The story of the Toronto mayor’s being out-of-touch and your observation as to why the homeless rate in NY is low are fascinating. I appreciate the principle of triangulation that uses “different field strategies (e.g., interviews, focus groups, archives) as checks on each other to arrive at a singular truth” (Ettlinger, 2014, p. 593), the number of strategies being “two [sic] or more” (Cohen, Manion, & Morrison, 2011, p. 112). A simple example can be given. We are both English language oral testers. I observe and record Student A at 72%. Your score for Student A is 74%. This constitutes “investigator triangulation” (Cohen et al., 2011, p. 114). It would seem that the validity of our independent scores is high. In this case our methods were similar, but “the more the methods contrast with each other, the greater the researcher’s confidence” (p. 112) in the veracity or validity of the results.
Although the above account seems unproblematic, a few questions do arise. If our scores were markedly different, would that invalidate the test scores. or point to different grading criteria held by both markers? Suppose we were trained in how to grade according to a similar criteria, would that training invalidate the results by introducing a bias (i.e. towards the criteria)? And would the attempt to reduce incompatibility between graders equally reduce the validating potential of having triangulation in the first place?
Lund (2005) argues that there is no definitive qualitative-quantitative divide. The underlying constructs turn out to be the same irrespective of whichever paradigm is being used. If this monist perspective is correct, triangulation hardly matters. Taking the prototypical statistical significance p-value of 0.05, it may readily be seen that one test in 20 is likely to fall outside significance and the other 19 inside. If triangulation comprises the same underlying constructs, various perspectives will have a 19-chance-in-20 of achieving significance. (The assumption here is that the constructs have been operationalised accurately and that there is a ‘truth’ to be observed.)
Ettlinger (2014) demonstrates a pluralist perspective and accordingly claims that triangulated data sources are “incompatible with non-totalizing and non-essentialist principles” (p. 593). In other words, triangulation rests on “the assumption of a positivistic view that social reality is objective” (Hesse-Biber, 2010, p. 457). If Gray’s (2009) assessment is right, that the positivist ontology sees “The world [as being] external and objective”, whereas a phenomenologist perspectives has “The world [as being] socially constructed and subjective (p. 25), triangulations that combine these paradigms runs serious risks of introducing inconsistencies and paradoxes into the research design. That risk may be present even in my simple example.
Cohen, L., Manion, L., & Morrison, K. (2011). Research Methods in Education (7th ed.). London: Routledge.
Ettlinger, N. (2014). Delivering on Poststructural Ontologies : Epistemological Challenges and Strategies. ACME, 13(4), 589–598.
Gray, D. E. (2009). Theoretical perspectives and research methodologies. In Doing research in the real world. Los Angeles: Sage.
Hesse-Biber, S. (2010). Qualitative Approaches to Mixed Methods Practice. Qualitative Inquiry, 16(6), 455–468. http://doi.org/10.1177/1077800410364611
Lund, T. (2005). The Qualitative–Quantitative Distinction: Some comments. Scandinavian Journal of Educational Research, 49(2), 115–132. http://doi.org/10.1080/00313830500048790