by Phil Bartle, PhD
A variety of approaches
Usually we think of seven different kinds of research methods in the social sciences:
Let us look at these in turn:
Documents offer a wide range of material.
A major difficulty is that the research has already been done, and you can not go back in time and ask different questions.
A telephone book provides a small amount of material for a large number of individuals.
A diary provides detailed material for only one person and a few others observed by that person.
Newspapers provide good historical data, although they are collected for what will sell the paper rather than for their quality of observation and scientific interest.
Experiments are more popular with micro sociologists and psychologists.
They tend to rely on setting up artificial environments which in themselves may affect the responses.
Participant observation means living with the observed people, learning how they see things and experience things.
It requires much time and language ability, and might not be suitable for researchers of every personality type.
A 90 pound weakling [sic] may find it difficult to run with the Hell’s Angels, for example.
That was the main method I used when I studied the Kwawu people of West Africa, supplemented by most of the other methods.
Qualitative interviews, with open ended questions, allow more subjects being observed but not as many as in a survey, and allows more in-depth investigation than will a wide survey.
One problem with them is that answers from different respondent may not match up or cover the same material.
The amount of material from each respondent is quite unpredictable.
Secondary analysis, a favourite of armchair sociologists, takes material collected by other observers, for other purposes, and rearranges it and interprets it for new purposes and new theses.
Like documents, it is difficult to go back and ask different questions, and you must be satisfied with whatever you get.
Surveys, one of the most popular methods favoured by the numbers crunchers, allows a limited number of bits of information for a potentially large number of respondents.
It requires determining methods of sampling, and well designed, unambivalent questions, perhaps a pilot test, and a list of questions that is short as possible to ensure respondent consistency; that all questions are answered by every respondent. It is not a good approach for in-depth questions.
Unobtrusive measures can apply to most of the above except questionnaires, where the persons being observed do not know that they are being observed.
Watching where people sit on getting on a bus, or which urinal they choose in men’s rooms are classical studies using unobtrusive measures.
Any study of documents can be done without letting the observed know that they are being observed.
Good research applies several of these methods on the same topic.
If you copy text from this site, please acknowledge the author(s)