PDF(Pubmed)
Abstract：
Background: In attempts to elucidate PTSD, recent factor analytic studies resulted in complex models with a proliferating number of factors that lack psychometrical and clinical utility. Recently, suggestions have been made to optimize factor analytic practices to meet a refined set of statistical and psychometric criteria.Objective: This study aims to assess the factorial structure of the German version of the PCL-5, implementing recent methodological advancements to address the risk of overfitting models. In doing so we diverge from traditional factor analytical research on PTSD.Method: On a large-scale sample of the German general population (n = 1625), exploratory factor analyses were run to investigate the dimensionality found within the data. Subsequently, we validated and compared all model suggestions from our preliminary analyses plus all standard and common alternative PTSD factor models (including the ICD-11 model) from previous literature with confirmatory factor analyses. We not only consider model fit indices based on WLSMV estimation but also deploy criteria such as favouring less complex models with a parsimonious number of factors, sufficient items per factor, low inter-factor correlations and number of model misspecifications.Results: All tested models showed adequate to excellent fit in respect to traditional model fit indices; however, models with two or more factors increasingly failed to meet other statistical and psychometric criteria.Conclusion: Based on the results we favour a two-factor bifactor model with a strong general PTSD factor and two less dominant specific factors - one factor with trauma-related symptoms (re-experiencing and avoidance) and one factor with global psychological symptoms (describing the trauma\'s higher-order impact on mood, cognition, behaviour and arousal).From the perspective of clinical utility, we recommend the cut-off scoring method for the German version of the PCL-5. Basic psychometric properties and scale characteristics are provided.
We contribute new insights to the debate on the factor structure of the PTSD Checklist (PCL-5) based on a large German general population sample deploying the newest methodological developments in a revised factor-analytical approach.Combining theoretical, statistical and practical considerations, we favour a two-factor bifactor model with a strong general PTSD factor and two less dominant specific factors – one factor with trauma-related symptoms and one factor with global psychological symptoms.For clinical practitioners, we recommend using the cut-off scoring method.
We contribute new insights to the debate on the factor structure of the PTSD Checklist (PCL-5) based on a large German general population sample deploying the newest methodological developments in a revised factor-analytical approach.Combining theoretical, statistical and practical considerations, we favour a two-factor bifactor model with a strong general PTSD factor and two less dominant specific factors – one factor with trauma-related symptoms and one factor with global psychological symptoms.For clinical practitioners, we recommend using the cut-off scoring method.
摘要：
背景：为了阐明创伤后应激障碍，最近的因素分析研究导致了复杂的模型，其中许多因素缺乏心理测量和临床实用性。最近,已经提出了优化因子分析实践的建议，以满足一套完善的统计和心理测量标准。目的：本研究旨在评估德语版PCL-5的阶乘结构，实施最新的方法进步，以解决过度拟合模型的风险。在此过程中，我们与传统的PTSD因素分析研究有所不同。方法：在德国普通人群的大规模样本中（n=1625），我们进行了探索性因素分析,以调查在数据中发现的维度.随后，我们验证并比较了我们初步分析的所有模型建议,以及以往文献中的所有标准和常见替代PTSD因子模型(包括ICD-11模型)和验证性因子分析.我们不仅考虑基于WLSMV估计的模型拟合指数，而且还部署标准，例如偏爱具有简约数量的因素的较不复杂的模型，每个因素足够的项目，低因素间相关性和模型错误规格的数量。结果：所有测试的模型都显示出与传统模型拟合指数的良好拟合；然而，具有两个或多个因素的模型越来越不符合其他统计和心理测量标准。结论：根据结果，我们倾向于双因素模型，具有强大的一般PTSD因素和两个不太占优势的特定因素-一个因素与创伤相关症状（重新体验和回避）和一个因素与整体心理症状（描述创伤对情绪的高阶影响，认知，行为和唤醒)。从临床效用的角度来看，我们建议使用德语版PCL-5的截止评分方法。提供了基本的心理测量特性和尺度特征。
我们基于大量德国普通人群样本，在修订的因素分析方法中部署最新的方法学发展，为关于PTSD清单（PCL-5）的因素结构的辩论提供了新的见解。结合理论，统计和实际考虑，我们支持双因素双因素模型,该模型具有强的一般PTSD因素和两个不太占优势的特异性因素-一个因素与创伤相关症状有关,一个因素与整体心理症状有关.对于临床医生来说，我们建议使用截止评分法。
我们基于大量德国普通人群样本，在修订的因素分析方法中部署最新的方法学发展，为关于PTSD清单（PCL-5）的因素结构的辩论提供了新的见解。结合理论，统计和实际考虑，我们支持双因素双因素模型,该模型具有强的一般PTSD因素和两个不太占优势的特异性因素-一个因素与创伤相关症状有关,一个因素与整体心理症状有关.对于临床医生来说，我们建议使用截止评分法。
