Complexity-Aware Methods

Author(s): MEASURE Evaluation

Year: 2017

Abstract:

Rigorous evaluations of health systems and services provide data that enable decision makers to improve programs. The question driving an evaluation can reveal if a health program or health system achieves its intended purpose and whether one approach works better than another. A typical evaluation design involves establishing intervention and control groups and randomizing subjects to the two groups, with baseline and follow-up quantitative surveys.

This and other evaluation designs carry us a long way toward answering questions about delivering effective health services. But these designs can’t answer all the important questions. Sometimes a comparable control group doesn’t exist, or randomization isn’t ethical. Sometimes there are too many variables, all affecting one another. Often, when an evaluation design won’t work, the problem is complexity—a dynamic that involves one or more of the following factors: many interacting variables; nonlinear chains of causation and unpredictability; and contextual factors such as cultural, political, structural, and environmental.

When we can’t use traditional evaluation designs and methods because of these factors, we need some other source of scientific guidance on which approach to take or which program to pursue. Methods that are “complexity-aware” enable us to address the inherent complexity in modern development programs, where environments are dynamic, multiple stakeholders intervene, and programs have many sectors of activity. This capacity statement examines ways that MEASURE Evaluation addresses complexity in designing evaluations that will yield reliable findings.