Is the internet causing half the rapes in Norway? I wanna see the scatterplot.

Ryan King writes:

This involves causal inference, hierarchical setup, small effect sizes (in absolute terms), and will doubtless be heavily reported in the media.

The article is by Manudeep Bhuller, Tarjei Havnes, Edwin Leuven, and Magne Mogstad and begins as follows:

Does internet use trigger sex crime? We use unique Norwegian data on crime and internet adoption to shed light on this question. A public program with limited funding rolled out broadband access points in 2000-2008, and provides plausibly exogenous variation in internet use. Our instrumental variables and fixed effect estimates show that internet use is associated with a substantial increase in reported incidences of rape and other sex crimes. We present a theoretical framework that highlights three mechanisms for how internet use may affect reported sex crime, namely a reporting effect, a matching effect on potential offenders and victims, and a direct effect on crime propensity. Our results indicate that the direct effect is non-negligible and positive, plausibly as a result of increased consumption of pornography.

How big is the effect?

A 1 percentage point increase in internet use is associated with an increase in rapes of 0.14 and an increase in child sex abuse of 0.16, per 100,000 inhabitants. The effects on the overall sex crime rate and rapes are precisely estimated, being statistically significant at the 5%-level.

How much is this? A 1% change in internet use seems uninteresting. Perhaps we should multiply by 50? A 50% increase in internet use (which is roughly what was observed in Norway, according to one of the graphs) is associated in their fitted model with an increase in rapes of 7 per 100,000 and in increase in child sex abuse of 8 per 100,000.

How big are these numbers? According to Figure 4, the overall rate of rape is about 15 per 100,000 and the overall rate of child sex abuse is about 25 per 100,000. So, unless I’m missing something here, the estimate is that about half the rapes and one-third of the child sex abuse are caused by the internet.

(Figures 2-4 confused me at first because the time axes are on different scales.)

I don’t have it in me to try to evaluate this paper in detail. But what I really want to see here is a scatterplot. I want to see what the data say: where is the leverage coming from? Presumably the sex crimes rates are increasing faster in certain cities than others, coinciding to some extent with internet coverage. I’d just like to see the basic pattern to understand where the conclusions are coming from. A scatterplot is not the end of the analysis but it would be a good beginning.

I’d also like to see a more direct analysis to go along with the instrumental variable model. You can think of the instrument has having a joint effect on the intermediate and final outcomes.

Finally, the seriousness of the topic aside, I can’t helped but be amused by the authors’ use of econ-jargon. As everybody knows, in an applied micro paper you can’t just run the regression and give your results, you have to start with a theoretical model. Which in this case begins with, “For a crime to happen, we need (at least) two individuals meeting each other.” What about burglary??

On the next page we see the delightful expression “dp/dporn,” which actually looks even better in real life because the d’s are those curly partial-derivative d’s and the first p is a rho. OK, here’s the full version; read it and weep:


Also, “most of what we know about dp/dporn is either circumstantial or anecdotal.”

All this theory doesn’t really hurt any, but I think it all comes down to the data analysis.

6 Responses to Is the internet causing half the rapes in Norway? I wanna see the scatterplot.

  1. Prison Rodeo May 17, 2011 at 3:29 pm #

    Pronounced “Dee Peedee Eye equals Dee Peedee Row Dee Row-dee Porn Dee Porndee Eye.”

    Sounds like the Swedish Chef.

  2. Andrew Gelman May 17, 2011 at 4:46 pm #

    The article explains that there’s more porn in Sweden than in Norway.

  3. Mark Rogers May 18, 2011 at 2:54 pm #

    If I was a policy maker I would want to see graphs showing an increase in crime that mirrored the broadband user and coverage rate graphs in Figure 2 (which is incorectly annotated). In fact a series of such graphs depicting the increase in crime by municipal area as broadband user and coverage areas spread across the country followed by an icrease in crime, would be better.

    Instead what I do see in figure 4 is that the velocity in the rate of crime increase actually decreased during the study period compared to the previous seven years possibly due to increased Internet use.

    1993-2000 per 100,000 people 2000-2008 per 100,000 people

    All sex crime rose 22 percent All sex crimes rose 17 percent
    Rapes rose 30.7 percent Rapes rose 31 percent
    Child abuse rose 20.7 percent. Child abuse rose 12 percent

    The introduction of broadband in this instance seems to me to have been beneficial contrary to the conclusions drawn by the authors.

  4. Mark Rogers May 18, 2011 at 3:16 pm #

    I tried to make a table showing the change in velocity of the rising crime rate with respect to the descriptive data found in Table 1 and Figure 4. It did not transfer well to the comments.

    Basicaly from 1993-2000 the crime rate per 100,000 people broke down as follows: all sex crime rose 22 percent, rapes rose 30.7 percent and child abuse rose 20.7 percent. In contrast, from 2000-2008 the crime rate per 100,000 people broke down as follows: all sex crime rose 17 percent, rapes rose 31 percent and child abuse rose 12 percent.

  5. naddah September 21, 2011 at 4:32 pm #

    There could be other causes. A disproportionate number of rapes in Norway are committed by Arab and other non-European immigrants. Maybe menacing immigrants are driving women indoors where they have nothing to do but surf the web.

  6. Mark Rogers September 22, 2011 at 3:04 pm #

    Naddah, just about all I know about Norway I learned from this paper. They do not break down who is committing rapes except by age. All of the socio-demographic variables they examined, unemployment, police density, poverty, urban settlement and education remained stable. Only immigration grew from a 3.9 percent population share to 6.3 percent. The paper did not indicate where these immigrants originated. So I would conclude that since the overall crime rate remained constant and the rate of increase in sexual crimes fell during the study period that could certainly be due to immigration and the Internet. Given that the authors conclude that a certain number of these sexual crimes are due to Internet usage I would further conclude that the benefits of immigration on Norway’s society are probably underestimated. With regard to your point about women staying home and surfing the Internet, the authors find this effect ambiguous as these women also would tend to use on line dating and matching services and meet criminals that way.