The hierarchical analysis on pair or group data

ペア シュウダン データ ニ オケル カイソウセイ ノ ブンセキ

シミズ, ヒロシ

Shimizu, Hiroshi

オオサカ ダイガク ダイガクイン ニンゲン カガク ケンキュウカ タイジン シャカイ シンリガク ケンキュウシツ

階層データ

階層線形モデル

混合モデル

ペアワイズ相関分析

Pair data

hierarchical data

Hierarchical linear model

Mixed model

pair-wise correlation analysis

The purpose of this article is to show the effective methodologies for analyzing data which have hierarchical structures such as pairs and groups, and compare advantages of those analyses. It is difficult to perform sufficient analysis to pair or group data by using conventional analysis, despite plenty opportunities to obtain such data in the field of social psychology. HLM and pair-wise correlation analysis not only avoid errors made the conventional analysis but also become possible to acquire the important information about dynamics of interpersonal relationships or groups. This article explained examples of application and interpretative method of HLM, using a MIXED procedure of SAS. Also comparison and limitation of each hierarchical analysis were described.