With the U.S. Presidential Election campaigns gaining momentum, the important issues are becoming more and more clear. One such issue that might make or brake a candidate, is the stance they take on abortion. McCain and Palin strongly object against the possibility of women having an abortion, Obama and Biden take a pro-choice stance. More interestingly, though, Obama and Biden explicitly state on their site that it is their goal to protect the Roe vs. Wade ruling, whereas McCain-Palin explicitly state their goal to overrule Roe vs. Wade.
From a social science perspective, I think it is interesting to investigate these issues to a bit more depth, especially to see whether abortion has always been such a dividing line in American Politics.
Roe vs. Wade
The Roe vs. Wade U.S. Supreme Court Ruling states that almost all of the state-level bans on abortion are unconstitutional. Basically, due to Roe vs. Wade American women in all states have the possibility to elect having an abortion. This ruling has not been uncontested since. It is expected that the U.S. Supreme Court will have to decide on new abortion issues in the coming years, for some states (i.e. South Dakota and Mississippi) attempt to press legislation that bans abortion, in order to ultimately overrule Roe vs. Wade. Induced abortion is now an especially precarious issue during the election campaigns, for the next president will have to appoint a number of new judges to the Supreme Court. Will they appoint pro-life, or pro-choice judges?
Evolution of the Abortion Issue
In their study on “The Role of Part Activists in the Evolution of the Abortion Issue“, Carmines and Woods make use of two different surveys to investigate trends in attitudes towards induced abortion amongst the general public, the party activists, and the party elite. One of their more interesting findings (to me), is that between 1972 and 1980 people associating themselves with the Democrats were objecting against abortion more than people intending to vote for the Republicans. Only since 1992 a clear difference — in the way we see it today — emerged. This holds to an even larger degree amongst people who actively campaign for either of the parties. In contrast with these findings, the convention delegates have always differed strongly on the abortion issue: Democrats have always been more strongly pro-choice, and the Republican convention delegates more pro-life; nevertheless, the degree to which these delegates differed on this issue has increased strongly over the years. So, the authors conclude, when the abortion issue was introduced to the political domain, both parties had to find their position in the debate. Because the party elite already had formulated slightly different positions, this trickled down to the party campaign activists and finally to people who actively campaign for either party.
A methodological shortcoming
The theoretical interpretation of these findings is derived from the “Model of Partisan Change“, previously developed by the main author. This model basically states that “The issue evolution process … unfolds gradually and incrementally over an extended period, and causality runs predominantly from elites to masses rather than vice versa“. Although the authors indeed argue to show exactly this direction of causality, this also directly touches my main methodological critique of this study.
All in all, I’m not impressed by the methodological quality of this study. To name a few small points of critique: the way data is handled is unclear: the survey they used (the NES) changed the questionnaire half-way this study. Other studies have shown that this has impacted the measurement of abortion attitudes, but the authors do not mention that. In one table they take this into account, but another they don’t, without explicating which measurement exactly has been used. Not un-importantly: the most eminent change in public opinion occurred exactly in the year the measurements changed. Also, the main theoretical argument, that the party activists translate ‘issue cues’ from party elites to the masses, remains untested.
A much stronger critique relates to the supposed direction of causality, however. I think their data did not support an actual test of this hypothesis of ‘partisan change’, for they only had access to cross-sectional data. Their statements are based on the observation that the elites had their opinions formed earlier (and stronger) than the party activists and the ‘masses’ had. So, indeed, this leaves room for the corroboration of their hypothesis. However, what about selection? An alternative hypothesis, that would also be perfectly backup up with exactly the same observations, would be that once the party elites form an opinion, that people elect a political party to identify themselves with that corresponds to their own attitudes towards abortion. Seeing the increased (electoral) importance of certain issues, the party elites will profile themselves on these issues, gaining more votes from people that agree with them on that issue.
If this even sounds a little reasonable, the direction of the causality might very well be the other way around. With the observations done by the authors, this cannot be ruled out. Interestingly, this possibility of reversed causality is (unintentionally) shown by the analyses in the article itself: they show the outcomes of a regression analysis, in which a 7-point party identification scale (running from Democrat to Republican) is regressed on several attitudes, amongst which the attitude on abortion. The relationship between abortion attitudes people have and their party identification becomes stronger over the years, possibly indicating that people increasingly elect the party they identify themselves with based on their own opinion on abortion.
To me, this last analysis only shows that the crucial arguments of the study remain untested. We have learned that attitudes on induces abortion indeed have changed over the years and that the Democrats and Republican have shown increasingly opposing attitudes on this subject. Party elites appeared to have more strongly expressed opinions than party activists and the masses.
The authors argue that the party elites have influenced the other party members and thereby have shaped the present-day political debate on induced abortion. However, as I have argued, they do not (properly) test this claim and the direction of causality might just as well run the other way around as well. Perhaps, a properly designed panel study might shed some light on this still unanswered question …
In the meantime, I would just argue that when Obama and McCain would debate on the abortion issue, it would not be to change public opinion. Their sole purpose would be to gain as much voters who, in the heated abortion debate, have already determined their own position.
Edward G. Carmines, James Woods (2002). The Role of Party Activists in the Evolution of the Abortion Issue Political Behavior, 24 (4), 361-377