for practical implementation. Also, the extent of ML use in real-world DMH remains unclear. This dissertation contributes through four investigations: three feasibility studies and one systematic review. Two studies explored ML applications in a chat-based DMH intervention for youth. The first analyzed nearly 20,000 cases of young help-seekers to predict recurrent chat contact as an indicator of further help needs, reaching an AUROC of 0.68 on unseen test data. The second achieved similar predictive performance for perceived helpfulness of chats, using ~2,500 conversations with feedback. Notably, a transformer-based model did not outperform a simpler baseline. The third study used HRV data from ~2,600 participants to predict depressive symptoms. While performance was above chance, it was notably below earlier studies and insufficient for practical use. The final study reviewed 94 digital interventions for depression, showing that although ML is considered promising, its real-world use for personalizing DMH remains limited. The encouraging results in two large-scale chat studies suggest potential for ML in DMH, especially for stepped care and automated quality control. Still, challenges remain. The weaker performance in the third study—compared to prior work with smaller samples and simpler validation—raises concerns about overly optimistic expectations. The review also underscored the limited implementation of ML in practice. Finally, the difficulty of translating predictive performance into clinical value remains a key limitation. Further work is urgently needed on ML-based interventions, clinical integration, and trials evaluating real-world impact. Only then can ML’s potential in DMH be fully realized
Winning Votes and Changing Minds: Do Populist Arguments Affect Candidate Evaluations and Issue Preferences?
No full textPopulist rhetoric – presenting arguments in people-centric, anti-elite and ‘good v. evil’ frames – is said to provide populist parties and candidates with an advantage in electoral competition. Yet, identifying the causal effect of populist rhetoric is complicated by its enmeshment with certain positions and issues. We implement a survey experiment in the UK (n≈9,000), in which hypothetical candidates with unknown policy positions randomly make (non-)populist arguments, taking different positions on various issues. Our findings show that, on average, populist arguments have a negative effect on voters’ evaluations of the candidate profiles and no effect on voters’ issue preferences. However, populist arguments sway voters’ issue preferences when made by a candidate profile that voters are inclined to support. Among voters with strong populist attitudes, populist arguments also do not dampen candidates’ electoral viability. These findings suggest that populist rhetoric is useful in convincing and mobilizing supporters but detrimental in expanding electoral support.Peer Reviewe
Essays in Labor Economics
No full textDiese Dissertation untersucht Arbeitsmarktdynamiken anhand von drei konzeptionell verbundenen, aber unabhängigen Aufsätzen. Sie liefert empirische Evidenz und methodische Innovationen, die unser Verständnis von Arbeitsmärkten erweitern. Der erste Aufsatz untersucht, wie reduzierte Arbeitslosenversicherungsleistungen die Wiederbeschäftigungs- und Gründungsrate beeinflussen. Wir zeigen mithilfe eines quasi-experimentellen Designs, basierend auf einer spanischen Arbeitsmarktreform, dass eine Senkung des Arbeitslosengeldes zwar die Wiederbeschäftigungsrate erhöht, jedoch die Gründungsrate deutlich verringert. Der negative Effekt auf die Gründungsrate überwiegt den positiven Effekt auf die Wiederbeschäftigung. Dies deutet darauf hin, dass politische Maßnahmen, die nur auf Wiederbeschäftigung fokussieren, Zielkonflikte zwischen schneller Wiederbeschäftigung und der Förderung des Unternehmertums übersehen. Der zweite Aufsatz untersucht die Rolle sozialer Bindungen hinsichtlich geografischer Arbeitsmobilität. Vor dem Hintergrund der deutschen Wiedervereinigung analysieren wir, wie ehemalige Arbeitskollegen aus der Deutschen Demokratischen Republik (DDR) die Migrationsentscheidungen von Ostdeutschen beeinflussen. Unsere Ergebnisse zeigen, dass Ostdeutsche, die bereits in den Westen übergesiedelt sind, die Abwanderungswahrscheinlichkeit ihrer ehemaligen Arbeitskollegen signifikant erhöhen. Unsere Studie verdeutlicht, wie soziale Bindungen Informationshemmnisse verringern können, indem sie verzerrte Erwartungen über Migrationsergebnisse korrigieren.
Der letzte Aufsatz leistet einen methodischen Beitrag zur Harmonisierung der Berufscodierung in Deutschland im Zeitverlauf. Mit der Einführung des Berufsklassifikationssystems von 2010 wurden ältere Beschäftigungsdaten schwer vergleichbar. Dieser Aufsatz entwickelt einen Algorithmus, der die Berufscodierung über 30 Jahre harmonisiert und die Zuverlässigkeit von Langzeitanalysen verbessert.This dissertation explores labor market dynamics through three independent yet conceptually linked essays. It provides empirical evidence and methodological innovations that enhance our understanding of labor markets. The first essay examines how reduced unemployment insurance (UI) generosity affects transitions into (self-)employment. Leveraging a quasi-experimental design based on a Spanish labor market reform, we find that although reducing UI benefits increases the job-finding rate, it significantly decreases the startup rate. The negative effect on self-employment outweighs the positive re-employment effect, suggesting that policies focusing solely on the job-finding rate may overlook important trade-offs between encouraging rapid re-employment and supporting entrepreneurship. The second essay investigates the role of social connections in facilitating geographic labor mobility. Set in the context of German reunification, this essay investigates how former coworkers from the German Democratic Republic (GDR) influenced East Germans’ relocation decisions. Our findings reveal that East Germans who had already relocated to the West significantly increase their former peers' likelihood of migration. Our study highlights the importance of social ties in reducing information frictions through updating biased expectations about migration outcomes.
The last essay offers a methodological contribution by addressing the challenge of harmonizing occupation codes in Germany over time. With the occupational classification system of 2010, Germany introduced a new classification that made older employment data harder to compare. This essay develops an algorithm that harmonizes occupation codes over 30 years and improves the reliability of long-term analyses
Studien- und Prüfungsordnung für die Studienanteile Bildungswissenschaften und Sprach-bildung im lehramtsbezogenen Masterstudiengang (Ein-Fach) (für das Lehramt an Integrierten Sekundarschulen und Gymnasien)
No full textVerb-final subject suffixes in Ngwi (West-Coastal Bantu B861, DRC): language-internal innovation or language-external imposition?
No full textThis chapter presents a preliminary analysis of the synchronic distribution and diachronic origins of subject suffixes in the West-Coastal Bantu language Ngwi (B861). We describe their present-day use in six different monoverbal affirmative future and past constructions and in one biverbal affirmative construction. Except for subject suffixes, which are highly unusual from a Bantu-centric perspective, monoverbal affirmative past and future constructions in Ngwi are significantly more conservative than aspectual constructions with present-time reference and prefixal subject indexes. Past morphology is inherited from Proto-Bantu, while future morphology is built on the past template with some innovative future-specific forms. In the biverbal affirmative construction, the subject index follows the erstwhile auxiliary, which is undergoing univerbation with the nominalized lexical verb. Even though they could presumably be a recent innovation, we are unable to identify a language-internal source construction from which subject suffixes could have originated in Ngwi. Hence, we speculate that this innovative feature may have a language-external source, possibly first-language imposition from shifters originally speaking a non-Bantu language with subject suffixes. Other features possibly diagnostic of substratum interference in Ngwi include phonemic interior vowels, diphthongization, heterosyllabic vowel sequences, final vowel loss, umlaut, a significantly reshuffled noun class system, and the rise of lability in noncausal/causal alternations
Survival of Eurasian lynx in the human‐dominated landscape of Europe
No full textSurvival and cause‐specific mortality rates are vital for evidence‐based population forecasting and conservation, particularly for large carnivores, whose populations are often vulnerable to human‐caused mortalities. It is therefore important to know the relationship between anthropogenic and natural mortality causes to evaluate whether they are additive or compensatory. Further, the relation between survival and environmental covariates could reveal whether specific landscape characteristics influence demographic performance. We used telemetry data on 681 Eurasian lynx ( Lynx lynx ), a model apex predator with large spatial requirements, that were tracked across their European distribution. Through time‐to‐event analyses, we sought to determine the variables associated with differences in their survival. Illegal killing was the main cause of mortality (33.8%), and mortality rates were similar in protected and hunted populations (8.6% and 7.0% per year, respectively). Survival varied greatly across populations (70–95% per year). Across all study sites, higher hunting and anthropogenic mortality rates were partially compensated by lower rates of other mortality causes but not by natural mortality alone. Variation in survival depended on sex (female survival was 1.5 times greater than male survival) and seasonality (highest risk during hunting season and winter), and lower survival rates were correlated with higher human modification of landscapes at both coarse (home range composition) and fine (habitat use within home range) scales. Some variation in survival was driven by unobserved factors, which, given the high rates of human‐caused mortalities, including illegal killing, are of foremost concern. Due to the low natural mortality rates in protected and hunted populations, we conclude that anthropogenic causes of mortality are likely close to additive, such that maintaining or increasing refuge habitat with little human disturbance is critical to lynx conservation.Supervivencia del lince euroasiático en los paisajes europeos dominados por humanos. Las tasas de supervivencia y mortalidad por causas específicas son vitales para la previsión y conservación de poblaciones basadas en evidencias, sobre todo en el caso de los grandes carnívoros, cuyas poblaciones suelen ser vulnerables a la mortalidad causada por el hombre. Por esto es importante conocer la relación entre las causas de mortalidad antropogénicas y naturales para evaluar si son aditivas o compensatorias. Además, la relación entre la supervivencia y las covariables ambientales podría revelar si las características específicas del paisaje influyen en el rendimiento demográfico. Utilizamos datos telemétricos de 681 linces euroasiáticos ( Lynx lynx ), un depredador ápice modelo con grandes necesidades espaciales, que fueron rastreados a lo largo de su distribución europea. Mediante análisis cronológicos, intentamos determinar las variables asociadas a las diferencias en su supervivencia. La matanza ilegal fue la principal causa de mortalidad (33.8%), y las tasas de mortalidad auales fueron similares en las poblaciones protegidas y cazadas (8.6 y 7.0%, respectivamente). La supervivencia anual varió mucho de una población a otra (del 70 al 95%). En todos los lugares de estudio, las mayores tasas de mortalidad antropogénica y por caza se vieron parcialmente compensadas por las menores tasas de otras causas de mortalidad, pero no por la mortalidad natural por sí sola. La variación en la supervivencia dependió del sexo (la supervivencia de las hembras fue 1.5 veces mayor que la de los machos) y de la estacionalidad (el mayor riesgo se registró durante la temporada de caza y el invierno), y las menores tasas de supervivencia se correlacionaron con una mayor modificación humana de los paisajes tanto a escala gruesa (composición del área de distribución) como fina (uso del hábitat dentro del área de distribución). Algunas variaciones en la supervivencia se debieron a factores no observados. Dadas las elevadas tasas de mortalidad causada por el hombre, incluidas las matanzas ilegales, estos factores son motivo de gran preocupación. Debido a las bajas tasas de mortalidad natural en poblaciones protegidas y cazadas, concluimos que las causas antropogénicas de mortalidad son probablemente casi aditivas, de modo que mantener o aumentar el hábitat de refugio con poca perturbación humana es crítico para la conservación del lince.European Union INTERREG IVMAVA Foundation 10.13039/100013324Norges Forskningsråd 10.13039/501100005416SCOPES programmeEuropean Commission 10.13039/501100000780Nature Protection Division of the County Governor's Office for Innlandet, Viken, Vestfold & Telemark, Trøndelag, Nordland, Troms & Finnmark CountiesCzech Academy of Sciences Institutional Research PlanSwiss Foundation for the Scientific ResearchNorwegian Directorate for Nature ManagementProf. Klaus Riepe und Ellen Riepe‐Brunnström‐StiftungStotzer‐Kästli FoundationDeutsche Bundesstiftung Umwelt 10.13039/100007636People's Trust for Endangered SpeciesAgassiz FoundationNature Conservation Agency of the Czech RepublicSociety for the Protection of Animals of the Canton of VaudJavna Agencija za Raziskovalno Dejavnost RSPolish Ministry of Science and Higher EducationNaturvårdsverket 10.13039/501100004357Operational Programme Environment through State Environmental Fund of the Czech RepublicEuropean Union Interreg IIIAWWF SwitzerlandHunting Administration of the Canton of Vaud, Fribourg and BernSwiss Federal Office for the EnvironmentSwiss League for Protection of Nature (Pro Natura)Svenska Forskningsrådet Formas 10.13039/501100001862Dr. Berthold Suhner FoundationOperational Programme Research, Development and EducationRegina‐Bauer StiftungPeer Reviewe
Does neural computation feel like something?
No full textArtificial neural networks are becoming more advanced and human-like in detail and behavior. The notion that machines mimicking human brain computations might be conscious has recently caused growing unease. Here, we explored a common computational functionalist view, which holds that consciousness emerges when the right computations occur—whether in a machine or a biological brain. To test this view, we simulated a simple computation in an artificial subject’s “brain” and recorded each neuron’s activity when the subject was presented with a visual stimulus. We then replayed these recorded signals back into the same neurons, degrading the computation by effectively eliminating all alternative activity patterns that otherwise might have occurred (i.e., the counterfactuals). We identified a special case in which the replay did nothing to the subject’s ongoing brain activity—allowing it to evolve naturally in response to a stimulus—but still degraded the computation by erasing the counterfactuals. This paradoxical outcome points to a disconnect between ongoing neural activity and the underlying computational structure, which challenges the notion that consciousness arises from computation in artificial or biological brains.Peer Reviewe