Performance evaluation should be viewed as an applied hybrid method distinct from program evaluation (defined by Peter Rossi) which combines methods from four different institutions each speaking a different policy language (Policy, Management, Programs and Accountability) as well as mixed methods. This includes two meta-data issues: Process Management ((the method of the “business process”) includes the management using participatory evaluation of a process-based chain of events leading to desired outcome(s)) and Research Metrics (a solid grounding in the “facts” about whether there is progress towards desired goals). Performance evaluation provides new ways of defining reliability and validity and the three types of performance evaluation inferences are reviewed: 1) relationships – how are two outcomes linked?; 2) outcomes – how good are the indicators of outcomes?; and 3) a process – what are the events and activities that produce outcomes? Each factor is critical for a robust, functioning performance evaluation system
AEA.Baer.Paper.Performance.Evaluation.as.Research.Method.October.20131.pdf