are not selected randomly and not representative of the larger population but
selected purposively—that is, each subject is selected precisely because he or she
possesses certain characteristics that are of interest to the researcher.
One problem with interview studies is not the size of the sample but the fact that
the sample is not a probability sample—that is, it is not a random sample, but rather
the sample is selectively drawn to make sure that specific characteristics are included
or excluded. Purposive samples do not allow sociologists to generalize about their
results as reliably as they can with random samples. However, they do enable
researchers to identify common themes in the data and can sensitize us to trends in
attitudes or behaviors among specifically targeted groups of people.
For example, let’s say you wanted to study feelings of guilt among new mothers,
to see how much these feelings were influenced by television shows and magazine arti-
cles that instruct women on how to be good mothers. It wouldn’t make much sense
to conduct a random sample, because you wouldn’t get enough new mothers in the
sample. You could use a “snowball” technique—asking one new mother to refer you
to others. Or you could draw a random sample from a nonrandom population—if,
for example, the manufacturers of baby foods could be persuaded to give you their
mailing lists of new mothers and you selected every hundredth name on the list. (We
discuss sampling further below.)
All the methods above involve actually interacting with real people—either in a
controlled environment or in their natural habitat. These methods give us a kind of
up-close and personal feel to the research, an intimate knowledge with fine nuance
and detail.
You know the old expression of being unable to see the forest for the trees. Field
methods such as ethnographies are often so focused on the minute patterns of leaves
and bark on an individual tree that they lose a sense of the shape and size of the for-
est. Because the researcher wants to understand broad patterns of behaviors and atti-
tudes, sociologists also use more quantitative methods involving our interaction not
with people but with data. Of course, these methods might reveal the larger patterns,
but it’s hard to make out the nuances and subtleties of the individual trees.Analysis of Quantitative Data
Quantitative data analysis involves the use of surveys and other instruments to under-
stand those larger patterns mentioned previously.Surveys.Surveysare the most common method that sociologists use to collect
information about attitudes and behaviors. For example, you might be interested in
how religion influences sexual behavior. A survey might be able to tell you whether
one’s religious beliefs influence whether or not an adolescent has had sex (it does),
or if a married person has committed adultery (it doesn’t). Or a survey might
address whether being a registered Republican or Democrat has any relationship to
the types of sports one likes to watch on television (it does).
To construct a survey, we first decide the sorts of questions we want to ask and
how best to ask them. While the simplest question would be a dichotomous question,
in which “yes” and “no” were the only choices, this form of question can provide
only limited information. For example, if you asked, “Do you believe that sex before
marriage is always wrong?” you might find out some distribution of moral beliefs,
but such answers would tell you little about how people usethat moral position,
whether they apply it to themselves or to others, and how they might deal with those
who transgress.118 CHAPTER 4HOW DO WE KNOW WHAT WE KNOW? THE METHODS OF THE SOCIOLOGIST