How did you conduct this research?
Jamieson: We have a long-standing, large-in-size content analysis project called CHAMP, which looks at health-risk behavior trends over time. A typical content analysis for many projects would study about 100 hours of media content, but we have 1,476 hours of TV coded. We go big, and we use a high-reliability standard. One of our earlier projects looked at gun violence in movies, and for this one we looked at gun violence in television, specifically dramas. Of the 33 shows we coded, about 60% are police dramas, with some medical dramas and other shows as well.
Why did you want to look at television gun violence this time?
Romer: Back in 2013, we published a piece in Pediatrics that first identified the fact that PG-13 movies had more gun violence in them than R-rated movies. When that came out, we got questioned by some who said, ‘Gun violence has gone down, especially in young people. Homicide rates have gone down and so as a result, gun violence rates have gone down. There’s very little reason to be worried about gun violence in the movies because it doesn’t relate to what’s happening in the real world.’ But given our previous work, we thought that if there’s a media influence from movies, it should be there in television as well.
For this research, how did you define ‘gun violence’?
Jamieson: It was a two-step process. We started off with a definition of ‘violence,’ which was more severe than aggression and included both intent to harm and involved contact. Then we coded each violent segment for ‘gun violence,’ which we defined as firing a gun and an animate being was hit.
Romer: Even if a segment had 100 instances of violence, we counted that only once. We wanted to keep it straightforward, ‘Did violence happen in a five-minute segment, period?’
Jamieson: Some projects like this have debated whether to code for verbal references of violence. Although it is a topic worthy of discussion because it does happen, we decided not to for this project. Once violence was coded in a five-minute segment, then we looked for things like presence of a gun or the firing of a gun. In other words, we coded for gun violence after we’d already found violence.
What did the research reveal?
Romer: We compared the homicide trend annually from 2000 to 2018 to the trends of violent segments with a gun in top-rated TV dramas, the ones that most people watch. We found a statistically reliable positive correlation between the two over that time period. It was the strongest for 15- to 24-year-olds. But it’s also there for 25- to 34-year-olds and those 35 and older. It’s true for all three age groups but strongest for the young people.
Jamieson: It’s not entirely surprising. Movie gun violence more than doubled in PG-13 movies between 1985 and 2015. So, it’s not entirely unexpected to see an increase in this violence on TV. We found the overall rate of TV gun violence doubled from 2000 to 2018.
Given these findings, what do you hope happens next?
Jamieson: In terms of the movie gun violence, in the past the Motion Picture Association has considered a PG-15 rating instead of a PG-13 rating. We think that should be a serious discussion. We also think TV gun violence matters because of the rate of media consumption. We see lots of exposure to TV, and we always have.
Romer: To that end, we think that CDC and NIH should start funding research to see whether there’s a causal relationship here. This was done back in the 2000s with respect to cigarette use in movies and television. The NIH spent quite a bit of money to look at that link, and they found that adolescents who watched movies with cigarette use were more likely to start smoking. There’s a tradition of looking at media influence on health-risk behavior influence, and you can do the same with guns. It’s a public health crisis, and there’s not enough money spent studying it.
