Keywords

JEL Classification

1 Introduction

Since it was first confirmed in December 2019 by Chinese health authorities,Footnote 1 COVID-19 has rapidly spread all over the world.Footnote 2 In March 2020, many European countries were locked down until May or June.Footnote 3 Similarly, in the US, President Trump declared a national emergency.Footnote 4 In this process, it has been revealed that elderlies and those with underlying illnesses tend to suffer from more severe illness.Footnote 5 Since then, medical research on COVID-19 has rapidly progressed, and various treatments have been developed; the quick recovery of President Trump from COVID-19 infection has been attributed to a new medicine treatment based on dexamethasone, Regeneron’s monoclonal antibody therapy, and remdesivir.Footnote 6 Despite all these efforts, however, 70.1 million people have been infected worldwide, with about 1.59 million dead as of December 12, 2020.Footnote 7

While medical research is a key to containing the COVID-19 outbreak, social science provides another key. An early study of Yano (2021) shows evidence supporting that political leadership and social learning matter in controlling the spread of COVID-19 infection. Focusing on the early outbreaks in Florida, he shows that the virus spread more in counties in which more voters support President Trump, who downplayed the danger of the virus, and that the spread is less serious in middle to small cities than in countryside. At the same time, many observers relate the increase in the number of suicides in Japan and Korea during the summer of 2020 to the fact that the economic and social state of women was most severely hit by COVID-19.Footnote 8

With these considerations, a group of social scientists at the Research Institute of Economy, Trade, and Industry (RIETI) and life scientists at the Graduate School of Medicine, Kyoto University, has initiated a socio-life scientific survey on COVID-19, targeting 3000 participants of the Nagahama Survey and 1000 medical workers at the Kyoto University Hospital. The survey is associated with the antibody test of COVID-19 administered to the respondents on the socio-life scientific survey. Although the survey can be adopted for various research theme, it is designed to highlight, in particular, behaviour change in the face of a pandemic in general, the effect on a system of values of the COVID-19 outbreak, and the actual spread of COVID-19 among the respondents.

The primary purpose of our survey is to find what may determine people’s propensity to behaviour change in the face of a pandemic. A more specific research questions could be: Is behavior change attributable to the state of health or to individual characteristics such as education, age, and the sensitivity to social pressures and to persuasion? Our experience shows that in order to prepare against the next pandemic, it is highly important to find what contributes to behaviour change. The outbreak of COVID-19 has shown that behaviour change is difficult not only for ordinary people but also political leaders. This fact was suggested as early as April 2020. Focusing on the 2016 US presidential election and the 2018 Florida gubernatorial election, Yano (2021) shows that the number of COVID-19 cases in April is positively correlated to the county-by-county number of votes received by the Republican candidate, Governor DeSantis, whose policy is closely in line with President Trump, heavily downplaying the potential danger of COVID-19. A similar relationship is not observed in Ohio, in which the Republican candidate, Governor Dewine, was elected in 2018; it has been known that Governor Dewine has kept a distance from President Trump. This finding is confirmed by the recent spread of COVID-19 in South Dakota and Vermont. A video posted by the Recount shows that since March 2020, the COVID-19 outbreak have followed completely different paths between South Dakota, in which the state governor downplayed the danger of COVID-19, and Vermont, in which the governor has taken a very careful approach, even though the two states are similar in size and in that both governors are Republican.Footnote 9 As of December 2, 2020, 80,912 people have contracted the virus with 948 deaths in South Dakota whereas 4239 have contracted it with 72 deaths in Vermont. By now, it is widely acknowledged that masks are important in dealing with the COVID-19 outbreak. Starting with the US, despite this, many countries are struggling to convince their people of the use of masks. This suggests how difficult a behaviour change is for some people, including leading politicians. These considerations are reflected in the design of our survey.

The secondary purpose of our survey is to look into the effect of the COVID-19 outbreak on the Japanese system of values. In order to compare different values across time and regions, Hirota et al. (2020) adopt what they call a system of values, which is defined as a functional relationship between an individual’s well-being and all the objective individual characteristics determining his well-being. Should, for example, the increase in the number of female suicides in Japan be in fact attributable to the COVID-19 outbreak, the change must somehow be reflected in the system of values. In order to capture such a change, panel data is necessary that makes it possible to compare people’s well-being before and after the outbreak. Our COVID-19 data is designed in such a way that it may constitute panel data with the socio-life science surveys that we have compiled in Nagahama since 2016.

Our survey is also designed to study the determinants of the COVID-19 outbreak itself. For that purpose, we conduct antibody tests for some 3000 Nagahama Survey participants and 1000 medical workers at the Kyoto University Hospital. In Nagahama city so far, 53 cases have been reported since March 2020, which is 0.045% of the Nagahama population; in contrast, in Kyoto city, 2018 cases have been reported, which is 0.14% of the Kyoto city population. At these ratios, we expect 1.4 cases either for the Nagahama respondents or for the Kyoto University medical worker respondents. Because our test is to detect the presence of COVID-19 antibodies, we may capture asymptomatic cases as well. Altogether, we hope to capture, for example, risk factors for each group of respondents.

In what follows, we explain the overall design of our survey in Sect. 2 and the survey questions and ideas behind them in Sect. 3. The actual survey questions and the options for answers are listed in the appendix, although its design differs from that of the actual survey, which was conducted online.

2 Overview of the Survey

Our survey questions can roughly be classified into two types. The first type consists of questions concerning what we want to explain by using our data (dependent variables). They are: those to capture behaviour change, those designed to measure well-being in the standard literature, and the results of COVID-19 and other antibody tests. The second type consists of those capturing factors that may explain a dependent variable. They are: basic individual characteristics such as age, family structure, etc., risk attitudes.

The survey covers two different groups of people. The first group consists of participants of the Nagahama Survey. The second group consists of medical workers at the Kyoto University Hospital.

2.1 Nagahama Group

As is discussed in Setoh and Matsuda (2021), the Nagahama Survey was started in 2007 to compile cohort data covering genomic and medical treats of people living in the city of Nagahama. In 2016, social scientific survey questions were added to capture the total aspect of human life. An individual’s medical history may affect his lifestyle and way of thinking at the same time that an individual’s economic status may influence his health. The current Nagahama Survey aims to capture this interaction between life scientific and social scientific aspects of life. As is discussed in Yano et al. (2021), Nagahama is a typical Japanese local city, in which the mobility of population is relatively low and in which traditional lifestyle is preserved.

As is discussed in Setoh and Matsuda (2021), the Nagahama Genome Cohort Survey was conducted in three waves. We conduct our survey online. The survey selects participants from the 10,115 people who participated in the second wave (2012–2016). We solicit participants who can participate in the online survey by announcing it through the local newspaper. Of these people, we randomly select 3000. In this selection, we exclude those who moved out of the Nagahama city, who refused the external management of medical examination information, those who withdrew their consent. After randomly dividing these 3000 participants into three groups of 1000 individuals, we conduct the questionnaire survey and the antibody test for each group in sequence in two- or three-week intervals.

2.2 Medical Workers Group

In contrast, medical workers are expected to be from many different parts of Japan, representing many different value systems. By investigating the determinants of behaviour change for this group of people, we may form a point of reference to make a comparison with the Nagahama group. Since the start of the outbreak, moreover, they have been exposed to the COVID-19 virus much more than those in the Nagahama group. By including the medical workers group, we may capture the relationship between behaviour change and COVID-19 much more clearly than in the Nagahama group.

Of the medical professionals working at Kyoto University Hospital, about 1000 people who agree with the purpose of this research and agree to participate are targeted. They are surveyed multiple times (about four times as a guide) to determine the presence or absence of COVID-19 antibodies, persistent infections, and reinfections. Multiple surveys will be conducted as follow-up surveys for the same subject.

2.3 Survey Method

Information on survey is delivered to potential participants, who are asked to express their willingness to participate through our website. The selected participants are asked to answer our survey questionnaire. To those who complete the questionnaire, we send equipment to collect saliva for the Nagahama group and blood for the medical workers group in order to conduct the antibody test. Survey participants will return the saliva sample in the enclosed return envelope according to the procedure manual. The test results are returned to participants in a secure manner by means of blockchain technology.

3 Survey Questions

Our survey questions can be divided into two types. The first type is primarily for dependent variables. The second type is primarily for explanatory variables. In addition, we ask a respondent’s personal views on various issues, which could be used as either dependent or explanatory variables. An English translation of the survey is set out in the appendix.

3.1 Primarily as Dependent Variables

3.1.1 Behaviour Change

As is discussed in the introduction, the primary purpose of the survey is to investigate what contributes to behaviour change. In order to capture the way in which people change their behaviour, we ask a number of questions.

In order to take into account the effect of the nationwide state of emergence starting on April 16, 2020, and being lifted at different time points in May in different prefectures, we divide the time period into five subperiods and ask respondents to answer the behaviour change related questions for each period. The five subperiods are (1) the period before the state of emergence (April 16), (2) the period from April 16 through the dates on which the state of emergency was lifted, which were May 14 in Shiga prefecture (where Nagahama is), and May 21 in Kyoto prefecture (where the Kyoto University Hospital is), (3) after the state of emergency was lifted through June 30 by when the outbreak first subsided, (4) from July 1, when the outbreak restarted through one month prior to the survey, and (5) for one month prior to the survey questions are answered. Note that, after July 1, the number of COVID cases hit a peak in August and stayed relatively low in September and October. It started to rise again after November. Most questions concerning behaviour change are asked for each of these periods.

  1. A.

    Changes in the frequency and ways in which daily activities are conducted

We ask if a respondent, facing the COVID-19 outbreak, has changed the frequency with which he/she engages in the following activities (increased/unchanged/decreased/stopped) and the precautions that a respondent took: general outings (Q8, Q9), outdoor exercise such as walking and jogging (Q11, Q12), drinking out and eating out with friends (Q14, Q15), non-urgent hospital visits (Q17, Q18), visits to parents (Q20, Q21), visits to karaoke, live houses, and game arcades (Q23, Q24), and shopping (Q26, Q27). The selections for precautions we offer in our survey vary according to the nature of activity.

We ask why a respondent chooses to, or not to, use a mask. As a reason why he/she uses a mask, a respondent is asked to rate how closely each of the following reasons holds true: “to prevent infection”, “to prevent infection to others”, and “because people around me wear them”, “because someone told me to wear it”, “because I would be criticized without a mask”, and “because I usually wear it”. As a reason why he/she does not use a mask, we present the following reasons. “Because I don’t think it works”, “because it was too expensive”, “because masks are unavailable”, “because it is uncomfortable”, “because it does not look good”, and “because it is unlikely that I will get infected”.

We also ask if the frequency of walking, biking, driving, and using taxis, buses and trains has changed compared to before the spread of COVID-19 infection.

  1. B.

    More careful daily conducts

We ask to what extent a respondent engages in activities to avoid infected. They are: washing hands, disinfect fingers, gargling, taking a temperature often, sunbathing, covering the mouth when coughing or sneezing, avoiding to touch the mouth or nose, not rubbing eyes, sleeping well, wearing a product that claims to disinfect the space, wearing a mask when going out, and taking enough space at the cash register. Although some of these are medically questionable, they are intentionally included to observe how much medical knowledge a respondent may have. For each of these items, we ask about the extent to which a respondent was concerned (not at all/not so much/to some extent/as much as possible).

  1. C.

    Travelling between badly and not badly affected areas

Whether or not one abstains from visiting areas in which COVID-19 spread badly may be an important measure to capture people’s willingness to change behaviour. With this consideration, we ask whether a respondent travelled to areas in which COVID-19 spread. That is, we ask whether a respondent has travelled or planned to travel to the eight prefectures in which COVID-19 badly spread (Tokyo, Hokkaido, Saitama, Chiba, Kanagawa, Osaka, Kyoto, and Hyogo); the options for an answer are: yes and no.

In order to assist the tourism industry, on July 22, 2020, the government launched the “Go To Travel” campaign under which the government subsidizes half of the domestic travel cost. We ask whether a respondent has travelled or planned to travel after July 22 by using the “Go To” subsidy. To those who answer yes, we ask if the respondent would travel or plan to travel even if the campaign did not start.

3.1.2 Well-Being

As is noted above, a system of values is defined as a functional relationship between people’s well-being and its determinants. In order to capture the way in which the COVID-19 outbreak affects the Japanese system of values, we ask questions focusing on people’s well-being, the dependent variable in a system of values.

As an index characterizing well-being, we adopt happiness and life-satisfaction. The question to measure happiness is: “How happy are you now?” We also ask about future happiness five years from now and life satisfaction. The question for life satisfaction is: “Overall, how satisfied are you with life as a whole these days?” These questions are recommended by the Organisation for Economic Co-operation and Development; see Annex B of OECD (2013). In addition, we ask about subjective health by the following question: “How would you rate your general health status?” Self-rated or self-assessed health that is asked by this type of questions has been widely used to measure an individual’s general health status and has been shown to be a powerful predictor of future morbidity and mortality.Footnote 10 A respondent is asked to rate the answers to the first two questions by 0 through 10 and the last question by five levels.

3.1.3 COVID-19 Antibody Test

As is noted above, in our research project, we offer respondents COVID-19 antibody tests to respondents. Its primary purpose is, obviously, to see how the virus is spread in a typical mid-size city (Nagahama) relative to a major city (Kyoto) and to medical workers in the major city. At the same time, we hope to attract more people to participate in our survey by offering an antibody test. Moreover, we may be able to explain the COVID-19 spread by means of various social scientific factors, including people’s readiness towards behaviour change. One caveat is, as is noted above, that the number of respondents who have COVID-19 antibody may be rather small.

3.2 Primarily as Explanatory Variables

We ask many types of questions to obtain different explanatory variables. They are basic characteristics of respondents, work, and economic status.

3.2.1 Basic Characteristics of Respondents

We ask about basic individual characteristics, including age, gender, and family structure. In addition to the basic family structure, we ask if a respondent has preschoolers and grandchildren living together. Moreover, in consideration of high COVID-19 risk groups, we ask if a respondent lives with people of age 65 and over and/or with underlying health problems (diabetes, heart failure, respiratory diseases, etc.). Moreover, we ask if a respondent lives with someone who needs dialysis or takes immunosuppressive or anti-cancer drugs. We also ask about basic characteristics on health and lifestyle, including subjective health, the K6 indexFootnote 11 measuring depression, drinking/smoking habits, and attitudes towards risks.

3.2.2 Work and Economic Status

We ask a respondent’s job status: The options for answer are: no jobs (meaning full-time housewives, students, and retired), employees, self-employed (operating restaurants, wholesale retailers, agriculture, etc.), professionals (doctors, lawyers, accountants, tax accountants, writers, etc.), employees of a family business (restaurants, wholesale retailers, agriculture, etc.), home-based work/internal work that has no employment relationship with a company, consignment/contract worker (without long-term employment relationship). Moreover, we ask if a respondent has a job that requires to make a direct contact with customers (floor staff, cash register staff, etc.) and if he has a job to provide face-to-face customer services at eateries (waiters, waitresses, etc.). We also ask the average number of people with whom a respondent makes direct contacts for colleagues, clients, and customers.

With respect to remote work and flexible time commuting, which has become popular due to the pandemic, we ask if a respondent is encouraged at the workplace (not encouraged at all/not so encouraged/neither encouraged nor discouraged/more or less encouraged/highly encouraged), how often the respondent takes up such an option (almost never/about one to two days a week/about three days a week/more than four days a week), and if he wants to have more or less opportunities for remote work/flexible time commuting.

We also ask about average weekly working hours, annual household income, and household wealth balance.

3.2.3 Risk Attitude

In the face of COVID-19, one’s readiness towards behaviour change may closely be related to one’s attitudes towards risk. With this consideration, we ask a couple of questions to quantify risk attitude.

The first question is as follows: “On the one hand, we have a saying that ‘nothing ventured, nothing gained.’ At the same time, we also have a saying that ‘a wise man keeps away from danger.’ Which do you think you are closer when you choose your own action?” A respondent is asked to identify which of these statements better describes their view. The other question is: “What is the probability of rain above which you take an umbrella when go out.” A respondent is asked to choose from 0, 10, 20 to 100%.Footnote 12

As related questions, we ask if a respondent agrees with the following statements. “In general, I feel very positive about myself”, “I’m always optimistic about my future”, “I am free to decide for myself how to live my life”, “I generally feel that what I do in my life is worthwhile”, “Most days I get a sense of accomplishment from what I do”, “When things go wrong in my life it generally takes me a long time to get back to normal”, “I try to live free from social constraints”.Footnote 13

3.2.4 Basic Health

State of health is also an important factor relating to behaviour change. In order to estimate a stable system of values, it may be important to control one’s physical health, which may affect one’s view on well-being temporarily.

With these considerations, we ask if a respondent has various symptoms, such as sore throat, fever, fatigue, and stomach ache. We also ask if a respondent suffer from various illness in 2020, including flu, COVID-19, ordinary cold, allergy, and asthma and if he has more basic health problems such as diabetes, high cholesterol, high blood pressure, respiratory disorders, etc. These questions also serve as explanatory variables for COVID-19 infection.

3.3 Additional Questions

In addition, we ask a respondent’s personal views on various aspects of life. Some of those questions are concerned with how people should live in the face of COVID-19. Others are concerned with one’s social capital. These questions can be, in and of themselves, something to be explained. At the same time, they can serve as explanatory (control) variables in explaining behaviour change and the system of values.

3.3.1 Personal Views on COVID-19

We ask a number of questions to capture a respondent’s views on COVID-19. First, we ask a respondent to evaluate the national and local governments’ dealings with COVID-19 outbreak. We also ask if a respondent thinks that the outbreak can be controlled if people simply follow governmental guidelines, which sources he relied on to obtain information on COVID-19 (friends/acquaintances, family, magazines/books, newspapers/TV, SNS, web information (news), web information (administrative agencies)), and the extent to which he rely on those sources.

We investigate how individuals view social restrictions on individual activities. That is, a respondent is asked to what extent he is uncomfortable with, or unhappy about, those who do not heed to social distancing in cases such as waiting to check out at a supermarket, whether he believes that those who caught COVID-19 can be socially criticized, and whether he supports making the use of a mask mandatory, prohibiting parties, and, more generally, restricting personal freedom.

We also ask if a respondent is usually vaccinated against the flu and if he wants to take a COVID-19 vaccine, if it becomes available.

3.3.2 General Temperament and Sociality

The survey includes some questions on general temperament and sociality. In relation to behaviour change, we ask if a respondent has a mental block against starting new activities such as visiting a new store and adopting new technology (for example, smartphone, online meeting). In addition, we ask if a respondent is reluctant to change his daily activities even if he knows he should change.

We measure social capital by means of standard questions recommended by the OECD (Scrivens and Smith 2013; Yodo and Yano 2017). On the trust one would have towards others in general, we ask if a respondent thinks that most people can be trusted or that it is more prudent to be careful in interacting with others. To measure social capital related to networks, we ask how reliable a respondent finds neighbours, relatives, and colleagues at work to confide daily problems and concerns, and the frequency with which he interacts with those group of people. We also ask the level of trust a respondent has in each of the national government, local governments, scientists such as infectious disease specialists, and news media such as broadcasters and newspapers. In addition, we ask about the donations that a respondent made in the last 12 months.