Explaining Attitudes Towards Obamacare: An Analysis of Factors that Impact the Formation of Public Opinion

            In March of 2010, President Barack Obama signed the Affordable Care Act (ACA), more popularly known as Obamacare, into law. The main aim of the legislation was to curb the growing costs of healthcare by forcing young people to purchase health insurance, to expand the public healthcare program Medicaid, and to create a more competitive marketplace for consumers (Wear, 2011). The act also banned insurance companies from denying coverage based on preexisting conditions. Proponents of the law claim that millions of people now have health insurance because of the ACA. According to the New York Times, “In 2016, there were 28.6 million Americans without health insurance, down from 48 million in 2010.”[1] However, opponents of the ACA argue that the law expanded the size of government and raised the cost of healthcare for millions of people who already had health insurance. According to Forbes, “…across the board, for all ages and family sizes, for HMO, PPO, and PSO plans, premium increases averaged 60 percent from 2013, the last year before ACA reforms took effect, to 2017.”[2] The enactment of the legislation has faced intense political opposition. Since the ACA was signed into law, various parts of the legislation have been struck down by the courts, and many states have refused to participate in the federal government’s expansion of Medicaid. It is even possible that the popular backlash against the ACA negatively affected Democrats in the run up to the 2016.[3] In light of these developments, what factors explain public opinion on Obamacare? Analyzing this research question will help us understand attitudes towards a controversial public policy and allow us to explore how public opinions are formed. Are citizens generally knowledgeable on public policy issues or are they largely apathetic and ignorant? Furthermore, what heuristic devises do citizens use when forming opinions on public policy?

            There are several theories that can help us answer these questions. First, past research has shown that public opinions and political behavior are affected by identity politics (Iyengar, Sood, and Lelkes, 2012). Since the 1970’s, America has witnessed the growth of a partisan divide between Republicans and Democrats. Politicians, media outlets, and citizens have increasingly become more divided by party (Druckman, Peterson, and Slothus, 2013). Politicians are more prone to voting along partisan lines with their respective party in Congress, and media outlets are more apt to frame the news to satisfy political partisans. Rank and file Democrats and Republicans are also more likely to favor a policy if it is supported by their own party regardless of personal socioeconomic status or ideology. According to the academic Stephen Wear, Obamacare was inspired by a health policy implemented by former Republican Governor of Massachusetts Mitt Romney, and Republicans are now only opposing the ACA so strongly because Democrats were responsible for enacting the law at the federal level. Past research has shown that many citizens will use their party identity as a heuristic device to make quick decisions in terms of voting and forming opinions on public policy. From this perspective, the great majority of citizens have neither the time nor the educational background to understand the complexity of public policy issues, but many will rely on political parties and other linkage institutions to make decisions for them.

            Furthermore, opinions on the ACA may also be effected by political ideology. Some citizens evaluate policy based not only on their party affiliation but a set of general beliefs about how the government should function (Jessee, 2009). Conservatives generally believe in small government, low taxes, and a minimal amount of regulation, and liberals generally believe in the necessity of a strong welfare state and higher taxes on the wealthy to pay for it. The enactment of the ACA included an expansion of Medicare and an increase of government regulation of the medical industry, so knowledgeable liberals and conservatives may have supported or opposed the law on this basis. While most citizens may have not understood the precise details of the law and how precisely it would impact individuals, partisans may have understood enough of the basics to form an opinion based on the ideological prism in which they view the world. Conservative and liberal media outlets have use simplified ideological arguments to criticize or defend the law. While past studies have shown that most citizens have very little knowledge of public policy, the opinions of political partisans who keep up with the news are influenced by their political ideology (Zaller, 1992). It should be noted that there is a significant amount of multicollinearity between party identification and ideological beliefs as Republicans tend to be conservative and Democrats tend to be liberal, although party partisans who follow the news less often and have lower levels of education are more like to have a political ideology that contradicts their party identification.  

            Opposition to Obamacare might also be a result of symbolic racism; in other words, conservative whites may oppose the enactment of welfare programs like the ACA because of its association with welfare for African Americans. While old fashioned or biological racism has declined sharply since the signing of the Civil Rights Act in 1965, a more subtle more of racial resentment called symbolic racism has become prevalent among white conservatives (Kinder and Sears, 1981; Nail, Helen, and Decker, 2003). Even though the great majority of whites no longer believe blacks are biologically inferior, a majority of conservative whites believe that socioeconomic problems in the black community are the result of cultural deficiencies. From this perspective, blacks only need to work harder and stop using the history of slavery and segregation as an excuse for their problems. For many conservative whites, poverty has become associated with race. Therefore, whites who hold this form of racial resentment will be less likely to support public programs intended to help the poor such as Obamacare. In addition, many conservatives negatively associate the ACA with Barack Obama, the country’s first African American president (Wear, 2011). In fact, the ACA’s nickname, Obamacare, was originally used by conservative pundits as a means to mock the law. Opinions on the law may be tied to the race of the former president. Past studies have shown that America has a weaker welfare state than Western Europe partially because of the existence of salient social cleavages (Alesina and Glaeser, 2004). Americans are less likely than Europeans to empathize with the poor because the lower classes in the United States are associated with minorities, and political entrepreneurs on the right use these social cleavages to their advantage. For these same reasons, blacks are more likely to support Obamacare than whites.

            In addition, opinions on Obamacare may also be influenced by the socioeconomic status of citizens; specifically, people suffering from economic or health problems may harbor negative attitudes towards the law. Obamacare was supposed to improve the cost of healthcare and increase access to it, but if individuals don’t see personal financial or health improvements, they may blame the ACA for their problems. Given the complexity of the law and its impact on the healthcare system, it is reasonable to assume that the average American would not understand the precise impact the law had on the price of health insurance or the quality of medical services while controlling for other factors. Instead, we would assume that people would associate their personal economic and health status with the efficacy of the law regardless of the existence of causation between the independent and dependent variables. According to the political scientist Achen Bartels, citizens tend to engage in what is called irrational retrospection (Bartels, 2016). If they perceive themselves to be worse off economically or if they read negative news stories prior to an election, they become more likely to vote against the incumbent or oppose the status quo even if it is irrational to blame the government for their problems. For example, Bartels found that citizens were likely to punish politicians for uncontrollable acts of nature like shark attacks and famines. We might assume the same would be true of how citizens evaluate Obamacare.

            Finally, ignorance of politics may explain why some people are more likely to have a neutral attitude on the ACA. The political scientist John Zaller has shown that independents are comprised mainly of people who are apathetic about politics and rarely follow the news (Zaller, 2012). Contrary to popular opinion, partisans tend to be more highly educated on politics and follow the news more frequently. When asked questions on public opinion polls, people who are ignorant of politics will either not give an answer or will be easily influenced by the wording of the question itself when stating their opinion. It is possible that people who don’t follow the news, regardless of demographics and party identification, are more likely to have a neutral opinion on the ACA. It is also possible that many respondents on a survey will give a random answer. Too embarrassed to admit that they know nothing about Obamacare, survey respondents might state they are for or against the law anyway. However, due to the salient nature of the healthcare debate in the news, it is possible that the majority of respondents have formed a basic opinion of the issue prior to being surveyed. Past studies have also shown that a majority of Americans have been able to form consistent and strong opinions on salient issues like abortion and gun rights.

            To test these theories, I will analyze data from the 2016 American National Election Survey. Since this survey was conducted two years after the law came into effect, it gives ample time for the enforcement of the law to directly affect peoples’ lives. To measure the dependent variable, which is attitudes toward Obamacare, the following survey question is used: “Do you favor or oppose the 2010 health care law?”[4] The wording of this question is neutral as it avoids using loaded terms like Obamacare or the Affordable Care Act. Zero is coded as being in favor, one as being neutral, and two as being against the law. Opinions on Obamacare are as follows:

Table 1: Opinion on Obamacare	Frequency	Percentage
Favor	1,578	36.98
Neutral	881	20.63
Oppose	1,808	42.33

As shown above, there are a significant number of responses for and against the law, which means there is sufficient variation on the dependent variable. Only four respondents declined to give an answer, so they were removed from the analysis.

            Numerous questions are utilized to operationalize the independent variables. For party identification, a dummy variable for Republican and Democrat is included with independent being the constant. To operationalize the independent variable political ideology, a question on the ANES is utilized that asked respondents to rank their ideology on a scale of one to seven with seven being the most conservative. To measure the independent variable symbolic racism, the following ANES survey question is used: “on a scale of one to seven, how much should the government help blacks?” One is labeled as “the government should help blacks,” and seven is labeled as “blacks should help themselves.” This is a common question that is found in past studies on symbolic racism, and answers to this question are highly correlated with other survey questions commonly used to measure the concept (Taber and Sears, 2005). Furthermore, dummy variables on race and ethnicity are included with separate coefficients for Blacks, Hispanics, and Asians, and whites are treated as the constant. To measure the variable economic status, a question from the ANES is used that asks respondents to state whether their economic status is better or worse from a year ago on a scale of one to five with five being much worse. Additionally, a question from the ANES is utilized that asks respondents to evaluate their health status on a scale of one to five with five being poor, and a dummy variable on whether or not people have health insurance is included. Finally, to measure political ignorance, a question is used that asks respondents to state how often they pay attention to politics, the news, and elections on a scale of one to five with five being never. Moreover, a dummy variable indicating whether or not a respondent has a bachelor’s degree or higher is included. The normal demographic control variables are also used in the model, which includes age, ethnicity, marital status, income, and gender.

            To conduct the analysis, I first used the ordered logit model, but after conducting the brant test, there was a problem with parallel regression. This was expected. The reason for this complication is that there are significant differences between people who were neutral on Obamacare and those who are either for or against the law as will be shown in the model below. The dependent variable cannot be treated as ordinal. As a result, I used the multinomial logit model with the baseline being those who oppose Obamacare. The baseline will be compared with people who favor Obamacare and respondents who are neutral on the public policy.  Using the multinomial logit model is a better choice since we can compare people who are neutral with those who have strong opinions on the ACA.

            It is my hypothesis that political party identification, political ideology, symbolic racism, and socioeconomic variables for health and economic status will be powerful explanatory variables to explain people who support and oppose Obamacare. In contrast, people who are ignorant of politics are far more likely to be neutral. Furthermore, weak partisans and independents are less likely to know anything about Obamacare and more likely to give random answers to a pollster’s question. Since the majority of Americans lack basic knowledge of important policy issues, we can expect a relatively weak pseudo R-Squared value for the model and weaker coefficients when comparing those who are neutral with those who oppose the law. The results of the analysis are below:

Opinions on Obamacare (Baseline: Opposes Obamacare)

Favors Obamacare				Odds Ratio
Republican	-0.75	-0.2	***	0.47
Democrat	1.21	-0.2	***	3.35
Ideology	-0.44	-0.1	***	0.64
Financial Status	-0.28	-0.1	***	0.75
Female	-0.21	-0.1		0.81
Health Insurance	1.02	-0.2	***	2.76
Poor Health	-0.13	-0.1	***	0.88
Age	0.02	0	***	1.02
Married	-0.11	-0.1		0.90
Education	0.11	-0.1		1.11
Income	-0.02	-0		0.98
Black	0.89	-0.3	***	2.43
Hispanic	0.29	-0.2		1.34
Asian	0.97	-0.3	**	2.64
Other race	-0.06	-0.3		0.94
Black Welfare	-0.27	-0	***	0.77
Economic Status	-0.43	-0.1	***	0.65
Political Ignorance	0.08	-0.1		1.09
Constant	3.32	-0.5		27.53
Neutral on Obamacare
Republican	-0.78	-0.2	***	0.46
Democrat	0.37	-0.2		1.45
Ideology	-0.24	-0.1	***	0.79
Financial Status	-0.07	-0.1		0.93
Female	-0.03	-0.1		0.97
Health Insurance	0.37	-0.2		1.45
Poor Health	0.01	-0.1		1.01
Age	0.01	0		1.01
Married	0.01	-0.1		1.01
Education	0.03	-0.1		1.03
Income	-0.04	-0	***	0.96
Black	0.85	-0.3	**	2.33
Hispanic	0.16	-0.2		1.18
Asian	0.57	-0.3		1.78
Other race	0.1	-0.3		1.11
Black Welfare	-0.09	-0	**	0.91
Economic Status	-0.28	-0.1	***	0.75
Political Ignorance	0.19	-0.1	***	1.21
Constant	1.39	-0.5		4.00
Pseudo R-Squared	0.285
Respondents	2807
*P<.05
**P<.01
***p<.001

            As predicted, the results show that partisanship, political ideology, symbolic racism, health status, and economic status are all statistically significant. Democrats are three times more likely than independents to favor Obamacare. Furthermore, moving up one point on the seven point ideological scale from liberal to conservative makes someone 1.56 times less likely to support Obamacare. Also, holding symbolically racist views makes someone .77 times less likely to support Obamacare, and being black makes someone 2.43 times more likely to support it. In addition, having health insurance makes someone 2.76 times more likely to support the ACA, and having poor health conditions and a self-reported decline in economic status is correlated with negative views of the law. Finally, political knowledge and education are not statistically significant when comparing respondents who favor Obamacare with those who oppose it. However, as expected, a lack of political knowledge is correlated with being neutral. Moving up one point in the scale of political ignorance makes someone 1.21 times more likely to be neutral. Moreover, almost all of the other coefficients in the neutral column are much smaller, and many of the independent variables are no longer statistically significant. The reason for this result is that people who are neutral are less likely to follow the news, and they have weaker partisan leanings regardless of their socioeconomic background and other factors. Being neutral on Obamacare is negatively correlated with being Republican, but it is not positively associated with being a Democrat. This indicates, as expected, that people who are neutral on Obamacare tend to be independents who do not follow the news often. Lastly, the weak r-squared of .28 indicates that over seventy percent of the variation in the dependent variable cannot be explained by the variables above. There is a high degree of randomness in people’s responses to the survey question. This may be the result of a majority of people not understanding the question. This is not surprising as past media reports have indicated that the majority of Americans are ignorant of health care policy in general.[5] The low pseudo r-squared could also partially be the result of independent variables that were not taken into account.

            The low pseudo r-squared could be a result of problems with the internal validity of this study. It would have been better to have included questions that asked respondents to state the economic and health impacts of the ACA on their lives, but there is no such question on the 2016 ANES. Self-reported health and economic status questions served as an imperfect substitute. It would have also been better to include questions that test the respondents’ knowledge of basic healthcare concepts and the ACA, but once again these questions are not found on the 2016 ANES. The questions on self-reported political knowledge and levels of education served as imperfect substitutes. Furthermore, it may have been better to have used a question that asked respondents to state their opinion on Obamacare on a scale of one to seven or to use a feeling thermometer, but given the large differences between people who are neutral and those who have strong opinions on the ACA, an ordered logit or ordinary least squares model would have not been appropriate. Finally, there are issues with external validity. Would we find similar results with public policy issues that are more or less salient than Obamacare?   

            In spite of these problems, the comprehensive number of independent variables in the model and the relatively low r-squared indicate that we can be sure of the finding that a large majority of people answering the ANES survey are ignorant of the details of the healthcare law and its impact on their lives. Furthermore, people who do follow the news regularly are more likely to use simplistic heuristic devices such as partisan political identity, political ideology, and symbolic racism to form their opinions on public policy issues. Only a small minority of people are judging Obamacare based on their personal economic and health status.

Work Cited

Alesina, Alberto, and Edward L. Glaeser. 2004. Fighting Poverty in the US and Europe: A World of Difference. Oxford: Oxford University Press.

Bartels, Achen. 2016.  Democracy for Realists.  Princeton University Press: Princeton.

DeSante, Christopher. 2013. “Working Twice as Hard to Get Half as Far: Race, Work Ethic, and America's Deserving Poor,” American Journal of Political Science, 57(2): 342-356.

Druckman, James, Erik Peterson, and Rune Slothus.  2013.  “How Elite Partisan Polarization Affects Public Opinion Formation,” American Political Science Review, 107: 57-79.

Iyengar, Shanto, Gaurav Sood, and Yphtach Lelkes.  2012.  “Affect, Not Ideology:  A Social Identity Perspective on Polarization,” Public Opinion Quarterly, 76: 405-431.

Jessee, Stephen.  2009.  "Spatial Voting in the 2004 Presidential Election," American Political Science Review, 103: 59-81.

Kinder, Donald and David Sears. 1981. “Prejudice and Politics: Symbolic Racism versus Racial Threats to the Good Life," Journal of Personality and Social Psychology, 40(3): 414-431.

Lau, Richard, David Andersen, and David Redlawsk.  2008.  “An Exploration of Correct Voting in Recent US Presidential Elections,” American Journal of Political Science, 52: 395-411.

Nail, Paul R., Helen C. Harton, and Brian P. Decker. 2003. "Political Orientation and Modern versus Aversive Racism: Tests of Dovidio and Gaertner's (1998) Integrated Model," Journal of Personality and Social Psychology, 84(4): 754.

Sniderman, Paul M., Thomas Piazza, Philip E. Tetlock, and Ann Kendrick. 1991. “The New Racism.” American Journal of Political Science, 35(2): 423-447.

Taber, Charles, and Milton Lodge. 2006. “Motivated Skepticism in the Evaluation of Political Beliefs,” American Journal of Political Science, 50:  755-769.

Tarman, Christopher and David Sears. 2005. “The Conceptualization and Measurement of Symbolic Racism,” The Journal of Politics, 67(3): 731–761.

 “User’s Guide and Codebook for the ANES 2016 Time Series Study,” http://www.electionstudies.org/studypages/anes_timeseries_2016/anes_timeseries_2016_userguidecodebook.pdf, 2016.

Wear, Stephen. 2011. “Sense and Nonsense in the Conservative Critique of Obamacare,” American Journal of Bioethics, 11(12): 17-20.

Zaller, John. 1992. The Nature and Origins of Mass Opinion. Cambridge University Press: Cambridge.

---

[1] “Nearly 20 million have gained health insurance since 2010,” https://www.nytimes.com/2017/05/22/health/obamacare-health-insurance-numbers-nchs.html, May 22^nd, 2017.

[2] Book, Robert. “Yes, it was the affordable care act that increased premiums,” https://www.forbes.com/sites/theapothecary/2017/03/22/yes-it-was-the-affordable-care-act-that-increased-premiums/#72c512b811d2, March 22, 2017.

[3] Gerson, Michael. “Obamacare’s effect on the 2016 election,” http://www.sandiegouniontribune.com/opinion/commentary/sd-oe-gerson-obamacare-20160919-story.html, September 20^, 2016.

[4] “User’s Guide and Codebook for the ANES 2016 Time Series Study,” http://www.electionstudies.org/studypages/anes_timeseries_2016/anes_timeseries_2016_userguidecodebook.pdf, 2016.

[5] Japsen, Bruche. “Americans don’t understand basic health terms, let alone Obamacare,” https://www.forbes.com/sites/brucejapsen/2016/03/13/americans-dont-understand-basic-health-terms-let-alone-obamacare-costs/#340ffcf37e62, March 13, 2016.