主成分分析・クラスタ分析による被災経験のあるA県B町住民の防災行動の類型化

Classification of Disaster Prevention Behaviors of Residents of Town B in Prefecture a Who Have Experienced Disasters Using Principal Component and Cluster Analysis

要旨

　目的：住民の防災行動の特徴を分類し，今後の防災教育のあり方について示唆を得ることを目的とした．

　方法：2021年にB町の20〜89歳の住民1,600人を対象に，防災行動に関する無記名自記式質問紙調査を実施し，主成分分析，クラスタ分析およびχ二乗検定を行った．

　結果：回答者は1,062人（回収率：66.3％）であった．住民の928人（99.3％）が何らかの防災行動を実施していた．防災行動の主成分分析から「確実な避難のための準備」「地域関係を維持する準備」「必須ツールのエネルギー確保」等の7成分が抽出された．クラスタ分析では各主成分の実施状況が高いものやいずれも低いものなど7クラスタに分類された．各クラスタは年齢，職業，家族状況，個人の生活習慣，災害自己効力感などにおいて有意な違いがみられた．

　結論：住民の防災行動はその主成分からクラスタ化することができ，各クラスタの特徴に応じた防災教育手法の必要性が示唆された．

　Objective: This study aimed to classify the characteristics of residents' disaster prevention behavior and to obtain suggestions for future disaster prevention education.

　Methods: In 2021, an anonymous, self-administered questionnaire survey about disaster prevention behaviors was conducted which included 1,600 residents aged between 20 and 89 years in Town B, with their responses subjected to principal component analysis, cluster analysis, and the chi-squared test.

　Results: Responses were obtained from 1,062 eligible residents(response rate: 66.3%). Of the respondents, 928(99.3%) implemented disaster prevention behaviors. In the principal component analysis of disaster prevention behaviors, seven components were extracted, including "preparation to ensure safe evacuation," "preparation to maintain community relationships," and "securing energy for essential tools." In the cluster analysis, residents were divided into seven clusters, including clusters in which all seven principal components were implemented at high and low rates, respectively. Significant differences among the clusters were noted in terms of age, occupation, family situation, personal lifestyle, and disaster self-efficacy.

　Conclusion: The disaster prevention behaviors of the residents could be divided into clusters based on the principal components, with the findings suggesting the need for educational techniques on disaster prevention to be tailored toward the characteristics of respective clusters.