Alarming Study on Teen Dating Violence Is Flawed

By Benjamin Radford, LiveScience's Bad Science Columnist

A recent study made headlines across the country: The "Tween and Teen Dating Violence and Abuse Study" was commissioned by Liz Claiborne and the National Teen Dating Abuse Helpline. Among the headlines: 62 percent of 11- to 14-year-olds know someone in an abusive dating relationship, and one in five of those age 13 to 14 knows someone who has been struck in anger by a dating partner.

According to one reporter from a Las Vegas newspaper, "The image of the innocence of youth was shattered by the new study," which "found shocking horrors in teen dating."

In an effort to show just how shocking and unexpected the findings were—even to teens themselves—the reporter interviewed two teens. One high schooler, Ryan Sniezyk, said that he doesn't think that any of his friends are being abused. "I don't know anything about that," he said. "Maybe they are keeping it from me." Another young man agreed, saying that his experience didn't reflect the new study's findings.

The dating violence may be a hidden epidemic, or there may be another reason that the statistics are shocking and the teens don't know anything about it: They aren't accurate.

Parents may want to remove their fingers from the panic button and take a closer look at the study. Some of the most alarming statistics are misleading.

First, a quick quiz: Let's say you read a statistic from a study that says 75 percent of students at Harvard say they know someone who has cheated on a test. What does that mean? Does that mean that three-quarters of Harvard students are cheaters? Many people will read it that way, but they are wrong. In an extreme hypothetical explanation for how wrong this could be, it's possible that only one student at Harvard cheated, but everyone knew about him.

The teen dating study contained many questions asking the respondent if they knew other people who experienced certain events. For example, question 11 is: "Do you know anyone among your friends and people your age who have been called names, put down, or insulted?"

That's a simple, clear question that does not yield a simple, clear meaning because the answer tells us very little about the prevalence of abusive behavior. It doesn't take into account multiple reporting of the same incident among survey respondents. For example, let's say there's a fight at a high school and someone gets stabbed. If you later take a survey of students at the school and ask them if they know or heard about anyone who was stabbed, hundreds of people will say yes. But that doesn't mean that hundreds of people were stabbed, it just means that all of the people asked had heard about the one person who was attacked.

Many of the teen dating study's questions suffered exactly this problem.

What is needed are valid numbers on the number of people actually being abused, not percentages of people who have heard about others' abuse. There's also the problem of definitions. The study includes being called names or being put down as abuse. By this definition, if anyone you have been involved with has ever put you down or criticized you, you were in an abusive relationship. With such a broad definition, the high abuse rates found are hardly "shocking."

Statistics don't speak for themselves, they must be interpreted with caution. If you don't know what questions were asked, how they were phrased, or don't understand what the answers mean, the numbers are meaningless. There may indeed be "shocking horrors" in teen dating, but these particular statistics do not reflect them. Teen dating violence and domestic abuse are serious issues, and deserve both credible research methods and good journalism.

Original here