Descriptive Research Methods (Ch.
12)
Case
Studies: Detailed analysis of
a single (or limited number) of people or events. Case studies are usually interesting because of the
unusualness of the case (Three Faces of Eve, Mind of a Mnemonist) and/or the detail
and apparent insightfulness of the conclusions drawn by the writer (e.g.,
Freud’s cases such as ‘Little Hans’). The major problem with case studies is
the problem of objectivity. The
person who is presenting the case usually has some theoretical orientation. It is acceptable for a theoretical
orientation to affect one’s interpretation of events. In a case study the theoretical
orientation can also lead to the selection of the facts to include in the
case. It is not surprising that
case studies often seem to provide very compelling evidence for a theory. (I discovered this when I tried to
provide alternative interpretations of classic cases described by Freud, Adler,
and Jung.)
Case
studies can therefore assist psychology by illustrating how a theory could be
applied to a person or events and by assisting with the development of
hypotheses for more systematic testing, e.g., Piaget’s case studies of
the cognitive development of his three children.
Observational
Research: Accounts of the
natural behavior of individuals or groups in some setting. Unless the observation is unobtrusive,
there may be some subject reactivity to being observed. This often decreases with time, a
process called habituation.
Observers cannot usually observe all behaviors all of the time. They may use a behavioral checklist and
may also use time sampling or event sampling procedures. It is important to assess observer bias
by the use of interobserver reliability.
Observational research may also pose ethical problems. These can arise when the behaviors
being observed are not public behaviors and when the observer joins a group in
order to observe the
members’ behavior – participant observation.
Survey
Research: Structured questions
to assess peoples beliefs, attitudes, and self-reports of behavior. If the researcher wishes to generalize
the responses to a population, it is important to have a representative
sample. Surveys that rely on
self-selection (respond if you are interested) produce non-generalizable
results. Surveys also provide
information for correlational research.
One can correlate responses to some questions (often demographic
questions) with responses to other questions (often attitudes or reports of
behavior). Survey question must be
clear and unambiguous. (See
Appendix B for a discussion of some issue in the framing of questions.) Even if the questions are unambiguous
and non-leading, people may display a social desirability bias and give
positive or socially acceptable and desirable answers. Survey methods include: (1) the
interview or face-to-face method which is generally viewed as the best method
for obtaining a high rate of responses but is also very costly; (2) phone
surveys, which are less expensive but have a higher non-response rate (which has
probably increased with caller ID); and (3) written or mail surveys, which are
least expensive but have a very high non-response rate. Follow-up messages can help increase
the response rate.
Archival
Research: Analysis of
pre-existing data or records.
E.g., studies of ‘lunar lunacy’ rely on records of
accidents, homicides, 911 calls, use of emergency rooms, etc. Archival research often involves content
analysis, a qualitative analysis of material. For example, one would use content analysis to determine
whether there had been an increase in the frequency with which women and
minorities were mentioned in US history books between 1920 and 2000. Some archival research is
quasi-experimental. An example of
this is Ulrich’s analysis of recovery from surgery in rooms with a view
of nature (trees) vs. rooms without such a view.