3.Are there any extraneous variables that might have contaminated the
data? Maybe the sharing bears were so boring that the children who
watched them are falling asleep.4.Is there an observer effect that might be contaminating the data?
Maybe I’m more likely to classify the behaviors of the ninja video kids
as aggressive.Any or all of these questions might render your assertion that watch-
ing ninja videos “causes” violent behavior unreliable. Sociologists must
constantly be aware of possible traps and biases in their research—even
in a controlled experimental setting like this one.
One must also always be on guard against logical fallacies that can
lead you in the wrong direction. One problem is what is called the “com-
positional fallacy” in logic: comparing two groups that are different,
assuming they are the same, and drawing an inference between them. Even
if all members of category Aare also members of category Bdoesn’t nec-
essarily mean that all members of category Bare members of category A. In its clas-
sic formulation: Just because all members of the Mafia (A) are Italian (B) doesn’t mean
that all Italians (B) are members of the Mafia (A). Just because virtually all those
arrested for child sexual abuse are heterosexual men doesn’t mean that all heterosex-
ual men are child abusers.
Issues in Conducting Research
No research project involving human beings is
without controversy. Debates have always raged
about the validity of studies, and we often come
to believe that we can explain anything by statis-
tics. That may be true—that you can prove even
the most outrageously false things by the use of
statistical manipulations—but not all “proofs”
will be equally valid or hold up in the court of
review by other social scientists. Most sociologi-
cal research is published in academic or scholarly
journals—such as the American Sociological
Review, Social Problems, Social Forces, or the
American Journal of Sociology. The American
Sociological Association sponsors several “flag-
ship” journals and controls the selection of edi-
tors to ensure that the entire range of topics and
perspectives is covered. Each subfield of sociology
has its own journals, devoted to those specific
areas of research. In the sociology of gender
alone, for example, there are dozens of journals,
includingGender & SocietyorMen and Mas-
culinities, a scholarly journal that I edit.
In all such reputable journals, articles are sub-
ject to “peer review”—that is, each article is eval-
uated by a set of reviewers who are, themselves,
competent researchers in that field. Peer review
ISSUES IN CONDUCTING RESEARCH 129Where there are more storks, there are more
babies. That’s true! The higher the number
of storks in an area, the higher the
birthrate. Could it be that storks actually do
bring babies? Well, no. It turns out that
storks tend to inhabit rural areas, and rural
areas have higher birthrates than urban
areas. That is, an extraneous variable
(urban versus rural) is the variable that
connects those two causally unrelated
variables.Didyouknow
?©The New Yorker Collection, 1977. Joseph Mirachi, from cartoonbank.com. All Rights
Reserved. Reprinted with permission.