Jumping to conclusions on Porn and crime
It is very difficult (actually almost impossible) to infer much of anything from one piece of time-series data since so many other factors can be changing over time. For example, may be as women are more willing to report rapes, the costs of committing rapes has gone up, thus reducing the number of rapes. (This is the National Crime Victimization survey data that tries to get around the problem of figuring out rapes that are not reported to police.) The aging population also seems like a possible factor. In general though it seems like pornography was easier to find in the 1960s and 1970s than Reynolds seems to remember, but, in any case, if there were big geographic differences in the availability of porn, we should be able to track where rapes rates started falling first. I have no idea of what one would find, but Reynolds claim is very weak.
Thanks to Don Kates for alerting me to this post.
Update: Clayton Cramer has some useful points on this too. Clayton also points out that Reynolds didn't properly read the article that he was reading from and that the data started in 1979. The first year that the data could have shown a drop in rape was thus 1980.