Methods for the Analysis of Longitudinal Dyadic Data with an Application to Intergenerational Exchanges of Family Support

在社会研究中，通常会收集成对被试的数据。例如，在家庭研究中，人们感兴趣的是家庭成员对关系质量的看法的一致程度，或者父母与子女关系的强度如何取决于父母和子女的特征。在组织研究中，同事之间的合作可能取决于与他们的相对角色和公司精神有关的因素。二元数据提供了关于人际关系过程的详细信息，但由于其高度复杂的结构，分析起来具有挑战性：由于人们对个人之间随时间的依赖关系感兴趣，二元数据往往是纵向的，二元数据可能被聚集到更大的群体中（例如在家庭或组织中），感兴趣的变量，如感知和态度，可能通过多个指标来衡量。该研究的主要目的是开发新的方法来分析纵向多元二元数据，虽然这些方法有许多潜在的应用，但本项目侧重于一个重要的案例研究：使用英国家庭小组研究和理解协会（或英国家庭纵向研究，UKHLS）的数据分析父母和他们的成年子女之间的支持交流。要调查的实质性问题包括：-有什么特点与给予和接受支持的respondent-parent和respondent-adult后代二人组？对于一个给定的二元体来说，给予和接受帮助在多大程度上是持续的？与生命周期事件相关的交换程度如何？交换的互惠水平和性质是什么？对于父母和成年子女非同居的受访者，互惠规范是否在对受访者的时间和资源有竞争需求的情况下受到限制？资金转移和“实物”转移（即花时间帮助某人）在多大程度上是互补的，在多大程度上是替代的？这如何取决于捐助者的社会经济情况？我们提出了一个通用的潜变量建模框架，随着时间的推移，受访者和他们的父母和成年子女之间的交流进行分析。我们的方法将在三个主要方面扩展现有的研究。首先，大多数以前的研究都研究了横截面或两个波之间的交换。我们将使用来自五波BHPS/UKHLS的纵向数据，跨越16年的时间，允许检查随着时间的推移，捐助者和受援国的情况下，交换支持和变化之间的关联。第二，通常将提供的支助与接受的支助分开分析，这就排除了对互惠性的研究。我们提出了一个联合模型的双向交流与支持给予和支持接受治疗作为一个多变量的响应，和响应之间的协方差测量世代之间的互惠程度。第三，家庭支持是通过一套不同类型帮助的二元指标来衡量的。对各个项目进行总分的普遍做法隐含着对每个指标给予同等权重。相反，我们将指定一个潜在变量模型，该模型将二进制项与代表整体帮助行为的基础变量联系起来。该模式将扩大到包括家庭成员之间资金转移的指标。项目产出将包括方法论和更注重实质内容的文章，以便将研究结果传达给社会科学家。R例程将被编写来实现所提出的方法。还将制作大量的在线资源，包括如何在R中进行分析的详细说明，建模决策的讨论以及结果的解释。

Data on pairs of subjects (dyads) are commonly collected in social research. In family research, for example, there is interest in the extent of agreement in family members' perceptions of relationship quality or how the strength of parent-child relationships depends on characteristics of parents and children. In organisational research, cooperation between coworkers may depend on factors relating to their relative roles and company ethos. Dyadic data provide detailed information on interpersonal processes, but they are challenging to analyse because of their highly complex structure: they are often longitudinal because of interest in dependencies between individuals over time, dyads may be clustered into larger groups (e.g. in families or organisations), and variables of interest such as perceptions and attitudes may be measured by multiple indicators. The principal aim of the proposed research is to develop new methods for the analysis of longitudinal multivariate dyadic data.While these methods have a number of potential applications, this project focuses on one important case study: the analysis of exchanges of support between parents and their adult children, using data from the British Household Panel Study and Understanding Society (or UK Household Longitudinal Study, UKHLS). Substantive questions to be investigated include:- What characteristics are associated with giving and receiving support for respondent-parent and respondent-adult offspring dyads? To what extent is the giving and receipt of help persistent over time for a given dyad? And to what extent is the level of exchange associated with lifecycle events?- What is the level and nature of reciprocity of exchanges?- For respondents with non-coresident parents and adult offspring, are norms of reciprocity strained where there are competing demands on respondents' time and resources?- To what extent are financial transfers and 'in kind' transfers (i.e. spending time to help someone) complementary and to what extent do they appear to be substitutes? How does this depend on the socio-economic circumstances of the donor? We propose a general latent variable modelling framework for the analysis of exchanges between respondents and their parents and adult children over time. Our approach will extend existing research in three major ways. First, most previous research has studied exchanges at a cross-section or between two waves. We will use longitudinal data from five waves of BHPS/UKHLS, spanning a 16-year period, which permits examination of associations between exchanges of support and changes in donor and recipient circumstances over time. Second, it is common to analyse support given separately from support received, which precludes the study of reciprocity. We propose a joint model of bidirectional exchanges with support given and support received treated as a multivariate response, and covariances between responses measuring the extent of reciprocation between generations. Third, family support is measured by a set of binary indicators of different kinds of help. The widespread practice of taking the sum score across items implicitly assigns equal weight to each indicator. We will instead specify a latent variable model that relates the binary items to underlying variables representing overall helping behaviour. The model will be extended to include indicators of financial transfers between family members. Sensitivity of results to measurement error and non-random dropout will also be considered.Project outputs will include both methodological and more substantively-oriented articles to communicate the research to social scientists. R routines will be written to implement the proposed methods. Extensive online resources will also be produced, including detailed instructions of how to conduct analyses in R, discussion of modelling decisions, and interpretation of results.