Dimensionality in canonical correlation of settlements characteristics and levels of infrastructure in rural Akwa Ibom State, Nigeria in 2014
This study examines the relationship between rural settlements characteristics and access levels to infrastructural facilities in Akwa Ibom State. A total of 50 rural settlements were randomly selected and utilized for this study. Access levels to five basic social infrastructural facilities and six...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Iranian Journal of Health, Safety and Environment
2015-11-01
|
| Series: | Iranian Journal of Health, Safety and Environment |
| Subjects: | |
| Online Access: | http://www.ijhse.ir/index.php/IJHSE/article/view/96 |
| Summary: | This study examines the relationship between rural settlements characteristics and access levels to infrastructural facilities in Akwa Ibom State. A total of 50 rural settlements were randomly selected and utilized for this study. Access levels to five basic social infrastructural facilities and six variables of settlements characteristics were examined simultaneously using multivariate method of canonical correlation analysis. The question this research set to address is: along how many dimensions are the settlements characteristics related to the levels of infrastructural facilities? Five canonical functions were produced out of which the first two were found to be significant at 0.001 levels. The first canonical correlation coefficient of 0.88 (0.86 adjusted), representing 78% overlapping variance for the first pair of canonical functions or variates was obtained while the second canonical correlation coefficient was 0.59 (0.50 adjusted) and thus, represent 34% overlapping variance for the second pair of canonical variates. These canonical correlations are highly significant and represent a substantial relationship between pairs of canonical variates. The result showed that in Akwa Ibom State, rural settlements characteristics and basic social infrastructure can be related at least in two major dimensions. This study therefore reaffirms the potentials of canonical correlation analysis as a useful tool for establishing empirically based linkages between two data sets. |
|---|---|
| ISSN: | 2345-3206 2345-5535 |