Response surface analysis (RSA) allows researchers to study whether the degree of congruence between two predictor variables is related to a potential psychological outcome. Here, we adapt RSA to the case in which the two predictor variables whose congruence is of interest refer to individual differences in within-person associations (WPAs) between variables that fluctuate over time. For example, a WPA-congruence hypothesis in research on romantic relationships could posit that partners are happier when they have similar social reactivities-that is, when they have similarly strong WPAs between the quantity of their social interactions and their momentary well-being. One method for testing a WPA-congruence hypothesis is a two-step approach in which the individuals' WPAs are first estimated as random slopes in respective multilevel models, and then these estimates are used as predictors in a regular RSA. As an alternative, we suggest combining RSA with multilevel structural equation modeling (MSEM) by specifying the WPAs as random slopes in the structural equation and using their latent second-order terms to predict the outcome on Level 2. We introduce both approaches and provide and explain their corresponding computer code templates. We also compared the two approaches with a simulation study and found that the MSEM model-despite its complexities (e.g., nonlinear functions of latent slopes)-has advantages over the two-step approach. We conclude that the MSEM approach should be used in practice. We demonstrate its application using data from a daily diary study and offer guidance for important decisions (e.g., about standardization). (PsycInfo Database Record (c) 2024 APA, all rights reserved).

中文翻译：

---

调查人内关联之间一致性的影响：响应面分析的两种扩展的比较。


响应面分析 (RSA) 允许研究人员研究两个预测变量之间的一致性程度是否与潜在的心理结果相关。在这里，我们将 RSA 应用于这样的情况：两个感兴趣的同余预测变量是指随时间波动的变量之间的人内关联 (WPA) 中的个体差异。例如，浪漫关系研究中的 WPA 一致性假设可能认为，当伴侣具有相似的社交反应时，即当他们的社交互动数量和瞬时幸福感之间具有相似强的 WPA 时，他们会更快乐。测试 WPA 同余假设的一种方法是两步法，其中首先将个体的 WPA 估计为各自多级模型中的随机斜率，然后将这些估计用作常规 RSA 中的预测变量。作为替代方案，我们建议将 RSA 与多级结构方程建模 (MSEM) 结合起来，方法是将 WPA 指定为结构方程中的随机斜率，并使用其潜在二阶项来预测第 2 级的结果。我们介绍了这两种方法并提供和解释其相应的计算机代码模板。我们还通过模拟研究对这两种方法进行了比较，发现 MSEM 模型尽管复杂（例如，潜在斜率的非线性函数），但比两步方法具有优势。我们的结论是 MSEM 方法应该在实践中使用。我们使用日常日记研究的数据来演示其应用，并为重要决策（例如，关于标准化）提供指导。 （PsycInfo 数据库记录 (c) 2024 APA，保留所有权利）。