>Yes, but like I said, even we assume that demon mothers used abortion more
>than they used all other forms of contraception combined (for which they
>give no evidence), the policy recommendation would be even starker:
>increasing contraception use among demon spawners would be a cheaper and
>easier way of achieving the same end. We would already be assuming unmet
>demand.
Michael, Of course S Levitt would make no policy recommendations given his prentensions as a positive scientist (by the way, is Stephen by any wild chance related to Norm of the Science Wars?). And I don't think we should accept this gift horse--that is, use their study to fight for full access for birth control.
First, the underlying theory of crime here is that bad mothers make criminals. That's the only way to support the counterfactual that their causal theory implies: that had these children been born, many of them would have been criminals (and the crime rate would have been up to 20% higher).
Second, there is every danger that this study will be read as argument that the best way to reduce crime is to kill off would be criminals young (or real young). This opens what Troy Duster has called the backdoor to eugenics. Aside from insidious uses of genetic screening, there is the risk of the spread of family caps and compulsory sterilization for welfare receipt. Levitt may argue that the public interest in crime reduction does not justify coercive abortions or birth control but others will surely weigh things differently It is not difficult to guess at the demagogic uses that this study could be put to in the context of crime escalation in a downturn.
We should not fight for reproductive freedom on the grounds that unwanted (or poor or minority) children can only become criminals in our society so it's best to give bad mothers every chance to abort them. That many of these children may well have ended up in the criminal underclass is not an indictment of their mothers or an argument for abortion; it's an indictment of the underlying socio economic machine, as I have been trying to suggest from the outset.
I am interested why Yoshie or Kelley have not commented on this study.
Yours, Rakesh