Statistical analysis
This study applied logit regression to identify factors affecting
each of the three levels of community participation under the
PHBM program. Each level of community participation can be
captured as binary-choice models, which presume that households
face a choice problem of participation and that the choice depends
on identifiable characteristics. Let yi denote a choice variable such
that yi ¼ 1 if respondent i chooses to participate in the program,
and yi ¼ 0 otherwise. The probability of participation of respondent
i, Prðyi ¼ 1Þ, is represented by the distribution function F evaluated
at Xib, where Xi is a vector of explanatory variables and b is a vector
of unknown parameters. The logit regression assumes a logit form