Social Science and
the Problem of “Truth”
One thing that is certain about social life is that nothing is certain about social life.
Sociology is both a social science, sharing basic strategies and perspectives with the
natural sciences, and a socialscience, attempting to study living creatures who often
behave unpredictably and irrationally, for complex rational, emotional, or psycho-
logical reasons. Because a single “truth” is neither knowable nor even possible, social
scientists approach their research with the humility of the curious, but armed with a
vast array of techniques that can help them approach “truths.”
Even if truth is impossible, we can approach it. Like all other sciences, we approach
it through addressing two central concerns, predictability and causality. Predictability
refers to the ability to generate testable hypotheses from data and to “predict” the out-
comes of some phenomenon or event. Causalityrefers to the relationship of some vari-
able to the effects it produces. According to scientific requirements, a cause is termed
“necessary” when it always precedes an effect and “sufficient” when it initiates or
produces the effect.
Predictability and Probability
Everybody knows, for example, that Titanic(1997), with Leonardo DiCaprio and Kate
Winslet as passengers on the doomed ship, is a “chick flick”: Women love it, and men
don’t. But when I invite 300 women to a free screening, something remarkable hap-
pens: Only 80 percent of them love it. What’s wrong with the other 20 percent?
Auguste Comte (1798–1857), often considered the founder of sociology, actu-
ally founded something that he called “social physics.” He believed that human soci-
ety follows permanent, unchangeable laws, just as the natural world does. If they
know just two variables, temperature and air pressure, chemists can predict with 100
percent certainty whether a vial of H 2 O will be solid, liquid, or gas. In the same way,
social physicists would be able to predict with 100 percent certainty the behavior of
any human population at any time. Will the crowd outside the football game get vio-
lent? What political party will win the election? Will women like Titanic? The answer
should be merely a matter of analyzing variables.
For 50 years, sociologists analyzed variables. They made a lot of predictions. Some
were accurate, many not particularly accurate at all. It turns out that human popu-
lations have many more variables than the natural world. Yet predictability is of cen-
tral concern to sociologists because we hope that if we can understand the variations
of enough variables—like race, ethnicity, age, religion, region, and the like—we can
reasonably guess what you would be more likely to do in a particular situation. And
that—being able to use these variables to predict future behavior—is the essence of
predictability.
Why do 20 percent of the women in my study dislike Titanic? Maybe gender is
not the only variable that can be the cause of the desired effect. So I also ask their
age, race, socioeconomic class, and sexual orientation. Of women who are aged 18
to 25, White, middle class, and heterosexual, 95 percent like Titanic. But that still
leaves 5 percent who do not; I still can’t predict whether any particular woman will
likeTitanicwith 100 percent accuracy.
What other potential variables are there? Who knows? Maybe one woman doesn’t
likeTitanicbecause her uncle drowned, and the movie brings back unhappy memories.
SOCIAL SCIENCE AND THE PROBLEM OF “TRUTH” 127