April 5th, 2006
And magazines. And music. And films.
That’s according to new research just out in Pediatrics journal (1), which claims that exposure to movies, music, television and magazines leads to teenagers having sex. It must be true because in the last couple of days has been covered in over 137 newspapers and websites worldwide.
Unfortunately none of those media outlets appeared to have read the original research paper. If they had, perhaps they’d be less eager to promote the study.
I’ve just finished reading the journal article and if it had been given to me in its current state to review for publication I’d have to respectfully tell the editor to reject it. It simply isn’t as good as it should be. And due to some major flaws in study design we cannot trust this research.
The research indicators were faulty
The study involved asking 1017 American teenagers complete a questionnaire on sexual behaviour and media consumption when they were aged between 12-14 and again two years later when they were aged 14-16. Unfortunately the researchers have failed to include a copy of their questionnaire within their journal article so it’s difficult to assess exactly what participants were asked.
Part of this research involved calculating the participants’ ‘Sex Media Diet’ or SMD. This invited participants to self-report their frequency of media use, and to indicate which specific media (magazines, music, television and films) they used regularly. This already introduces a problem – ‘frequently’ and ‘regularly’ are both terms best avoided in surveys because their interpretation varies widely between individuals.
Self-report studies in this area are also unreliable. Some teenagers in this study may have overemphasised some media use to appear more worldly wise, whilst others may downplay use for fear of being judged. Also it’s asking them to recall what movies, television, magazines and music they’d watched, read and listened to in the preceding month. Could you do that with any accuracy? I know I couldn’t. There are more appropriate and accurate ways to measure media consumption/exposure. Sadly these researchers didn’t use them.
It’s difficult to glean from the paper exactly what happened next, but it seems analysis of the questionnaires identified 264 magazines, television shows, music and films that were consistently mentioned. Of these 264 individual sources references or portrayals to puberty, romance, nudity, sexual innuendo, touching, kissing and penetrative sex were identified. And here’s where it gets even more confusing (and flawed), the researchers say within these 264 individual sources “each paragraph, headline and photograph in magazine; each lyric line in a song; each nonbreak sequence in a television show; and each nonbreak sequence in a movie was treated as a unit” (p1020). They identified 28000 separate units within these original 264 units.
Are you still with me?
The problems with this approach are numerous. For starters, it’s just a clumsy way of analysing media. It assumes that media can be broken down to discrete units. And it assumes each of those units are equal in value, weight, and interpretation. But critical media analysis shows this approach to be naïve. One striking song lyric could have more of an impact than a whole album of songs. This study also presents the units as independent. Research indicates there’s more impact when music and lyrics, or music and images, or text and images are combined, but again that doesn’t mean that images and text automatically have a greater impact when presented together than images or text individually. And research also shows that often people don’t process media in a line-by-line or scene-by-scene format. Nor are they often able to recall or even identify individual lyrics, scenes or images. The current study ignored all of this evidence.
Context wasn’t considered
The research didn’t ask about how the media identified in this study was used. If I ask you the question ‘how often do you watch television?’ it may seem simple. However what does that mean? It could mean having the TV on as background noise whilst you carry out other tasks or chat to friends or family. It could involve a TV playing in another room, or more than one TV in the house simultaneously tuned to different channels. Or it may mean sitting down glued to a particular show. Or none of those things. Over a period of a month the participants in this study would have consumed media in many different ways, and they wouldn’t have paid the same level of attention to television, movies, magazines and music all the time.
The study counted up possible exposure to media that participants were scored on, but it didn’t appear to ask people the way in which they interacted with or responded to different types of media.
The measures of sexual behaviour were incomplete
Aside from the clumsy measurement of media consumption, the sexual behaviour questionnaire was also inadequate. Female participants were asked to indicate ‘yes’ or ‘no’ if they’d ever “1. kissed a guy lightly on the lips, 2. kissed a guy by using your tongue, 3. had my breasts touched by a guy, 4. had my vagina touched by a guy, 5. had oral sex. Male response categories were equivalent but referred to relations with girls” (p.1021).
Well, where to start with these confused questions?
Firstly, the question asks have you ‘ever’ had the sexual experiences in the list above. This doesn’t differentiate between those who’ve tried one or more of the list just once, or those who are more sexually active. And since it doesn’t give a timeline we can’t tell if they’ve just had a sexual experience, had one at the start of the study, or for the full two years of the research have been engaging in sexual activity. Secondly, it sets up a sexual hierarchy, assuming that a ‘light kiss on the lips’ is of lesser value than having breasts touched or oral sex. But a light kiss on the lips could arguably be far more erotic and powerful than a clumsy boob grope.
Thirdly, the questions for girls (which are the only examples given in the paper) are confused in their delivery. The first two questions imply the girl is initiating sex – she’s asked to say if she’s kissed a guy (rather than been kissed), but the next two questions refer to the guy initiating breast or vaginal touching, whilst the final question ‘oral sex’ could apply to giving, receiving or sharing oral contact. The survey uses initiating sex as an indicator of sexual activity but fails in this objective.
Fourthly the questionnaire sets up the study to exclude teens that are lesbian, gay or bi. There’s no need to include gender in the questions – although the researchers stress the inclusion of gender somehow makes it ‘gender neutral’. What it actually does is set up questions that prevent certain teenagers from answering. The researchers claim their sample wasn’t large enough to find out about homosexual participants, unfortunately they’d lost those participants due to poor question phrasing.
Background evidence was used selectively
The researchers do cite evidence about media effects in the introduction of their paper. However, they only discuss research that promotes causal links between media and behaviour, and implies media sex content always causes negative sexual behaviours (e.g. early and unprotected sexual activity). There is evidence to this effect, but other studies suggest media can be a major and positive source of sex information, particularly in the developing world.
In fact it’s as though no research has happened in media research since about 1980. There’s no idea about critical media theories, how people engage with media, or how we negotiate our way through an increasingly mediated culture.
Social factors were overlooked
This study approached the media as though it has a direct and causal link with sexual behaviour. More sophisticated media research acknowledges this approach is simplistic and also allows for other factors to be included. So you may have a young person who does consume a lot of sex-related media, but they may also have a good relationship with friends and family and are empowered by school sex education so they delay sex until they’re ready and use condoms when they do have sex. Or you could have a young person who has little media exposure but little social support and education, so they end up having coercive and unprotected sex at a young age. The current research overlooked any of these possibilities.
Although this research paper contains several tables of data, and argues it’s proved that the more sexual media exposure you have the more likely you are to engage in sexual behaviour, in fact the study is so flawed the data collected is really meaningless, no matter how you report it.
In an editorial in the same journal (2) it’s stated “media research is not easy to perform, and this new study is ‘as good as it gets’” (p. 1428). Whilst it’s true media research isn’t easy, this study is nowhere near as good as it gets. And what’s truly worrying is the researchers have obviously not read the research that’s in the ‘good as it gets’ category. If they had, they may have produced a more thoughtful and useful study. Or more likely they’d have never embarked on this research at all.
Sexual behaviour is influenced by many factors. School and home sex education, family environment, peer support, income, health, education level, and self-esteem (to name a few). You can’t isolate any one of these and claim they alone cause or prevent sexual activity. It’s time we started to campaign for holistic approaches towards sex education, stop expecting the media to fill the information gaps that school and home sex education isn’t providing, and blaming it when young people have sex.
1. Brown,J.D., L’Engle,K.L., Pardun,C.J., Guo,G., Kenneavy,K. and Jackson,C. Sexy media matter: exposure to sexual content in music, movies, television and magazines predicts black and white adolescents’ sexual behaviour. Pediatrics. 2006; 117; 1018-1027.
2. Strasberger, V. “Clueless”: Why do pediatricians underestimate the media’s influence on children and adolescents? Pediatrics. 2006; 117; 1427-1431.