This report examines in detail the cost-estimating model developed for a study of reading-program costs, undertaken as part of a large-scale evaluation of compensatory reading programs. Results of the cost-analysis were presented in a separate report. Here, the confidence which can be placed in the cost-assessment model is evaluated in terms of the probable effects of erroneous allocation rules, missing data, and measurement or data-collection errors. The range of probable error in the analyses is judged to be within an acceptable range for cost-estimation studies. A final chapter makes recommendations to improve future data-collection instruments in an attempt to reduce model sensitivity and measurement error. (AA)