Motivated beliefs and overconfidence in health insurance: a survey study with married individuals
How to cite
Zhang, Chang; Jia, Ruo; Zheng, Jiakun; Mo, Jiawei. 2023. "Motivated beliefs and overconfidence in health insurance: a survey study with married individuals." AEA RCT Registry. AEARCTR-0011139. https://doi.org/10.1257/rct.11139-2.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
We conduct a survey-based experiment to examine health insurance demand and overconfidence in health among married individuals. The aim is to provide direct evidence that individuals are overconfident about their own and their spouse's health risks and that they believe they are healthier than their spouses. Additionally, we will investigate how individuals update health-related information about themselves compared to their spouses. The questionnaire comprises six parts. The first module aims to collect basic demographic data from respondents, including their gender, number of children, and weekly exercise time, as well as the same information about their spouses. The second module aims to elicit respondents' perceptions of the health risks. We ask respondents to compare their own health risks with those of their spouses, their own health risks with those of average people in their age and gender group, and their spouses' health risks with those of average people in their spouses' age and gender group. The third module introduces a survey experiment on belief updating. The fourth module collects health information on both respondents and their spouses. The fifth module gathers data on respondents' education levels, lifestyle habits, risk preferences, and private insurance purchases. Finally, the sixth module addresses potential experimenter demand effects; and we inquire whether the respondents trust the cancer news presented in the survey experiment and the extent to which they find others' health information helpful in shaping their beliefs about their own and their spouses' health risks.
Intervention
We work on the following stages for both subjects and the spouses of the subjects.
In the first stage, we randomly selected 9 people at the subjects’ gender and age from previous surveys. And ask the subjects’, "Of the 10 people, including you, what do you think the chance is that your odds of getting cancer is higher than half of them (5 selected people)?"
In the second stage, we then present a piece of health-related news to the subjects. Recent studies published in top international oncology journals have shown that increasing physical activity can significantly lower the risk of developing many types of cancer. Specifically, engaging in at least 3 hours of moderate-intensity exercise per week has been shown to reduce the incidence of liver cancer by 27%, lung cancer by 26%, and breast cancer by 10%. Afterward, we asked subjects to choose one of the previously selected 9 people and informed them about how much that person exercises each week. We then asked for their perceived chance that their odds of getting cancer is higher than half of the 10 people again.
The updating order of self and spouse is randomized but consistent in the two stages.
Experimental design
To compare the perceived health risks, we inquire about individuals' perceived likelihood of being hospitalized due to illness in the next year, in comparison to people of their age and gender, for both themselves and their spouses. We expect that on average, individuals believe that both themselves and their spouses are healthier than the general population with the same age and gender.
To compare perceived health risks, we also inquire about participants' perceived cancer risk and that of their spouses. Specifically, we ask about the likelihood that participants think they or their spouses are healthier than a half of selected people of their age and gender in terms of cancer risk. We anticipate that, on average, individuals will perceive themselves and their spouses to be healthier than a half of people in terms of cancer risk. We then demonstrate the high correlation between the two measures of overconfidence.
To compare individuals with their spouses, we ask participants, "Compared to your spouse, how likely do you think you are to be hospitalized due to illness in the next year?" We expect to find that, on average, individuals perceive themselves as healthier than their spouses.
To capture the information updating process, we present a piece of health-related news. Before and after sharing the news, we ask participants about their perceived chance that their and their spouses odds of getting cancer is higher than a half of the 10 people. We then compare the participants' response to the news when estimating their own and their spouses’ cancer risk.
Design details
Cited by
No citation data available for this registration yet.