Abortion: Evidence of an Issue Evolution
Abortion: Evidence of an Issue Evolution (DOI: https://doi.org/10.2307/2111673) was written by Greg D. Adams in 1997. It was published in the American Journal of Political Science (vol. 41, no. 3).
The author applies the issue evolution model to American politics and the abortion legalization issue
There are two analyses:
176 roll call votes in the House and Senate, collected between 1973 (Roe v Wade) and 1994.
In terms of the overall party, Democrats in both houses have trended towards legalization, while Republicans in the Senate have trended against legalization.
- In terms of individual legislators, no clear trend that people are changing their position.
- Implication is that change comes from "retirements and turnovers"
- Certain events are effectively weighted up because several roll call votes were triggered. There is no clear way to identify the duplicative/ repetitive votes, and the authors claim their inclusion does not affect conclusions.
GSS responses collected between 1972 and 1994.
- Note: interpolation used for 1979, 1981, 1986, 1992
- Proportions of categorical response to a standardized abortion question appear consistent.
- Author's interpretation is that opinions on abortion legalization are informed by morals, and are relatively unchanged over time.
Splitting by party identification and presidential vote (both self reported), author finds that Republicans were initially more propensed to support legalization. Especially when comparing the groups that reported voting for Humphrey vs Nixon in 1968. 1980s are an important inflection point for abortion legalization opinions and the parties.
- Author suggests that a lag between "changes in party identification" and "changes in voting behavior" explains differences between the twp splitting strategies.
Correlations between the two trends logically imply causation from congresspeople to the electorate. (Since individuals do not seem to change their opinions on the issue of abortion legalization, this should be interpretted as: changes of who is run in elections lead to changes of how voters identify politically) Democratic voters were relatively more anti-legalization than Democratic congresspeople until the 1980s. Causation in the opposite direction would have had the opposite outcome.
Reading notes
Author claims that responses to abortion questions are consistent over time. I'd like to see that claim tested, and more importantly I'd like to see it tested within important subpopulations that I expect are correlated (esp. sex and marital status). I also wonder if the GSS is calibrated; otherwise I'd like to see these tests with poststratification applied.