Improving Data Quality in Large-Scale, Performance-Based Programs Evaluations

This article examines a systematic education & skill development strategy used for improving data quality in a large-scale, statewide program evaluation. This approach includes streamlining data collection procedures, operationalizing measures, developing & disseminating codebook information, & training those responsible for data entry. This data quality improvement strategy was examined in a process & outcome evaluation of North Carolina's statewide comprehensive tobacco control program. We calculated error rates in data (i.e., data did not match code book definition or construct) submitted by program coordinators before & after implementation, & we observed a 62% decrease in the total number of errors after the intervention was implemented (p = .014; 95% confidence interval [CI]: 17.6, 144.5). This article demonstrates that the intervention was effective in improving data quality, so that data entered matched operational definitions, thereby improving confidence in & validity of evaluation results. Adapted from the source document.