East Carolina University
Department of Psychology
Research Triangulation
You wish to measure the distance from your current location to a landmark across the field. You measure exactly the distance between two points at your current location (this defines the base line of a triangle) and then the angles between that base line and the two other sides of the triangle formed by drawing a line from each side of the base line to the distant landmark. Given the length of the base line and measurement of the two base angles, you can compute the distance to the landmark.
In a similar fashion a social scientist may better be able to describe/measure/manipulate/understand a concept if she can look at it from two (or more) different perspectives. If I reach essentially the same conclusion from a second perspective that I did from the first perspective, I likely will feel more comfortable with my conclusion, as if I have validated the first conclusion by checking from a different angle and seeing the same thing again. Such "research triangulation" may take many different forms in the social sciences. For example:
I want to study the effect of hunger on attitude to poodles. I decide to manipulate hunger experimentally. I could achieve such manipulation by depriving some of my human subjects of food for 24 hours while giving others ad lib access to food. Alternatively, I could attempt to induce hunger in some subjects by exposing them to the sight and smell of tasty pizza, while reducing hunger in others by exposing them to the sight and smell of vomitus. Or I could manipulate hunger by injecting certain substances in the body known to affect hunger or stimulating certain areas of the brain known to be involved in the control of hunger. If my research involved two or more of these methods of manipulating hunger, I would probably learn more about hunger than if it involved only one method.
I might also consider nonexperimental methods. For example, rather than trying to manipulate hunger, I could just ask people how hungry they are on a scale of 0-10. If I study hunger with both experimental and nonexperimental methods, I am likely to learn more about it than if I study it with only one or the other.
So, how am I going to measure attitude towards poodles? I could have subjects complete a survey in which they answer questions designed to measure attitudes towards poodles. Alternatively I could have trained observers watch subjects interacting with poodles and then assign poodle attitude scores derived from those behavioral observations. Or I could convene focus groups and ask them to discuss poodles. Or I could interview each subject and bring up the topic of poodles and record everything the subject says when talking about poodles. Or I could give the subjects poodle Rorschach tests or other projective tests to discover things about their relationships with poodles that they would be unlikely to reveal otherwise. If I use two or more different ways of measuring attitudes to poodles I am likely to learn more about such attitudes than if I use only one method.
Suppose I send out five research colleagues to observe the behavior of people interacting with poodles. From five observers I am likely to get a richer description of the observed behavior than if I use only one observer.
Suppose that instead of sending out five anthropologists to make the observations I send out one anthropologist, one nurse, one musician, one zoologist, and one social worker. These different observers are likely to interpret what they see from different theoretical perspectives, and that interpretative diversity may give me a richer understanding of what is going on out there in poodle land.
Suppose that I observe only persons at a meeting of the local poodle fanciers club. I might well get a biased picture of attitudes to poodles. Better to observe different sorts of persons at different places and different times.
By the way, in my experience, psychologists are unfamiliar with the term "triangulation" as used in this document, but they do address the some of same concerns that social scientists are addressing when they use this term. A Google search for "+triangulation +research" turned up many hits from various social sciences, but I did not find a single one from a psychologist (of course, I did not look at all 1,100,000 hits, just the first hundred or so). Here are links to a few of these hits:
Return to Wuensch's Research Design Lessons
