Triangulation in sociological research is the use of 3 or more contrasting methods in a study to produce 3 different sets or even types of data. Its purpose is to reduce the weight given to any individual set of results. Instead, results obtained by different methods are contrasted in order to learn more about their validity. Different types of triangulation are used to test validity based on different variables.
Data obtained through contrasting methods reinforce confidence if it confirms the original conclusion, and removes confidence if it opposes the original conclusion. The reason for using 3 different methods is to create the potential for a tiebreaker.
Most sociological methods contain biases which cannot be completely excluded. Even the data source may have an inherent bias. For example, people answering a telephone survey may give different responses from those answering a mail survey, an Internet survey, or in person.
These kinds of biases may cause an atypical outcome. While the researcher can try to guess what might cause an atypical result, he has no way of realizing that the outcome was atypical on the basis of a single data source. By using 3 different methods which compensate for each other's biases, it is much less likely that all 3 sets of results will show the same atypical data.
Even in an ideal study, a result is considered statistically accurate when there is less than a 5% probability of that result having occurred by chance. In fact, a small probability that the results occurred by chance cannot be eliminated in sociologcial research. However, the cumulative probability of a chance finding over 3 studies, where each study meets the 5% rule, is 0.0125% or less.
An immediate disadvantage of triangulation in sociological research is the cost. Multiple methodologies require larger budgets. Analyzing 3 sets of results is more time-consuming than analyzing a single set. Statistical comparison between the sets adds another layer of time and cost.
The source model for the term "triangulation" is also a problem for its use in sociology. In surveying, the second measurement is not used to verify the first, while in sociological research, verification of method is the only justification for using triangulation at all. In surveying, the correct position ies at the point of intersecting lines. In sociology, the map itself is in question, and its redrawing depends on experimental data.
This difference is between checking an answer against a known structure and producing an answer to examine an unknown structure. However, the metaphor of triangulating does not always make this clear to the researchers or to those reading the research. This is a particular source of tension between sociological researchers who focus on either qualitative or quantitative methods, especially where some sources of data have traditionally been accepted without proper testing.
While verification of a sociological method is always ongoing, conflicting results are sometimes inevitable. Trying to find the correct interpretation between these conflicting data sets could result in a neverending process of triangulation. If the researcher finds himself in this kind of loop, it could indicate that the wrong question is being asked in the first place.