A method is described for increasing residual degrees of freedom in a community experiment without substantially increasing cost or difficulty by dividing experimental subunits into batches. Theoretical advantages of batch sampling are described and illustrated with data from the Pawtucket Heart Health Program. (SLD)

Two studies were conducted to determine the stability of self-assessment of evaluation performance and use by 276 evaluators of community mental health centers over a 6-week period. The test-retest reliability of two instruments developed to assess internal evaluation was examined. The value of these instruments is discussed. (SLD)

Articles in this special issue consider different aspects of community-level evaluation research and demonstrate the emerging sophistication and promise of this field. Process evaluation, qualitative research, and meta-analysis are promising new directions for ensuring more appropriate levels of analysis and more appropriate outcomes. (SLD)

A methodology is presented for collection of cost data in evaluation of community-level health interventions that coincides with the intervention implementation. The seven discrete steps of the analysis are described, and the Minimal Contact Education for Cholesterol Change study is used as an example of how the strategy is used. (SLD)