Doing the difficult and dangerous: the community program practice scale.

The CPPS uses staff respondents to portray practices and program climate of nonresidential mental health programs. We report psychometric analyses of 1,533 respondents in 165 programs. Confirmatory factor and partial credit analyses identified eight practice and five climate subscales, all of which show adequate psychometric properties. Program types are distinguished better by practices (R (2) values .37 to .52) than by climate (R (2) values .09 to .23), as expected. Multiple discriminant analysis and K-means cluster analysis illustrate how well the CPPS distinguishes program differences. The CPPS offers a promising, economical approach to measuring program practices in clinical trials comparing service programs.