Definition
Triangulation is the practice of using multiple data sources, methods, or perspectives to cross-verify findings and strengthen the validity of evaluation conclusions. Rather than relying on a single piece of evidence, triangulation deliberately seeks converging lines of inquiry to confirm that observed patterns are real and not artifacts of a particular method or source. The term borrows from navigation and surveying, where multiple reference points are used to pinpoint a location with greater accuracy.
In M&E practice, triangulation typically involves combining qualitative data with quantitative data, drawing on multiple stakeholder perspectives, or applying different methods to investigate the same phenomenon. The goal is not simply to collect more data, but to collect different kinds of data that can speak to each other and either reinforce or challenge initial findings.
Why It Matters
Triangulation directly addresses one of the most persistent challenges in evaluation: establishing credibility. When a single method or data source supports a conclusion, stakeholders may reasonably question whether the finding reflects reality or is an artifact of how the data was collected. Triangulation builds confidence by demonstrating that multiple independent lines of inquiry point to the same conclusion.
Beyond credibility, triangulation serves a critical diagnostic function. When different data sources converge, evaluators can be more confident in their findings. When they diverge, this signals important questions about context, implementation, or measurement that warrant further investigation. Rather than masking complexity, good triangulation illuminates where and why different perspectives differ.
For practitioners, triangulation is a practical strategy for strengthening evaluation findings without necessarily increasing sample sizes or budgets. It is a core component of mixed-methods evaluation design and is particularly valuable when making claims about program impact or effectiveness.
In Practice
Triangulation appears in M&E work in several common forms:
Source triangulation involves gathering data from different stakeholder groups about the same phenomenon. For example, when evaluating a training program, you might compare participant self-reports of learning with supervisor observations of behavior change and objective performance metrics. Each source has different biases and blind spots; together they provide a more complete picture.
Method triangulation applies different data collection approaches to the same question. A typical pattern combines surveys (quantitative data) with focus group discussions or key informant interviews (qualitative data). The survey identifies patterns across a population; the qualitative methods explain why those patterns exist and what they mean in context.
Investigator triangulation brings multiple evaluators or analysts to bear on the same data, reducing the influence of individual bias. Theory triangulation examines findings through different analytical frameworks, testing whether conclusions hold across different interpretive lenses.
The key to effective triangulation is planning it into the evaluation design from the start, not attempting it as a post-hoc exercise. Data sources must be genuinely independent - two surveys from the same population using similar questions do not constitute triangulation. The most powerful triangulation combines methods with different strengths: breadth from quantitative approaches, depth from qualitative inquiry, and multiple perspectives from diverse stakeholders.
Proposal Context
Triangulation commitments in proposals signal rigor but need to be operationalized to be credible. Common proposal pitfalls: (a) claiming triangulation but sourcing all data from the same type (three household surveys do not triangulate), (b) no plan for what happens when triangulated sources diverge (the three sources agree, most of the time; when they do not, the methodology needs to specify how findings get adjudicated), (c) using the word "triangulation" as a general credibility signal without naming the specific triangulation type (data source, method, theory, investigator), (d) budget does not support multi-source data collection (triangulation costs more than single-source; this must be budgeted), (e) qualitative and quantitative used as labels rather than integrated in an actual mixed-methods triangulation design. A strong proposal names the specific type of triangulation, identifies the distinct sources or methods, and specifies how divergent findings will be resolved. Pair with mixed-methods and data-quality-assurance.
Related Topics
- Mixed Methods Evaluation: The broader design approach that enables triangulation
- Focus Group Discussions: Qualitative method often used in triangulation
- Key Informant Interviews: Qualitative method for multi-perspective data
- Qualitative Data: One dimension of triangulation
- Quantitative Data: Complementary dimension for triangulation