501: Applied Research Mixed Methods and Triangulation Pam Brown & Joanna Wiebe (Originally presented as Prezi slideshow, which doesn’t work on SlideShare.net)
Combines different methods - whether qualitative or quantitative - to enhance validity and reliability while allowing deeper interpretations of what you’re studying.
Studies that compare quantitative and qualitative observations at the conceptual or propositional level.
Central to the development and testing of theory.
Triangulation is NOT
3 points of view on a phenomenon
Intended to look like a triangle :)
Mixed methods (that is, it’s not about the mixing of the methods but of analyzing the data)
Describes how, along with the use of multiple approaches to a research question (mixed methods), the researcher can zero-in on the answers or information sought in the data.
Origins of triangulation in social research
1959: Campbell & Fiske
Refers to using more than one research method (e.g., in-person survey, phone survey) in measuring the same object of interest
Denzin (1978) Refers to using the same approach for different sets of data in order to verify or falsify generalizable trends detected in one data set Data triangulation
Denzin (1978) Refers to making use of different investigators with different backgrounds Investigator triangulation
Denzin (1978) Refers to using at least 2 of the other methods of triangulation in combination Multiple triangulation
Denzin marries methodological approaches - and their ontologies & epistemologies - in his understanding of triangulation.
Denzin’s goal: To create a complete picture after using mixed methods (after combining qual and quant)
The primary reason of triangulation is the recognition that data-set or investigator survey bias can be introduced by using only one research method.
Data & investigator triangulation are argued to overcome problems of bias and validity.
Assumption of triangulation
In social research, if one relies on a single piece of data, there’s the danger that undetected error in the data-production process may render the analysis incorrect.
If, on the other hand, diverse kinds of data lead to the same conclusion, you can be a little more confident in that conclusion. (If 2 points cross, you can be more confident that your boat’s going in the right direction.)
Different things to different people
Triangulation is regarded differently by adherents to various perspectives, which have their own ontological and epistemological approaches.
Positivists seek to remove bias
Interpretivists seek to increase validity
Positivists and interpretivists conceive and act on subjective data differently.
The deficiencies of any one method can be overcome by combining methods and thus capitalizing on their individual strengths.
Overcome bias issues (e.g., data bias, investigator bias)
ASSUMPTION: The more different the instruments/methods, data sets and investigators used in analyzing a specific problem, the greater will be the confidence in the final result
Keep in mind…
Theoretical triangulation does not necessarily reduce bias
Methodological triangulation does not necessarily increase validity
Triangulation isn’t meant to combine qualitative and quantitative data
Denzin’s approach had strong elements of a positivist frame of reference - assumes a single reality that multiple accounts can be mapped on to pinpoint “absolute truths”
The effectiveness of triangulation rests on the assumption that the methods or strategies used will not share the same biases.
INSIGHT: Think about the significance of the role of methodological perspectives in social research generally and in the use of triangulation in particular.
SUCCESSFUL CROSS-PHILOSOPHY TRIANGULATION IS NOT POSSIBLE
How triangulation SHOULD be
All measurement should be of the same kind and based on the same ontology and epistemology
Triangulation is possible only within paradigms; any effort to compare or integrate findings from different methods requires the prior adoption of one paradigm or another
It should describe diverse, severe quantitative tests (not qualitative)
Should you avoid triangulation, then?
The sociologist should not necessarily avoid generating data in multiple ways and analysing it; rather, the sociologist should just avoid the mistake of using data to adjudicate between accounts
You shouldn’t adopt a naively optimistic view that the aggregation of data from different sources will unproblematically add up to produce a more complete picture.
Remember this when using triangulation
What’s involved in triangulation is not just a matter of checking whether inferences are valid but of discovering which inferences are valid.
Don’t take any data at face value. And don’t regard some data as true and some as false.
Use multiple methods if your research requires it. And use data-analysis approaches like triangulation if it fits. But just use triangulation consistently within one perspective - without drawing from different methodological perspectives.
Towards a less problematic approach
Triangulation constitutes the first phase of analysis; it’s used to estimate the error inherent in quantitative measurements. It also serves as a building block for the incorporation of qualitative data. Sequential mixed method:
QUANT > TRIANGULATION > QUAL (to deepen the analysis of quantitative findings)