[FONT=&] It has been found that there are significant differences in child homicide rates according to the state abortion policies that are employed (e.g. Sen et al. [2012, Social Science and Medicine] write that "between 1983 and 2002, the average increase in the number of homicide deaths for children under 5 years of age was 5.70 per state among states that implemented stricter abortion policies over that time, and 2.00 per state for states that did not"). To what extent can we explain these differences by how anti-abortion policy increases the chances of children born into high risk environments?[/FONT]
It isn't a statistical analysis, it's a single instance of a correlation. It could well be that a fuller analysis was in the report you referred to but it's not in the single sentence you quoted. You'd need to explain "stricter" abortion policies and look at other potential factors in child homicide rates through the long period being averaged across for a start.
The data analysis isn't that interesting (as its been confirmed elsewhere). I'm more interested in the possible explanations (which aren't necessarily palatable to either pro-life or pro-choice positions). However, happy to offer more info.. The authors look at 3 policies that restrict abortion access: public funding of abortion, mandatory delay laws, and parental involvement laws for minors seeking abortion. We're not referring to simple correlations. The paper uses multivariate regression analysis (i.e. a conditional maximum likelihood negative binomial model with state fixed effects). To test for robustness it measures state restrictions on abortion access with 3 different approaches: "In the first approach, the empirical model included a binary variable to indicate whether any of the abortion-restrictions under consideration were present in that state and year for that age-cohort. In the second approach, the binary variable was replaced by a restriction index, created by adding the number of restrictions present in that state and year for that age-cohort. This index ranged from 0 if no restrictions were present, up to 3 if all three restrictions were present (that is, no public funding, mandatory delays, and either parental-consent or parental-notification). In the third and final approach, the empirical model included four separate binary variables to respectively denote the presence of no public funding laws, mandatory delay laws, parental-consent laws, parental-notification laws. In this approach we also controlled for the number of neighboring states that did not have any form of parental involvement law." Numerous control variables are included: state poverty rate, the state unemployment rate, the rural residence rate, the state's apparent per capita alcohol consumption, Aid to Families with Dependent Children benefit levels for a family of three, adult homicide death rates etc.
If parents do not want a child yet can not get an abortion then it makes sense that the child would be more at risk.
We shouldn't underplay the numerous factors at play though. For example, there is some evidence (mainly from Medoff, but its disputed elsewhere) that state abortion restrictions increase the use of contraception. The overall effect on unwanted pregnancies could then well cancel out. And, even if we reject that, there are possible differences in abortion policy effects. For example, the authors refer to how parental-consent laws increase the number of single-parent households (and the problems that can then be associated with such family structures)
