Triangulation

Triangulation is typically perceived to be a strategy for improving the validity of evaluation findings; a strategy that will aid in the elimination of bias and allow the dismissal of rival alternative explanations of conclusions and propositions. Triangulation as a strategy in the social sciences dates to a 1959 paper by Campbell and Fiske in which they lay out the foundation for establishing the validity of measures through the application of a multitrait-multimethod matrix, a procedure involving both convergent and discriminant measures of traits. Eugene Webb coined the term, however, in 1966 in his now famous work, Unobtrusive Measures.

There are generally understood to be four types of triangulation:

1.
Data triangulation, including across time, space, and person, or using several data sources
2.
Investigator triangulation, or using more than one investigator
3.
Theory triangulation, or using more than one theoretical perspective
4.
Methodological triangulation, or using multiple methods

The basic underlying assumption in triangulation is that its use will lead to the elimination of error and bias, leaving nothing but the truth. This assumption is embedded within particular ontological and epistemological frameworks. Other conceptions of triangulation focus on an underlying assumption of clarification and elucidation as a result of triangulation. Thus, for example, different data sources or methods may lead to inconsistency and contradiction, but these are equally valuable understandings in rendering sensible accounts of what something means. Indeed, most evaluators understand that it is atypical for a single conclusion to emerge and that this does not preclude the making of evaluative claims or the drawing of evaluative conclusions.

Triangulation is often thought of as a technological solution for ensuring the validity of claims and conclusions, but it should be thought of more as a step in the process of embedding complex empirical data in a more holistic understanding of that specific situation. [Page 424]In the end, the evaluator still has responsibility for the construction of plausible and justifiable claims.