Geänderte Inhalte

Wir trauern um Dr. Heinz Liebeck, unseren langjährigen, sehr geschätzten Kollegen und Supervisor. Wir danken Dr. Peter Zezula für das Verfassen des u.s. Nachrufs, den Sie auch auf der GEMI-Homepage unter https://www.psych.uni-goettingen.de/de/news/nachrichten-items/heinz-liebeck-ist-verstorben lesen können.

None

Objective Blended cognitive behavioral therapy (bCBT) combines face-to-face therapy with digital elements, such as digital health apps. This pilot study aimed to explore the effectiveness and safety of a novel bCBT application for treating unipolar depression in adults combined with cognitive behavioral therapy (CBT) compared to CBT alone in routine care. Methods Patients (N = 82) were randomly assigned to bCBT (n = 42) or CBT (n = 40) over 12 weeks. bCBT consisted of weekly CBT sessions accompanied by the elona therapy depression module (a bCBT application for unipolar depression) for use between sessions. Standard CBT consisted of weekly CBT sessions. Outcomes (6,12 weeks) were analyzed with linear mixed models. Results Improvements in depressive symptoms (BDI-II, PHQ-9) were descriptively larger for the bCBT group. Yet, this difference did not reach statistical significance. bCBT was superior to standard CBT in secondary outcome measures of psychological health (d = .50) and generalized anxiety symptoms (d = -.45). In other secondary outcomes (BAI, PSWQ, GSE, WHOQOL-BREF), improvements were descriptively larger for bCBT compared to CBT. Conclusion This pilot study provided preliminary evidence that bCBT might be advantageous in comparison to CBT alone in the treatment of depression, but larger RCTs of the bCBT application are needed.

When assessing text, supervised natural language processing (NLP) models have traditionally been used to measure targeted constructs in the organizational sciences. However, these models require significant resources to develop. Emerging “off-the-shelf” large language models (LLM) offer a way to evaluate organizational constructs without building customized models. However, it is unclear whether off-the-shelf LLMs accurately score organizational constructs and what evidence is necessary to infer validity. In this study, we compared the validity of supervised NLP models to off-the-shelf LLM models (ChatGPT-3.5 and ChatGPT-4). Across six organizational datasets and thousands of comments, we found that supervised NLP produced scores were more reliable than human coders. However, and even though not specifically developed for this purpose, we found that off-the-shelf LLMs produce similar psychometric properties as supervised models, though with slightly less favorable psychometric properties. We connect these findings to broader validation considerations and present a decision chart to guide researchers and practitioners on how they can use off-the-shelf LLM models to score targeted constructs, including guidance on how psychometric evidence can be “transported” to new contexts.