Methodology
Case-control geographic clustering for residential histories accounting for risk factors and covariates
-
* Corresponding author: Geoffrey M Jacquez jacquez@biomedware.com
1 BioMedware, 516 North State Street, Ann Arbor, MI 48104-1236, USA
2 The University of Michigan, School of Public Health, Ann Arbor, MI, USA
International Journal of Health Geographics 2006, 5:32 doi:10.1186/1476-072X-5-32
Published: 3 August 2006Additional files
Additional File 1:
Animation of local clustering of cases at k = 7 accounting for smoking, gender, age, race and education. Controls indicated with a "+", cases by circles. Blue is a not significant local Q-statistic for that case, yellow p < 0.05, orange p < 0.01, Red p < 0.001.
Format: ADDI Size: 6.6MB Download file
Additional File 2:
Animation of focused clustering about business address histories at k = 8. Industries are indicated by triangles, controls by a "+", cases by circles. Triangles are color-coded based on statistical significance after covariate adjustment. Green is not significant, yellow p < 0.05, orange p < 0.01, Red p < 0.001.
Format: ADDI Size: 4.9MB Download file