Probability vs. Certainty

By Karen Sternheimer

In a recent class discussion, we talked about the connection between children who have parents who are incarcerated and the likelihood of future incarceration for those children. One student had trouble understanding how all kids in this situation don’t end up in prison someday. After all, don’t we all just follow our parents’ examples?

Children with parents in prison do have a greater likelihood of getting arrested in the future, for a number of reasons beyond the scope of this post. The real issue that this student needed to understand was the concept of probability, or the notion of how likely an event is to occur.

Certainties are rare in science, especially in the social sciences, but even in the physical sciences it is difficult to know with certainty if one event predicts another. Some people who smoke cigarettes don’t get lung cancer, but the likelihood of contracting lung cancer rises tremendously with regular cigarette smoking. It is highly likely that students who don’t attend class or read the text won’t pass, but there is some variability there too.

Certainties are common in stereotyping and in making assumptions. Within stereotypes, we discount factors that contribute to the existence of certain behaviors and attribute them to group membership.

Understanding the distinctions between probability and certainty is one of the keys to developing a sociological imagination (and becoming an educated citizen, for that matter). One of the fascinating aspects of social science is using research tools to test assumptions through collected data—typically through multiple studies in a variety of settings.

The concept of probability is central to any statistics course. Through statistics, we can quantify the likelihood of something happening based on our data, and test whether it is the result of chance or if a true pattern exists between two or more variables. Tests of significance are based on the principle of probability, allowing us to measure whether the coexistence of two events is a coincidence or not. If something is statistically significant, it is unlikely to be the result of chance, and thus a real relationship exists between two events.

For example, one study tested the relationship between parents’ criminality and children’s. The likelihood of children’s criminal offending increased if parents had more convictions, and children with parents who were convicted after the child’s birth went on to have more convictions than those whose parents had convictions only before the child was born. The risk was most acute if the child was between 7 and 12 years old when their parents were convicted. The authors also note that specific risk factors, such as living in a violent neighborhood, poverty, and poor parenting practices are important factors in determining future criminality.

What does this teach us about probability? This study highlights the complexity through which events take place. It is also important to remember that people make choices in different ways, and in ways that are not always perfectly predictable. That’s why there are rarely perfect certainties within human behavior.

Even something that seems obvious on the surface—that children of parents with criminal pasts are more likely to be involved in the criminal justice system—requires deeper exploration. When is this relationship stronger? When is it weaker? Why is there a relationship at all?

Thinking about probabilities, rather than certainties, leads us to ask questions that help us understand sociological phenomena in much more depth than assumptions do. What other assumptions can be better understood as probabilities, rather than certainties?

Comments

[Yonason Goldson]

Thanks for the post.

FYI:
http://yonasongoldson.com/2015/06/05/are-you-too-sure-for-your-own-good/