This literature review investigates thirty-one years of studies on thermostatic politics, a framework wherein political scientists trace how changes in policy affect public mood through the analogy of a thermostat. The development of this theory traces its origin from the foundational theoretical work of Wlezien (1994) to the various contemporary applications of the framework in other policy realms and polities.

This review considers three of the four major groups into which thermostatic studies can be categorised and collected. Firstly, there are foundational texts that define the thermostatic framework. Secondly, there are comparative applications in which the concept is applied to different contexts with the aim of determining whether the framework can be generalised. Thirdly, there are studies examining media and communications, which further our understanding of how the framework operates. The only group excluded from this study is the application of the thermostatic concept to elections, and this was done to allow this review to remain focused solely on the framework’s original intended purpose, namely public feedback to policy. Nevertheless, this area requires further study and deserves a review of its own.

Group One: Foundational Texts

Wlezien’s contributions, particularly his original 1995 article, appear as the cornerstone in understanding thermostatic mechanisms in politics. Through the lens of United States (US) defence spending, Wlezien illustrates a framework through which political scientists can understand how linear changes in policy impact public mood using the analogy of a thermostat.

A domestic thermostat registers the temperature of a home. When the temperature is too low, it calls for the heating system to produce more heat until the temperature rises sufficiently, and then it stops calling for heat. Equally, when a room is too warm, the thermostat calls for air conditioning to cool it down until the room is cold enough, and then it stops calling for cooling.

Wlezien (1995) argues that this same relationship is true of the public’s relationship with government policy, using the temperature as an analogy for the level of government spending and public opinion taking the place of the thermostat. He contends that “the public would behave like a thermostat; when the actual policy ‘temperature’ differs from the preferred policy temperature, the public would send a signal to adjust policy accordingly, and once sufficiently adjusted, the signal would stop” (Wlezien, 1995, p. 981). Wlezien (1995, p. 998) finds that, like a thermostat, when looking at the relationship between public opinion and defence spending in the US from 1973 to 1991, the “representation of preferences is not perfectly coordinated with the process of feedback.”

Wlezien’s (1995, 1996) original studies into this area test the hypothesis that if one combined an informed public, able to articulate their wishes, with electorally motivated politicians for whom responsiveness to public preference was a top priority, a thermostatic system would ensue. His 1996 article on the same topic concludes that “it is clear… that there are strong reciprocal connections over time between public inputs into the policy-making process and the outputs the process generates, at least in the defence spending domain” (Wlezien, 1996, p. 100).

Group Two: Comparative Applications

Following Wlezien’s initial formulation, a second group of scholars sought to test whether the thermostatic model could be applied to broader domains and political systems beyond US defence spending.

Pacheco’s (2013) article investigates thermostatic relationships between the public and US State legislators with specific regard to education and welfare spending, finding that “public opinion and policy exhibit a significant reactive relationship over time at the subnational level, at least in relation to education and welfare expenditures and in certain situations” (p. 326). However, Pacheco (2013) finds that representation is conditional on “the legislative professionalism of the states” (p. 326), meaning states lacking the ability to effectively measure public sentiment revert to using the national rather than state-level thermostats, thereby giving greater salience to the importance of public opinion data.

Anderson’s (2015) study stretches the thermostatic model further, broadening the application in terms of both polity and policy, by studying immigration policy in Sweden from 1990 to 2014. Anderson (2013) found that immigration in Sweden was not “as clean-cut as other policy domains” (pp. 27-29), with public responsiveness existing only within the ideological sub-groups on the left and centre-right, as these housed the main pockets of salience for immigration debates. Additionally, his research showed that changes in policy had no significant effect on the general public’s attitude toward immigration or the salience of the issue in public discourse. Anderson (2013) found the thermostatic properties to be limited in this context, which they explained as being due to “the unique nature of the immigration policy domain” (p. 26).

Group Three: Media and Communications

In contrast to the scholars previously discussed, the third group of scholars in this review explores the public’s informational requirements for thermostatic responsiveness from a new angle. Pacheco’s (2013) study investigates how well-informed politicians are of public opinion and the importance of this to a thermostatic response, whereas Atkinson’s (2011) article essentially studies the opposite, exploring how deep an understanding of government policy is needed amongst the public for a resultant thermostatic response. Atkinson (2011) poses that “public affairs journalists and editors are incentivised to focus on the partisan conflict and debate inherent in the legislative process,” suggesting therefore that  “the public [is] more aware of and more likely to respond to the contentious nature of the policymaking process than the substance of new laws” (p. 2). Atkinson (2011) tests this by examining thermostatic politics through the lens of American healthcare policy from 1980 to 2010, arguing that within the zeitgeist of thermostatic politics, the media has been viewed as a conveyor belt of information from policymaker to public, rather than as distinct “institutions comprised of strategic actors pursuing their goals” as it is conceived in the media sphere (p. 5). Atkinson then compares data from the Clinton and Obama administrations, both of which attempted to pass healthcare legislation, with only the Obama administration succeeding.

The article compares the decrease in the public appetite for healthcare spending predicted by the thermostatic model with the number of articles published on health policy and annual healthcare spending. Atkinson (2011) found that public mood was “somewhat responsive” to the actual passage of new policies, but a stronger correlation was visible between changes in public mood and the volume of news coverage a proposed policy received, regardless of whether the policy became law (pp. 22-23). This leads Atkinson (2011) to argue that if this could be generalised to other policy realms, then “public mood has more to do with the conflict inherent in how policies are made than with what policies are made,” meaning the thermostatic effect has been “misinterpreted by both policymakers and researchers” (p. 3).

Neuner, Soroka and Wlezien’s (2016) study revisits Wlezien’s original field of defence spending and thermostatic politics, focusing on the understanding that “thermostatic responsiveness does not require a high level of information” and investigates the extent to which the public is able to extract budgetary or policy change from mass news media (p. 6). They build on Soroka and Wlezien’s (2010) study, and argued that “in many domains people only need to have a sense for the direction of policy change” (Neuner, Soroka and Wlezien, 2016, p. 30).

Their conclusions challenge Atkinson’s ideas, arguing that the public is able to extract policy cues from mass media content, information which “matters for perceptions of government spending, and thus to preferences for policy change,” thereby suggesting a “very real possibility that the basic information needed for thermostatic responsiveness to function is readily available” (Neuner, Soroka & Wlezien, 2015, p. 28). However, they do concede that if focusing on “a low-salience domain in which a large proportion of policy change is regulatory rather than budgetary,” their study would have produced “much weaker results” (Neuner, Soroka & Wlezien, 2015, p. 30). Unlike Atkinson, they do not go so far as to question the accuracy of public perceptions, a qualm of increasing salience in the modern political-media sphere.

Conclusion

Over the last thirty-one years, the thermostatic public opinion feedback model has been stretched, developed and applied to numerous different political systems and policy domains, further refining the analogy and providing a fuller picture of its competencies and its limitations. Two areas stand out as worthy of further study, the second dependent on the results of the first.

Firstly, given the vastly different mass media and social media landscape we currently live in, the incorporation of more up-to-date studies looking at misinformation, media biases and the accuracy of public opinion on new policies would be invaluable for assessing the model’s current relevance. Furthermore, a comprehensive review of modern social media campaign communications requires a re-evaluation of the model’s application to elections more generally. Overall, the study of thermostatic politics offers an invaluable and accessible means for understanding the mechanisms which underpin the impact of public reflection on government policy, and it is clear that despite its humble beginnings discussing defence spending, the literature is still evolving to meet the mores of modern polities and policy domains.