Multipaper Session 641: When a Triangle Becomes a Three-Sided Square: Measuring Students' Science, Technology, Engineering and Math (STEM) Interest and Learning through Data Triangulation Although triangulation or the use of multiple methods is widely considered to be good practice in evaluation, there is relatively little guidance in the evaluation literature about how to design a multi-method evaluation to triangulate data appropriately, or how to interpret incongruent findings. Measuring whether and how students' interest in science, technology, engineering and math (STEM) changes in response to program interventions can be particularly challenging, and triangulation is one approach to try improve the confidence and consistency of findings. In this paper, we describe how multiple methods were used in an evaluation of TechREACH, an out-of-school-time STEM program serving low-income and underrepresented middle school students in Washington State. We present examples of data about students’ STEM interests that were congruent (i.e., the data from all sources generally agreed in direction and magnitude), and examples of data that were incongruent (i.e., one or more sources did not agree with the others), and discuss how to interpret the lack of congruence. We recommend steps to avoid problems with incongruity prior to beginning data collection and methods to manage incongruity following data collection.#triangulation #afterschool #2009Conference #Prek-12EducationalEvaluation #technology #STEM #Evaluation2009 #NSF #ITEST