Relationship Between Difference of Motivation and Behavior Change Caused by Visualization

The purpose of this study is to verify the hypothesis that altruism makes others change their behavior. Recently, in economics, the heterogeneity of human beings is considered for modelling. In computer science, the agent-based model also accounts for the heterogeneity of human beings. Heterogeneity is being considered because traditional public policy does not work well as different motivations result in different behaviors. For instance, traditional economics assumes that in a model with representative rational human beings, when the public policy to levy tobacco tax is implemented, the consumer stops smoking. However, in our society, when tax is levied on tobacco, some people stop smoking, but others do (or can) not stop smoking. Thus, as the traditional economic model faces the abovementioned limitation, researchers need to consider various types of human beings. In behavioral economics, the government can create a policy to induce the desired direction while allowing other people to have a choice. This is called “libertarian paternalism.” In this example, levying a tobacco tax is a nudge.

This research compares two maps that elicit different motivations. One map focuses on selfish motivation through the desire for recognition, while the other map focuses on altruism. The common feature of the two maps is that they are made by the residents and shared among themselves. We developed an original application to enable this feature and to clarify that visualization of information by residents brings about indirect changes in their behavior.

The remainder of this paper is organized as follows. In Sect. 2, we refer to previous studies, and we explain our model in Sect. 3. Section 4 describes the system of collecting data and our original application. Section 5 describes the results, and we present our conclusion and remarks in Sect. 6.

In public economics, provision of public goods is an important task for the government. This is because without government intervention, households engage in utility maximization behavior, and public goods become under supplied. Private provision of public goods cannot be achieved due to the social dilemma. For example, if all the residents spend time only to do whatever they want, the utility of individuals will be maximized but no one will participate in volunteers such as regional traffic safety campaigns. The absence of traffic accidents in the area is desirable for the residents, but because traffic safety campaigns are insufficient only with police patrols, which are local public goods, traffic accidents increase and the utility of resident declines. However, Lazo et al. (1997) shows that residents supply local public goods based on selfish motivation caused by time inconsistency. Shiozu et al. (2017) shows that private provision of public goods is not successful even if there are leaders with ultra-altruism, under the condition that provision of public goods causes a financial burden on the residents.

On the other hand, Thaler and Sunstein (2003) states that households never engage in utility maximization behavior. The authors suggest that, while allowing residents to make self-decisions, the government can lead their behavior to the desired direction for the society through libertarian paternalism. Moreover, Thaler and Sunstein (2008) recommends the utilization of this nudge approach.

3.1 Concept for Behavior Change Through Visualization and Sharing of Information

We assume that people try to share information because of two incentives. One is the selfish incentive caused by the desire for recognition, and the other is the altruistic incentive to help people. Both incentives make a person upload his/her information. But self-satisfaction is obtained when others suggest sympathy by recommending a place as a good view. This is the reason why selfish motivation makes a person upload one’s information. On the other hand, sharing the information of dangerous places occurs because it seems that the information is also beneficial to others. If other people sympathize with this information, they will change their behavior. For example, someone shows agreement on a social networking service (SNS) or changes his/her behavior. Of course, these actions such as to express sympathize and change behavior may take place simultaneously. Figure 1 illustrates this concept.

Concept of our model

Full size image

3.2 Nudge for This Study

In this study, we designed a nudge for the study subjects by asking them to upload a picture and an icon on two maps with two different meanings. The first map is developed to determine a recommended course for a walk. We intend this meaning to capture the subjects’ need for approval. As the picture and icon on this map simply express the subjects’ feelings, they do not always make others sympathize. Thus, this map focuses on the subjects’ desire for approval rather than the desire for helping others. The second map involves safety for children. We intend this meaning to capture the subjects’ altruism. The pictures and icons on this map show places that are dangerous for children, such as areas with heavy traffic. The information is uploaded as a call to attention. If a subject agrees with the information, he/she may change his/her behavior to avoid the danger. Figure 2 shows an example of the icon and picture on a map. We performed image processing on Fig. 2 for privacy protection. “いいね!” means “nice!” in Japanese.

Example of the icon and picture on the map

Full size image

4.1 Overview of Survey District and Subjects from the Questionnaire

The survey district is located in the southern part of Kyoto prefecture in Japan. There are 20 subjects, whose average age is 70 years. We conducted the experiment from June 12, 2018, to August 31, 2018. Doshisha University conducted an ethical examination of the study before the experiment and provided approval.

First, we asked subjects about their cellphone usage through a questionnaire, and 18 persons replied. Of them, eight have their own smartphones, and 10 have feature phones. They use their cellphones mainly for making phone calls, and all of them make calls at least once per day. Most subjects, except for two, send SMSs. Over half of them do not use an SNS, such as LINE, Skype, or Twitter (see Fig. 3).

Usage of social networking services (SNSs)

Full size image

We asked the subjects about the type and frequency of their participation in volunteering activities. Figure 4 shows the types of volunteering activities they engaged in—five subjects do not participate in any volunteering activity. Figure 4 includes multiple answers.

Participation in volunteering activities

Full size image

In Fig. 5, the frequency tends to be of three types: several times per month, once or twice per week and three or more times per week.

Frequency of participation in volunteer activities

Full size image

From these results, we find that the subjects use cellphones for making calls and sending SMSs; however, they are not used to SNSs. In addition, over 70% of the subjects participate in some type of volunteering activity. As some of these activities, such as safety patrol for children, need frequent participation, subjects may engage in it three or more times per week.

4.2 System for the Study

To confirm that our model works, we develop a photo contribution system and collect global positioning system (GPS) data using a smartphone. In this section, we describe the system for collecting GPS data with a smartphone. Details of our original photo contribution system are presented in the next section.

We loaned smartphones to the subjects to collect their GPS data. The specifications of our devices and the network carrier are included in Table 1.

4.3 Application

We developed our own application called the “Community System Design” (CSD). This application shows the map of the survey area only. For this study, the subject can upload a picture taken by the smartphone camera on two maps with different purposes. Thus, in this experiment, we operate two maps: (1) map recommending the course for a walk and (2) a safety map for children. Figure 6 shows the user interface of this application.

User interface of the Community System Design application

Full size image

If no location data are available for the area, the picture cannot be uploaded. Subjects also choose a suitable icon from the list, which is uploaded with the picture. An example of an uploaded picture is shown in Fig. 2. Next, the picture and icon are shown where the pin is dropped on the map. These pins and pictures are shared among subjects. In addition, other subjects can express their approval about this picture if they tap the button called “いいね!.”

5.1 Overview of the Results

A total of 62 pictures were posted. Half of them are on the map of the recommend course for a walk. The rest of them are on the map of safety for children. Figure 7 shows the map of the recommend course for a walk^{Footnote 1}. This was especially significant as contributed pictures of the location of the subjects’ base got 13 approvals. Moreover, four places gained three approvals, but the remaining 27 places gained only 1 or no approval.

Map of the recommend course for a walk

Full size image

The map of safety for children is shown in Fig. 8. This map has two types of icons: one indicates a dangerous spot, and the other indicates a good place for children. A total of 30 pictures were posted on the map of safety for children. Three places gained approval as being good, 26 places gained approval as being dangerous, and one did not gain approval. Some of them gained multiple approvals; however, three is the maximum number of approvals. This is the difference between the map of the recommended course for a walk and the map of safety for children.

Map of safety for children

Full size image

Table 2 summarizes the results. More number of places gained two to three approvals on the map of safety for children than on the map of the recommended course for a walk. In addition, the map of the recommend course for a walk had multiple places that gained no approval.

5.2 Statistical Analysis of the Posted Picture and Behavior

In our system, we can obtain the time and GPS data of the posted picture. Using these data, we apply the before-after trial. If it can be confirmed that the subject visited the posted place through GPS data, the posted pictures may be considered to function as a nudge. As we assume in the model in Sect. 3, it is expected that the need for approval caused by selfish motivation makes others change their behavior less.

First, we found the places that were contributed on the maps. Then, we determined the frequency with which other subjects visited the same place before and after the contribution date based on GPS data. We considered that other subjects had visited a place when GPS data showed they were within a radius of 5 m from that place. Table 3 shows results for the map of the recommended course for a walk. There are 35 contributed pictures on this map; however, it was observed that other subjects had visited only five of those places. As the remaining 30 pictures had no visiting records, we omitted those results from Table 3. Place Nos. 1, 2, 3, and 5 indicate the base of the research collaborator, and No. 4 is the community center.

Table 4 shows the result of the safety map for children. There are 33 contributed pictures; however, it was observed that other subjects had visited only 12 of those places. As the remaining 21 pictures have no visiting records, we omitted those results from Table 4. Except for Place No. 6, no subject had visited these places before the pictures were contributed. After the pictures were contributed, some subjects visited those places. Someone visited Place Nos. 1, 2, 11, and 12 only once. The other places recorded multiple visits.

Comparison of Tables 3 and 4 show a large difference between in the number of visits. The subjects had visited all the places in Table 3 very frequently, because these places indicate the base of the Nonprofit Organization (NPO). Considering that these pictures were contributed relatively early and indicated the abovementioned base, it is conceivable that a large number of visits were recorded. The ratio of visits by other subjects to all the contributed pictures is 14.3% in Table 3 and 36.4% in Table 4.

To verify the hypothesis that altruism makes others to change their behavior, it is clarified by a t-test whether the number of visits of people changed before and after the subjects contributed the pictures. If the p-value exceeds 0.05, there is no difference between the visiting frequency before and after the picture’s contribution. As the map of the recommend course for a walk is based on the desire for recognition caused by a selfish motivation, the p-value is expected to exceed 0.05. Similarly, since it is assumed that the map of safety for children is based on an altruistic motivation, the p-value is expected to be less than 0.05. The data shown in Tables 3 and 4 and data on places that were not visited are used in the t-test.

Table 5 shows the result of the t-test. As the p-value of the recommend course for a walk exceeds 0.05, it means that the subjects did not change their behavior. On the other hand, because the p-value of safety for children is lower than 0.05, it supports the hypothesis.

This experiment suggests that posting pictures on the map of safety for children makes subjects sympathize with the information, because much higher number of places on this map gained approvals. Especially, dangerous places pointed out on the map and shared between the members gained multiple approvals. Before this trial, the subjects made similar maps with paper. Some places were pointed out on both the paper map and our digital map; however, some places were recognized as dangerous on our digital map for the first time. Considering the average age of the subjects was 70 years, posting or expressing approval on the map of safety for children is not likely to depend on selfish motivation but on altruism. In other words, participation in making this map captures the subjects’ altruism. In contrast, when making the map of the recommend course for a walk, the number of places that gained no approval suggests that, motivation is very important for solving local problems.

The result of the t-test shows that behavior could be changed if the visualization makes someone sympathize. In our study, because posting a picture on the map of safety for children made the subjects sympathize, some of them visited the place after the picture was posted. Since the subjects have a choice about whether to check the map and visit the place or not, visualization of the information can be considered as a nudge. It can be said that visualization of the map of safety for children works well as a nudge. In other words, visualization with altruism can encourage the provision of local public goods among the residents.

In this research, we did not attach short messages to pictures. However, adding short messages may cause changes in people’s behavior. In addition, because we conducted experiments among acquaintances, malicious posts were not observed. Moreover, the posted picture has the possibility of be used for crime. In the future, we would like to extend the current research as well as address these issues.