Multitrait-multimethod (MTMM) analyses are used in psychology to assess convergent and discriminant validity and to study method effects. Most current MTMM approaches assume that measures have equal convergent and discriminant validity across the entire range of trait values and thus do not account for potential trait × method interactions. A novel approach is presented that allows analyzing trait × method interactions using factor mixture modeling. The new MTMM mixture model allows identifying latent classes of individuals who differ with respect to convergent and discriminant validity. The new approach was applied to mother’s and father’s ratings of children’s attention deficit hyperactivity disorder (ADHD) symptoms (N = 618). Results revealed four latent classes: one with no symptom levels, two with low symptom levels, and one with moderate symptom levels. Three classes showed evidence for convergent and discriminant validity, whereas a low symptom class lacked convergent validity for ratings of inattention. Keywords: mixture distribution, factor mixture model, convergent and discriminant validity, multitrait-multimethod analysis

中文翻译：

---

使用混合分布多特征-多方法模型检查特征×方法相互作用

多特质-多方法 (MTMM) 分析在心理学中用于评估收敛效度和判别效度并研究方法效果。大多数当前的 MTMM 方法假设度量在整个特征值范围内具有相同的收敛效度和区分效度，因此不考虑潜在的特征 × 方法相互作用。提出了一种新方法，允许使用因子混合建模分析特征×方法相互作用。新的 MTMM 混合模型允许识别在收敛效度和判别效度方面不同的个体的潜在类别。新方法应用于母亲和父亲对儿童注意力缺陷多动障碍 (ADHD) 症状的评分 (N = 618)。结果揭示了四个潜在类别：一个没有症状水平，两个症状水平低，和一个具有中等症状水平。三个类别显示出收敛效度和区分效度的证据，而低症状类别缺乏注意力不集中评级的收敛效度。关键词：混合分布，因子混合模型，收敛和判别效度，多特征多方法分析