2003 Fiscal Year Final Research Report Summary
Statistical assessment of Matching and Other Strategies for Selecting Comparison Subjects in Epidemiologic Studies
| Project/Area Number |
14580356
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Statistical science
|
|---|
| Research Institution | Radiation Effects Research Foundation |
|---|
Principal Investigator |
COLOGNE John Radiation Effects Research Foundation, Department of Statistics, Assistant Department Chief, 統計部, 副部長(研究員) (50344411)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
FUJII Yoshinori Miyazaki University, Department of Mathematics Education, Assistant Professor, 教育文化学部, 助教授 (10218985)
IZUMI Shizue Radiation Effects Research Foundation, Department of Statistics, Research Scientist, 統計部, 研究員 (70344413)
|
|---|
| Project Period (FY) |
2002 – 2003
|
| Keywords | Epidemiologic study design / Cohort study / Case-control study / Risk factor interaction / Control selection methods / Statistical efficiency / Confounding factor / Statistical interaction |
|---|
| Research Abstract |
The major results are related to control selection and studies of interaction. Matching on a risk factor is not generally effective for studies of interaction. Rather, counter matching on a known risk factor when selecting controls is preferred. We compared the efficiency with matching to that with counter matching using simulation when the risk factor known at the time of control selection is dichotomous or continuous. We also evaluated efficiency of the case-cohort design via simulation. We investigated statistical power using calculations based on the likelihood. The power calculation work is in collaboration with Dr.Charles Land of the U.S.National Cancer Institute.
|
