关键词: 成对数据, APIM, 非独立性/相依性, 可区分, 不可区分

Abstract: The dyad is the fundamental unit of interpersonal interaction and interpersonal relations, and the phenomenon laying the dyad is an important topic in social science. The data collected from the dyad is characterized with the interdependence feature of interpersonal relationship, which manifests in the target effect and attracts researches’ interests. The actor-partner interdependence model (APIM) is a simple yet compelling model of dyadic data, which takes the interdependence into consideration and estimates actor effect and target effect simultaneously. There are three kinds of variables according to the sources of variance, between-dyads variables, within-dyads variables and mix variables, and data can be distinct into exchangeable and non-exchangeable one depending on distinguishing variable existing or not. Dyadic data analysis always begins with an analysis of the degree of nonindependence. Then three methods can be used to estimate the effects, pooled regression, multilevel modeling, and structural equation modeling. Next, we illustrate the analysis with LISREL computer program for APIM path analysis without or with latent variables. Finally, the applications of APIM are brief reviewed and future development and implications for relationship research in China are presented.

Key words: dyadic data, Actor-Partner Interdependence Model (APIM), interdependence/nonindependence, distinguishable, indistinguishable