Review: Long-term Life Recovery Processes Among Survivors of the 1995 Kobe Earthquake: 1999, 2001, 2003, and 2005 Life Recovery Social Survey Results Shigeo Tatsuki Professor, Department of Sociology, Doshisha Universty Kamigyo-ku, Kyoto 602-8580, Japan E-mail: tatsuki@gold.ocn.ne.jp [Received September 25, 2007; accepted December 4, 2007] This paper summarizes findings from life recovery surveys conducted in 1999, 2001, 2003, and 2005 among 1995 Kobe earthquake survivors. The 1999 survey (N=915) developed some of the key scales for the project, including life recovery, physical and psychological stress, family relations, and civic-mindedness. The 2001 study (N=1203) integrated 1999 study findings and those from 1999 grassroots assessment of life recovery, from which a seven critical element model of life recovery was constructed. The effects of these seven critical elements on life recovery were empirically tested and validated by general linear model (GLM) analysis. The 2003 (N=1203) and 2005 (N=1028) studies focused both on life recovery outcomes and on intervening life recovery processes. Structural equation modeling (SEM) identified causal links among recovery-promotion factors, recovery processes such as event impact stabilization, and event evaluation through community empowerment, and recovery outcomes. Event impact was a process through which impact caused by earthquake damage, loss, and/or stress was alleviated by housing, household finances, and stress management. Through event evaluation, social ties and community rebuilding efforts directly or indirectly facilitated the reframing of earthquake experiences into positive narratives. Research and policy implications of these findings are discussed in the end. Keywords: The 1995 Kobe earthquake, social survey, long-term life recovery, GLM, SEM 1. Introduction This paper summarizes major findings from the Hyogo Life Recovery Survey Project, a series of four consecutive cross-sectional social surveys, in order to study long-term life recovery processes among 1995 Kobe earthquake survivors. The first survey was conducted in 1999, five years after the Kobe earthquake. Questionnaires were repeatedly administered to those who resided in the same 330 affected areas out of which ten respondents were randomly selected each time in 2001, 2003, and 2005. The research framework evolved in three stages over the project’s seven-year period. The first stage was exploratory because no preceding long-term life recovery model applicable to an urban mega-disaster (Kawata, 1995 [1]) existed either nationally or internationally. The first survey, administered in March 1999, was designed to develop several key scales that could be repeatedly used in subsequent surveys. Among these measures, an important development was that of life recovery, which was designed to quantify the dependent variable for the project. The 1999 survey also examined demographic, disaster impact, and social characteristics of those who showed a higher level of life recovery (Tatsuki and Hayashi, 2000 [2]). The second stage integrated findings from the 1999 study and from the grassroots assessment workshop on life recovery conducted in summer 1999 (Tatsuki and Hayashi, 2001 [3]); refer also to the article by Tamura in this issue). Based on these findings, the seven critical element model of life recovery was constructed. This model guided research framework building for the 2001 Hyogo Life Recovery Survey. The third stage refined the research strategy that was then used in the 2003 and 2005 surveys. These two surveys focused on both life recovery outcomes and intervening life recovery process variables. The 2003 survey resulted in the construction of the long-term life recovery process and outcome model from a mega-disaster (Tatsuki et al, 2004 [4]). The 2005 survey together with the 2003 and 2004 grassroots assessment workshops conducted in Kobe and southern Hyogo areas (Tatsuki, 2004 [5]) confirmed, in general, both internal and external validity of the final life recovery process and outcome model. 1.1. Preceding Studies on the Recovery Process in Japan and the US Based on ethnographic interviews (cf., Shigekawa and Hayashi, 1997 [6]) of Kobe earthquake victims in Nishinomiya, Aono, Tanaka, Hayashi, Shigekawa, and Miyano (1998 [7]) found three distinctive timeframes in victim disaster response behavior. This provided the basis for subsequent quantitative analysis that Tatsuki, S. 484 Journal of Disaster Research Vol.2 No.6, 2007