|Relative importance (relative to the most important, getting an importance of 100) of explanatory variables with ‘ever refused condoms’ and ‘ever used condoms’ as response variables (order according to the importance of ‘ever use condoms’)|
|Variable||Ever used condom||Ever refused condom|
|Perceived need for condoms||100||0|
|Knowledge about correct use of condoms||43.6||0|
|Knowledge about HIV status of partner||0.12||4.35|
|Presence of stigma on condom||0||100|
|Presence of sexual risk behaviour||0||37.52|
|Knowledge about own HIV status||0||12.01|
|Knowledge about condoms||0||9.49|
|Knowledge about HIV||0||3.81|
|Source of condoms||0||0|
|Belief that condoms can prevent HIV||0||0|
Note: Importance, for a particular variable, is the sum, across all nodes in the tree, of the improvement scores between this variable and the best splitter at a particular node. This means that a variable that does not occur in the tree because always “second best” and not selected as the main splitter, can occur as very important.
Chandran et al.
Chandran et al. BMC Public Health 2012 12:381 doi:10.1186/1471-2458-12-381