Can Television Get You Pregnant?
While science has not yet allowed electronics to impregnate us, this question circulated in the news the day before the presidential election on November 3. Lost in the frenzy of the election, it was easy to miss reports of a new study in the journal Pediatrics that found a relationship between teen pregnancy and watching sexually laden television shows.
News organizations like Time, CNN, MSNBC, and the Washington Post and others around the world covered this study, validating its importance and suggesting that sex on television is a central cause of teen pregnancy. But is it really?
Let’s start with the study itself. The researchers studied twelve- to seventeen-year-olds over three years. They conducted a survey by phone, starting with 2,003 respondents; by the end of three years, the sample size had shrunk to 1,461. (In long-term studies it isn’t uncommon for researchers to lose track of participants who may move and/or change phone numbers.)
The authors document that those most likely to have remained in the study were younger, female, and had parents with less education. Among black teens in particular, those who remained were more likely to have been sexually active at the start of the study. By the end of the study, 91 teens, or fourteen percent of the remaining sample became pregnant or had gotten someone pregnant.
In addition to gathering information about sexual behavior, the survey asked teens how often they watched one of 23 predetermined television shows that the researchers determined had high levels of sexual content (including flirting, touching, kissing, talk about sex, implied sex, and depictions of actual intercourse).
The authors’ initial findings suggest that the most important factors associated with teen pregnancy are being black, female, exhibiting problem behavior, and stating an intention to have a child before the age of 22. Living in a two-parent family was negatively associated with teen pregnancy; in other words, these teens were significantly less likely to become pregnant or to impregnate someone. Television viewing on its own was not statistically significant.
When the authors performed a multivariate analysis, or controlled for these other factors, television viewing became statistically significant—although the strongest predictors remained being black, female, not living in a two-parent household, and exhibiting problem behavior. Television viewing itself was actually negatively associated with pregnancy; the more television watched, the less likely a pregnancy would occur.
And yes, they found that watching the shows they deemed high in sexual content was statistically significant when controlling for the other factors, but it was less of a factor than those mentioned above.
If you have taken statistics, you might remember the term p value, which means the probability of this finding being the result of chance or an actual relationship between the concepts we are studying.
Generally speaking, social scientists want to make sure that our findings are unlikely to be the result of chance. If we want to be 99.9 percent certain, we use a p-value of .001 or lower, 99 percent certain means our p-value must be no greater than .01, and to be 95 percent certain our p-value cannot exceed .05. If we are not at least 95 percent certain that our findings are not the result of chance, we typically will concede our results are not statistically significant.
Now back to sex and TV. According to the Pediatrics study, we can be 99 percent certain that being black and exhibiting problem behavior is linked with teen pregnancy, and 99.9 percent that being female and not living in a two-parent family is also linked with teen pregnancy. The p-value for these variables is about .01 or less.
When we look at the television p-values, they are higher, which means the strength of their significance is lower. Total television exposure, a factor that predicts lower pregnancy likelihood, has a p-value of .02, meaning we can be 98 percent certain of this relationship. Age also has a p-value of .02, indicating that getting older is also a predictor of pregnancy.
And what of the smoking gun of sex on television, the finding that heard ‘round the world? It’s p-value is .03, the weakest of the statistically significant findings.
This does not mean these findings don’t matter, but the authors’ conclusion that the “results suggest that television may have a substantial role in the high rates of teenage pregnancy in the United States” is likely an overstatement.
The authors do concede that they cannot make any causal conclusions here; correlation is not the same as causation. Yet their concluding suggestions focus on television: they ask industry leaders to modify content, encourage pediatricians to inquire more about their patients’ media use and suggest parents to watch television with teens more often.
While it is very possible that teens who are sexually active might be more drawn to sexual media content than their abstinent peers, there are a number of unexplored issues here. First is the racial/ethnic disparity. According to the Centers for Disease Control and Prevention (CDC), in 2006 African American teens aged fifteen to nineteen had a birthrate of 63.7 per thousand, Latina teens 83.0 per thousand, Native American teens 54.7 per thousand, Asian American teens 16.7 per thousand and white teens 26.6 per thousand. Something is going on here, and it’s not just that the white and Asian American parents are watching television with their kids.
Teen pregnancy is highest in racial/ethnic groups that experience the highest levels of poverty. Yes, lower income people might also watch more television, but the central causal factor is not what teens see on TV. Lack of access to and information about birth control, lack of access to the educational and job opportunities that might be derailed by early pregnancy, as well as cultural factors come into play to explain teen pregnancy far better than what is on TV. For instance, a recent New Yorker article described how evangelical teens are actually more likely than other teens to be sexually active but less likely to use birth control. Sociologists have hypothesized that cultural attitudes condemning premarital sex might make these teens less likely to use birth control, which teens might interpret of proof of a plan to have sex rather than something that "just happened."
Teen birth rates are also much higher in the United States than they are in Europe, which has far more sexually explicit TV than we do here. But we have much higher poverty rates. If we truly want to reduce the number of teen pregnancies—and there are good reasons to try and do this—television is a tantalizing, but ultimately futile diversion. Studies like this make for great sound bites, but reducing teen pregnancy is far more complex than changing the channel.