Page 52 - Essencials of Sociology

Research Methods (Designs) 25

of your sample. (Tables of random numbers are available in statistics
books and online, or they can be generated by a computer.)
A random sample will represent your target population fairly—in
this case, married women enrolled at your college. This means that you
will be able to generalize your findings to all the married women stu-
dents on your campus, even if they were not included in your sample.
What if you want to know only about certain subgroups, such as the
freshmen and seniors? You could use a stratified random sample. You
would need a list of the freshmen and senior married women. Then,
using random numbers, you would select a sample from each group.
This would allow you to generalize to all the freshmen and senior mar-
ried women at your college, but you would not be able to draw any
conclusions about the sophomores or juniors.

Asking Neutral Questions. After you have decided on your popula-
tion and sample, the next task is to make certain that your questions
are neutral. The questions must allow respondents, the people who
answer your questions, to express their own opinions. Otherwise, you
will end up with biased answers, which are worthless. For example, if
you were to ask, “Don’t you think that men who beat their wives should
go to prison?” you would be tilting the answer toward agreement with
a prison sentence. The Doonesbury cartoon on the next page illustrates
another blatant example of biased questions. For other examples of
flawed research, see the Down-to-Earth Sociology box on the next page.
Types of Questions. You must also decide whether to use closed-
or open-ended questions. Closed-ended questions are followed by a
list of possible answers. This format would work for questions about
someone’s age (possible ages would be listed), but not for many other If sociologists were to study land
items. For example, how could you list all the opinions that people hold about what diving on Pentecost Island in Vanuatu,
should be done to spouse abusers? The choices provided for closed-ended questions can they could use a variety of methods.
miss the respondent’s opinions. Based on what you have learned
in this chapter, how do you think
As Table 1.4 below illustrates, you can use open-ended questions, which allow peo-
this activity should be studied?
ple to answer in their own words. Although open-ended questions allow you to tap the Remember that there are both
full range of people’s opinions, they make it difficult to compare answers. For example, participants and observers.
how would you compare these answers to the question “Why do you think men abuse
their wives?”
“They’re sick.”
“I think they must have had problems with their mother.”
“We ought to string them up!”

sample the individuals intended
TABLE 1.4 Closed- and Open-Ended Questions to represent the population to be
studied

A. Closed-Ended Question B. Open-Ended Question random sample a sample in
which everyone in the target
Which of the following best fits your What do you think should be done population has the same chance of
idea of what should be done to some- to someone who has been convicted being included in the study
one who has been convicted of spouse of spouse abuse? stratified random sample a
abuse? sample from selected subgroups
1. Probation of the target population in which
2. Jail time everyone in those subgroups has
3. Community service an equal chance of being included
4. Counseling in the research
5. Divorce respondents people who respond
6. Nothing—It’s a family matter to a survey, either in interviews or
by self-administered questionnaires