Predictors of Retention in Longitudinal Study - MIDJA (Midlife in the Japan) longitudinal survey case -

[Speaker] Nakajima, Reiko:1
[Co-author] Karasawa, Mayumi:1
1:Tokyo Woman's Christian University (Japan)

In longitudinal study design, how predict the retention of the study is one of the key issues. This study explored predictors for continuing participation, using data from MIDJA two-wave surveys, conducting in 2009 (N=1027) and 2012 (N=657) respondents aged 30-85.
Logistic regression analyses of sociodemographic factors and respondents' health variables were used to predict participation. Results showed that (1) Being older, female, or having better functional health were predictors of higher participation rates. (2) The participation rate was higher with age, however it dropped in the oldest age group, 75 or older, when using 5 age categories as variables. (3) Interaction analysis clarified that better education predicted participation among females compared to males.
The findings suggested that sociodemographic and health factors such as younger, male and poorer health impacted on nonparticipation, however, further analyses including interplay of those variables regarding nonparticipations are needed when considering adjustments.
Advanced Search