Inequality in Life Expectancy – (Mis)Perception and Policy Demand

Information provision to a randomly selected group of survey participants, revealing the true extent of inequality in life expectancy of their country.

2022-08-16

2022-11-30

Misperceptions of inequality in life expectancy, magnitude of and concerns about differences in life expectancy, policy preferences (general policy demand, health care access, economic stability, living conditions, education, working conditions), real outcome (donation to charities).

Misperceptions of inequality in life expectancy: We document the perceived difference in life expectancy between the bottom and top 10%, averaging across men and women. This outcome is descriptive and not affected by the treatment, but we expect that the effect of the information treatment depends on initial misperceptions. Therefore, we differentiate between participants who underestimate, accurately estimate, and overestimate inequality in life expectancy in the analysis.

Magnitude of and concerns about differences in life expectancy: We ask participants whether they think that socioeconomic differences in life expectancy are small or large and how concerned they are about socioeconomic differences in life expectancy. The concrete questions are: 1.) Do you think differences in life expectancy between rich and poor people are: very small; small; neither small nor large; large; very large. 2.) Do you think differences in life expectancy between rich and poor people are: not a problem at all; a small problem; a problem; a serious problem; a very serious problem.

Policy preferences: We elicit policy preferences in three steps. First, we ask participants about their general demand for policy action. The concrete question is: In your opinion, should the government do more to improve the life expectancy of poor people?
Second, we ask participants in an open question to state the measures they can think off that would improve life expectancy of poor people. Respondents should state the measures that they find most important. We use this information to examine the kind of policies that people are thinking of in general when it comes to the question how the government should address socioeconomic differences in life expectancy.
Third, we ask about support for specific policies related to the social determinants of health, as outlined by both the WHO Europe (https://health.gov/healthypeople/objectives-and-data/browse-objectives/economic-stability) and the US Department of Health and Human Services (https://health.gov/healthypeople/priority-areas/social-determinants-health). These policies cover the areas of health care access, economic stability, living conditions, education and working conditions of the poor. We introduce political consequentiality of the respondents’ answers by stating that the average level of approval to the policies is passed on to the politicians in their State Legislature / General Assembly after the survey has been completed. Besides studying support for each individual policy, we will also create indices for each policy area as well as an overall policy index from the five policy area indices. We also construct an additional index for health literacy policies, covering the expansion of education on health-conscious living (policy area education) and increasing taxes on unhealthy food, alcohol, and cigarettes (living conditions).
Real outcome (donation to charities): Respondents are enrolled in a lottery to win $500 in the United States and €500 in Germany. Before they know whether they have won or not, they have the option to donate to one or more charities. These charities help low-income people deal with the hurdles of everyday life. We use the probability of donating and the amount participants wish to donate as real measures of support for low-income households.

Concerns about income inequality: We ask participants how concerned they are about income inequality. The concrete question is: Do you think income differences between rich and poor people in the United States/Germany are: not a problem at all; a small problem; a problem; a serious problem; a very serious problem.
Perceived Covid-19 effect on differences in live expectancy: We ask participants about their beliefs to what extent socioeconomic differences in life expectancy have changed due to the Covid-19 pandemic.
Heterogeneity analysis: Depending on the distribution of prior perceptions, we may not observe an average treatment effect on outcomes. We therefore generally differentiate between the treatment effects for participants who overestimate, correctly estimate, and underestimate the true (gender-averaged) socioeconomic inequality in life expectancy. We also study how treatment effects vary with or are mediated by:
1. Certainty about perceptions
2. Beliefs about a person’s degree of control about one’s own life expectancy (measured with a luck-effort belief scale and with an item battery about causes of health problems)
3. Attitude towards government (measuring trust, ability, and scope of government).
4. Political orientation (measured once with a left/right scale and once with party affiliation)
5. Socioeconomic background (such as actual income, employment status, age, gender, migration background, education, social mobility, health insurance status, …)
6. Perceived own income position and beliefs about a person’s degree of control about one’s own income position
7. Responsibility of government and other actors
8. Own health and health behavior
9. Individual preferences on patience, altruism, and risk.
Mediation: To explain mechanisms of any potential treatment effects, we use a mediation analysis. The main potential mediator outcomes refer to the magnitude of and the concerns about socioeconomic differences in life expectancy. Moreover, we may also observe that the treatment affects luck/effort beliefs, trust in government, and the responsibility of the government for addressing inequalities.
Country comparison: We also study why (if at all) misperceptions and treatment effects on concerns and policy preferences differ between the United States and Germany. Potential explanations for any differential effects of information provision in the two countries may come from differences in the attitude towards government, luck/effort beliefs, political orientation, perceptions of health, health care access and income position, health behavior, responsibility of government, and other factors that differ between the two countries.

We conduct survey experiments in the United States and Germany. The experiment covers four stages: (1) Elicitation of prior perceptions of life expectancy inequality, (2) an information treatment to half the sample, (3) a post treatment survey on concerns, policy preferences, and background information, and (4) a follow-up survey.

We design and conduct online survey experiments in the United States and Germany with representative samples of the population between 18 and 70 years.
The experiment covers four stages. In the first stage, we ask all participants about their perceptions of the life expectancy of the bottom and top 10% in terms of household income, with half the sample randomly assigned to an incentivized belief elicitation.
In a second stage, we randomly assign half the sample an information treatment revealing the true extent in the difference in life expectancy between the bottom and top 10% in the income distribution.
In the third stage, we conduct a post-treatment survey to learn about participants’ concerns on socioeconomic inequality in life expectancy, their policy preferences on topics related to these differences in life expectancy (general policy demand, health care access, economic stability, living conditions, education, and working conditions). Participants then participate in a lottery with the chance to win $500/€500 (United States/Germany) and give them the opportunity to donate to charities supporting low-income individuals. The survey further covers participants' beliefs about the origins of differences in live expectancy, beliefs about the government and society, participant’s health and further socioeconomic characteristics.
In the fourth stage, two weeks after the main survey, participants are invited to an obfuscated follow-up survey, where we again ask about their perceptions of the socioeconomic inequality in life expectancy (this time directly asking about the difference in life expectancy between the bottom 10% and top 10%), their concerns and general policy preference. This allows us to study the persistence of any treatment effect and to consider the relevance of potential demand and attention effects of the initial survey.

Simple randomization, done by a computer.

The randomization is going to take place at the level of the individual survey participant.

No

n/a

12000 individuals in main survey. United States: 6000 individuals. Germany: 6000 individuals. Expected recontact rate in follow-up survey: 60% in both countries

3000 individuals per treatment for each country

With a Cohens D of 0.09, we should be able to detect a small effect size on concerns about differences in life expectancy.