Search
Browse By Day
Browse By Time
Browse By Person
Browse By Mini-Conference
Browse By Division
Browse By Session or Event Type
Browse Sessions by Fields of Interest
Browse Papers by Fields of Interest
Search Tips
Conference
Location
About APSA
Personal Schedule
Change Preferences / Time Zone
Sign In
X (Twitter)
In American politics there is a longstanding tendency for public opinion to drift toward the views of the party out of power on various issues. Described by political scientist Christopher Wlezien as the thermostatic model, public opinion tends to shift away from the president when he begins moving the country one way.
Although existing studies of the thermostatic model (Wlezien 1995; Soroka and Wlezien 2009; Pacheco 2013) have demonstrated a clear, direct relationship characterized by the negative effect of policy change on public opinion, the optimistic conclusions stand in contrast to existing policy feedback literature which have generally failed to identify changes in mass preference as an outcome of policy change.
An alternate dynamic thermostatic model developed by Caughey and Warshaw (2020), argue that changes in mass policy preference is not necessarily the direct outcome of policy change itself, but is instead driven by the partisan control of government, which subsequently influences policy preference.
In this study, I examine whether changes in state abortion and gun control policies--electorally salient issues, produce a thermostatic shift in public support and electoral outcomes? I put together roughly five decades’ worth of public opinion data to identify whether new policies lead to changes in public support for abortion and gun control.
Findings from the aggregated analysis offer little to support the hypothesis that policy changes produce a thermostatic shift in public opinion. The absence of a feedback effect on relative preference as a result of states passing gun control policies, suggest that the thermostatic model is unable to explain the effects of social policy change. In addition, the null effects of abortion policies on abortion conservatism, an absolute measure of policy preference is also consistent with earlier findings from Caughey and Warshaw (2020).
As the aggregated measures of policy preference only indicate whether changes in public opinion are due to fewer citizens whose preferences are greater than the new policy, the same measures are unable to show whether changes in policy support is due to a shift in policy support at the individual level.
To identify whether the individual impact of policy change, I propose a novel survey experiment that is designed to isolate the effect of policy change on individual opinion. By manipulating policy changes at the state level, I am able to estimate the extent to which extreme policy changes reduce relative support for the policy in question. Besides relative support, I also examine the influence of policy change on absolute or dichotomous support. Additionally, I also analyze the heterogeneous effects of the thermostatic model among those with higher education levels, their partisan and ideological identification. Finally, I examine whether voters who have experienced extreme policy change are more likely to vote for politicians who are out of power.
I find that the impact on electoral outcomes is substantial and meaningful. On average, the contemporaneous impact on national and state politicians was a loss between 0.03 and 0.05 percentage points from the previous election, though national politicians more often experienced electoral loss than state politicians. Thermostatic electoral backlash due to abortion and gun control policies also usually lasted 1 to 2 election cycles after they were first implemented in states, though the pattern was more prevalent with abortion policies. Additionally, the thermostatic backlash was particularly obvious in midterm elections, where electoral losses were roughly 1 to 2 times greater than on-cycle presidential elections.
Overall, the results suggest that while a thermostatic model as characterized by a direct relationship between policy change and mass preference may not hold, a dynamic perspective of the model that incorporates the role of government partisanship may be necessary.