The purposes of this study were: (i) to compare the extent of depression in the nursing home and community-dwelling elderly people, and (ii) to find the variables including residential status and other related variables explaining depression in Korea. Two sets of secondary data were combined and used to achieve the objectives of this study. One data set relating to elderly people in nursing homes was from a part of 2002 Kyunggi Long-term Care System Construction Study by Kyunggi Research Institute, Kyunggi province, Korea. The other data set for community-dwelling elderly people was from a part of 2001 National Long-term Care Study by Korea Institute for Health and Social Affairs. The data set for this study included 307 elderly people living in nursing homes and 166 elderly people living in the community. Depression was measured and determined using the geriatric depression scale-short form (GDS-SF) Korean version, with scores of 8 or higher to indicate possible depression. A total of 39.3%[95% confidence interval (CI); 32.1-46.9%] of the sample in the community elderly showed symptoms of depression, higher than the rate found in the nursing home elderly (24.0%) (95% CI; 19.5-29.2%). The mean (standard deviation) GDS-SF score for the elderly in the nursing home was 6.1 (3.4), and 7.4 (4.3) for elderly in the community, the difference being statistically significant [t (275.5) = -3.33; P < 001]. As a result of multiple logistic regression analysis, residential status has appeared as an important predictor after controlling other related variables. The adjusted odds ratio of depression associated with the nursing home residents in residential status, all other factors being equal, was 3.14 (95% CI; 1.30-7.58). Community-dwelling elderly people have higher odds of depression. These findings suggest that there is a need to provide adequate health-related care services for the elderly people in the community.