Correlation is not causation, but...

[elfs]

Patterns of the spread of Internet access across the United States strongly correlate with a steep drop in rape cases in those states; states with the fastest Internet adoption show the greatest decline, and a 10% increase in household network penetration in any given region correlates to an average 7.3% decrease in rates of rape within that same region.

The results suggest that potential rapists perceive pornography as a substitute for rape. With the mass market introduction of the world wide web in the late 1990's, both pecuniary and non-pecuniary prices for pornography fell. The associated decline in rape illustrated in the analysis here is consistent with a theory, such as that in Posner (1994), in which pornography is a complement for masturbation or consensual sex, which are themselves substitutes for rape, making pornography a net substitute for rape.

Given the limitations of the data, policy prescriptions based on these results must be made with extreme care. Nevertheless, the results suggest that, in contrast to previous theories to the contrary, liberalization of pornography access may lead to declines in sexual victimization of women. The results suggest that the internet has had large effects on important social behaviors; further exploration of these effects is necessary to fully understand these results, however.

Kendall, Todd: Pornography, Rape, and the Internet, Clemson University Law & Economics Seminar, 2006