Analysis of the Activities of the Website “Question and Answer about Radiation in Daily Life” after the Accident at the Fukushima Daiichi Nuclear Power Plant and Some Lessons Learned from It: To Pass on This Experience to the Future

1. Stance

We decided on the stance in the Q&A website based on three principles in consideration of the public’s deep distrust for the central government, the municipal authorities, and the electric operators at the time. First, we will say “there is no need to be worried about the risk from radiation” where appropriate, depending on circumstances and the available information. Second, we will provide answers based on scientific findings and our expertise, recognizing the importance of gaining trust from the public and reducing their concerns. Third, we will need to be accountable for our answers as a team of experts (it is directly related to our credibility as professionals on radiation protection).

In the specific work process, we took care to ensure the following attitudes: (1) answering all questions, (2) address any and all questions with care, (3) remain objective, and (4) maintain the dignity of the JHPS. Furthermore, in order to ensure the objectivity of the answers, we presented the data as quantitatively as possible by showing mathematical formulas and calculation processes, as well as citing the sources of the data and figures used (dose conversion coefficients, etc.). In addition, we focused on responding to the questions instead of using the website as a platform for our opinion.

The contents of the questions started to change after about a month after the website’s launch, including an increase in the number of questions regarding the risk of health effects from radiation. In responding to those questions, we made sure to address the following points: (1) There is no “zero risk,” (2) uncertainty is unavoidable, and (3) consider and calculate the representative population risk as a basis of the answer (as it was not possible to evaluate the risk of a certain individual in a specific circumstance).

The Q&A website was our attempt to provide interactive communication on radiation, but it would still appear as one-way from the viewpoint of the citizens who did not send the questions. To address this problem, we utilized the social networking service (SNS) Twitter. Twitter users can post short messages (tweets) of up to 140 characters, and other Twitter users can “Reply,” “Retweet,” and “Like” them. The term “Retweet” refers to the act of re-distributing a tweet, whereas “Like” refers to a positive reaction to a tweet. The JHPS created a Twitter account for Q&A [8] and tweeted all Q&A website updates to promote our activities, to invite Twitter users to our website, and to possibly generate exchanges on Twitter.

2. Procedure for Responding to the Q&A website

The details of the establishment of this website will be referred to in the articles such as “Activities in the early stage after the occurrence of the accident,” [1, 3, 9]. This paper will provide a summary of the development after the launch.

As explained earlier, the JHPS volunteer team initially drafted the answers to the questions. Out of the 22 team members, 16 members were involved in writing the answers before the activities were transferred to the Committee on July 20, 2011. Six members of them were retirees and the rest of them worked full-time in universities or research institutes (As mentioned in the Introduction section, JHPS members working for corporations or authorities were busy in their professional capacity, thus researchers played a central role in developing the answers for the Q&A website).

The names of the people coordinating the Q&A website were published, but we chose not to disclose the name of the individual who wrote the specific answer to spare him/her from extra burden of responding to additional inquiries amid the already busy work schedule (Also, we set up a system so that the Committee as a group was responsible for the answer from the Committee).

An answer was prepared through the following process. After receiving a question, it was sent to a team member who wrote and sent the initial draft to the coordinator. The coordinator would review the draft and ask for any addition, revision, etc. as necessary. The coordinators took care to correct obvious errors and looked for deviations from the viewpoints of JHPS but respected the works of the fellow members. After reflecting these comments and making necessary revision, the writers of the original draft posted the final answer on the website.

As explained above, the Q&A Committee was officially established in the the JHPS in August 2011. Thereafter, risk communication activities involving radiation to the public became JHPS’s responsibility, until the website’s closing in February 2013. The Committee had approximately 50 members. Younger members of JHPS played key roles with approximately 40 members from the JHPS Young Researchers’ Association and the Students’ Association. They were supported by about ten senior experts who played advisory roles. The young researchers drafted the answers and the advisors reviewed and commented the draft. A diagram describing the workflow is shown in Fig. 1 [10].

Fig. 1

The correspondence process of answering.

1. The Questioners

The questions sent from the public in the first month or so after the launch clearly showed how the questioners refused to accept ambiguous explanations such as “Not dangerous, but not entirely safe.” This was a direct reaction to the government’s overuse of the phrase, “There would be no immediate threat.” Therefore, we adopted a principle to explain that “zero risk” was impossible in our answers and aimed to provide an estimate of the health risk using available data at the time.

After a few months, we started to see some changes in the people’s attitude, judging from the tweets in response to our Q&A. It could be summarized as an answer vouching for safety is actually suspicious, and an honest answer pointing out the true risk is more trustworthy and thus welcome. The change may reflect the fact that more information on the accident and the situation in Fukushima became available over time via the internet and the mass media. It is understandable how the public struggled with the many and often conflicting information. And it is natural for people to be skeptical when even authorities and experts appeared to disagree with each other. To address this issue, we introduced and emphasized the view of the International Commission on Radiological Protection (ICRP) about the probabilistic effects of radiation below 100 mSv in our answers, as it is the consensus of leading experts. Specifically, we assessed the radiation dose in the particular circumstance that the question referred to, and if it was below 100 mSv, stated that “(scientific findings to this date show that) there is no need to worry if the dose does not exceed several tens of mSv.”

In retrospect, our response may have been confusing for those without knowledge of radiation. However, in the early days of the accident, when there was no foreseeable future and the situation was urgent, we believed it was more important to secure a minimum standard of living while utilizing scientific knowledge to avoid the health effects of radiation to the extent possible. This meant we tried to alleviate the anxiety of those who sent us questions by providing them answers based on scientific findings. It should be noted that in the initial days after the accident, radiation measurement data was significantly limited. Thus, we had to estimate exposure doses using extreme assumptions, such as calculating the dose assuming a person stayed outdoors 24 hours being exposed to the highest dose rate measured on the day. We knew the result would grossly overestimate the exposure, but we chose this method to show that if the calculated dose did not exceed several tens of mSv, the actual exposure would have been significantly smaller. We used the expression “several tens of mSv,” which was independently established by the JHPS members, taking into consideration both the risk of radiation (the health effects of 100 mSv or less are not well understood scientifically, and they are much lower risk than the risk of cancer caused due to lifestyle factors) and the risk of other social damage.

Reflecting the situation, our answers became increasingly careful, polite, and lengthy. This tendency became more pronounced after the Ministry of Education issued the “Tentative Concept for Determining the Use of School Buildings and Grounds in Fukushima Prefecture (3.8 μSv/h),” and the government’s plan to reduce the dose with a long-term goal of 1 mSv/yr was widely shared [11].

It should be noted that these reference levels are only guidelines for the effective implementation of radiation protection measures; they do not indicate exposure dose limits or the line between safety and danger.

2. Situation of the Respondents

The 16 members of the volunteer JHPS team came from a diverse range of fields, including radiation protection, radiation control, the environment, and biology. We posted more than a dozen questions with answers on the website each day, and sometimes even more. Because the members were radiation protection researchers and engineers with sufficient knowledge and experience, the team could respond to questions slightly beyond their specialties, but we also took care to match the team member with the knowledge/expertise required to provide an appropriate response.

There were questions that were directly related to our daily lives as well as questions that experts would not have thought of, and we were frequently stumped as to how to respond to them. In hindsight, the need to provide answers quickly led to some hastily written responses. The coordinators’ role as gatekeepers was also limited by time-constraint, and there were cases where the answers did not directly address the concerns raised by the question.

3. The Number of Questions and the Characteristics of Questioners

Our website allowed the person sending questions to enter their attributes including gender, age, occupation, and place of residence. After compiling the data, it was estimated that women made up the majority, according to the gender entry and the contents and the style of the questions. Since we received many questions on babies and children, we assume many were mothers in their late 20s to their 30s. If they did not provide their occupation information, we had no way of knowing if they were homemakers or working mothers. As for the background and expertise, as well as their level of knowledge on radiation, we could only assume these information based on the content of the question.

Fig. 2 shows the number of questions received in the early stages of the response (from the launch until July 2011) for each 10-day period. The graph shows that the Q&A website was closed from April 26 to May 5 because of Golden week holiday period, early May. We also closed it from May 26 to June 5 and June 16 to June 25 in order to handle a large number of questions. After the website reopened, we received about double the number of questions and struggled to post timely answers.

Fig. 2

Distribution of the number of inquiries for 10 day-periods.

Fig. 3 shows the number of questions received between June 6 and July 1, 2011. We recognized that people from Tokyo, Chiba, and Fukushima sent in high number of questions. It is understandable that the residents of Fukushima took a keen interest, but the residents of the Tokyo capital region―Tokyo, Kanagawa, Chiba, and Saitama―sent many questions as well.

Fig. 3

Distribution of the number of inquiries over regions.

After the Committee was established, the demand was strong, and there was a week when we received approximately 100 questions. However, the number of questions decreased over time. In February 2013, we decided to terminate the website as few questions came in. Throughout the website, housewives with children were most likely to send questions. The data shows their keen interest in the health effects on their children. As previously stated, the information such as gender, age, and place of residence was self-reported, and the results of the analysis described here only allow us to draw a rough picture.

4. Characteristics of the Questions

The Q&A website received a wide range of questions, including simple questions, opinions, criticism of the answers posted, distrust of the government and experts. Some sought advice on life. We analyzed 246 questions posted from June 6 to July 1, 2011, in the initial days of the website, to count the frequency of the keywords being used. Table 1 shows the keywords (The Committee developed a set of keywords. The member who maintained the website assigned appropriate keywords for the individual Q&A.) that appeared in the questions, in order of frequency (The total is 276, because some questions used more than one keywords). The result can be used to understand the issues the public was most concerned about at the time. Our analysis shows that the top concerns were children’s health and food, followed by contamination and decontamination.

Table 1

Keywords Used in the Questions on our Q&A Website, in Order of Frequency

Whenever there was a major government announcement (e.g., a policy on exposures in a school environment or data on dose) or a significant change in the situation (e.g., the appearance of hot spots), we received multiple questions with similar concerns. As a result, listing the frequency of the keywords appearing in chronological order allowed us to reconstruct what the public was interested in at a specific point in time.

5. Analysis and Discussion of the Questions

The questions revealed that the public had the following concerns:

We can see from the above that education on radiation has not reached the general public and the mechanism of how radiation affects health is difficult to be understood. Also, it is hard to grasp how the risk of radiation is represented, either as a product of radiation weighting factor and absorbed dose or as a product of tissue weighting factor and equivalent dose. In particular, understanding the health effects of low-dose radiation was a huge challenge, as the mechanism is still being debated among experts. If one wants to be scientifically accurate, one cannot determine there is no effect from even a small dose. The expression the experts use for probabilistic effects is difficult and can be misleading for the public, which caused anxiety and led to a suspicion that the experts are conspiring to underplay or hide the real risk. It is natural for the public to have many unanswered questions in such circumstance, and the questions actually sent to our website may be a tip of the iceberg.

On the other hand, those who sent questions showed positive responses to the answers posted on the website. This demonstrates how the interactive communication (i.e., the public could send further question/comment if they were not satisfied with the answer) through the Q&A website contributed to reducing anxiety and distrust from the lack of, or, inappropriately presented information. This was not instantaneous but happened over time through our effort to continue answering all questions. Kinoshita pointed out that the bidirectional nature of information ensures credibility [12], and although earning trust was not our initial objective, our activities seemed to have contributed to this aspect.

1. Compilation of Responses on Twitter

Tsubokura et al. [13] analyzed and summarized the Twitter information related to the accident and radiation in about six months after the Fukushima Daiichi Nuclear Power Plant accident. We also referenced the result in examining the response from the public who did not send questions, using the twitterers’ responses to our website, as mentioned in our stance. In this section, we look at the contents of the Q&A that caused high engagement by counting the number of “Reply,” “Retweet,” and “Like” to the tweets of the Q&A account. The scatter plot matrix of the number of “Reply,” “Retweet,” and “Like” is shown in Fig. 4. The results of Pearson product-moment correlation coefficient [14] for each response are shown in Table 2. Here, one point in the scatterplot of the number of “Retweet” and “Like” in Fig. 4 was confirmed to have an outstandingly large value, so we excluded this value and calculated the correlation coefficient. Table 2 shows the results of calculating the correlation coefficient by Pearson product-moment correlation coefficient with this value. Regardless of the existence of the data, the correlation coefficient is approximately 0.70, which means that there is a correlation between “Retweet” and “Like.”

Fig. 4

Matrix of scatter plots regarding “Reply,” “Retweet,” and “Like.”

Table 2

Pearson Correlation Coefficient (r) among “Reply,” Retweet,” and “Like” without outlier or including outlier

Table 2 shows that there is a strong correlation between “Retweet” and “Like,” followed by a loose correlation between “Reply” and “Retweet.” Table 2 shows that there is a strong correlation between retweet and like, and a loose correlation between reply and retweet. In addition, Fig. 5 shows that the number of “Retweets” and “Likes” was small during the period immediately after the accident, and the number of responses increased more than one year after the accident.

Fig. 5

Chronological change regarding the number of “Retweet” and “Like” for our activity period.

Table 2 shows that there is a positive correlation among the three types of responses. Therefore, the total number of “Reply,” “Retweet,” and “Like” is defined as the “score,” and the tweets with the largest response are extracted using the score. Table 3 shows the contents of Q&A with a score of 31 or higher for the entire period. In addition, Table 4 shows the Q&A with a score of 12 or higher during the first year of the activity, when there were few responses on Twitter.

Table 3

Inquiries with Highest Twitter Engagement (Score≥31)

Table 4

Inquiries with High Twitter Engagement (Total Response ≥12 and <30)

Tables 3 and 4 show that our Q&A-related tweet that received the most “Retweet” and “Like” were those where the questioner expressed his/her lack of knowledge and anxiety. The website received many questions asking for our views on the problem (shown in Tables 3 and 4 are reflection of this situation). The website also received comments and complaints; although not in such large numbers compared to the questions (shown in Tables 3 and 4), these received relatively a large number of “Retweet” and “Like.”

2. Analysis and Discussion

Shown from the questioner’s anxiety based on the keywords used in the question (Table 1), these keywords represent topics which the questioners wanted to know about. Exposure related keywords appeared most often (8 out of 61 in Table 1), such as “exposure-related,” “exposure effects,” and “internal exposure,” in order of the frequency of use. In the questions, people wanted to learn about the reference values of radioactive materials in food in Japan, internal exposure through breathing, and health effects of exposure to low doses of radiation such as 1 mSv and 100 mSv.

The second most sought-after topic was food-related (7 out of 61 in Table 1), linked to keywords, such as “safety,” “standards,” and “radioactivity in food.” Specifically, they wanted to know about the new regulation values for radioactive materials in food and the amount that can be eaten in a day, the concentration and amounts of radioactive materials in food, and the safety of food.

Keywords related to children appeared the same number of times (7 out of 61). The questioners wanted to learn the actual situation of their children’s exposure to radiation, and what will happen to them in the future.

Some questions were related to expertise in radiation protection, such as “What is a ‘radiation controlled area’?,” “The difference between the lower limit of detection, detection limit, and measurement limit, and the number of becquerels in food,” “I want to know the screening level,” “I want to know how to convert Bq/kg to Bq/m²,” and “What is the lower limit of quantification?”

Questions asking for the opinions of websites included: “I would like to know the website’s opinion on thyroid abnormalities in children,” “What do you think of the opinion on thyroid disease after Chernobyl in the TV program of NHK (Nippon Hoso Kyokai), the Japanese public broadcasting company?,” “I would like to know your opinion on the contents of a blog written by an expert,” “I would like to know the credibility and position of a certain academic paper,” and “I would like to know your opinion on the articles by Dr. Christopher Busby and Professor Yury Bandazhevsky.”

Other questions not related to radiation, but classified as comments and complaints on the Q&A website included “The dangers should be communicated in full” (score=43; Reply=4, Retweet=24, Like=15), “We should not say that something is safe if the effects are not clear” (score=42; Reply=2, Retweet=20, Like=20), and “Is this a confident answer that will not be a problem even if it is seen after 10 years from now” (score=15; Retweet=7, Like=8).

The above analysis indicates that the following principles/practices of our activities were welcomed: (1) the format of “answering questions from the general public” worked well, (2) in the initial stage, it was effective to reply that something (asked by the questioner) was safe, as long as it was supported by science, (3) the speed of answering a question was important, even at the cost of some accuracy, and (4) we should explain as clearly as possible, trying to understand and respect the questioner’s point of view (specifically, include full details without abbreviating explanations, and show numerical values and calculation formulas to show our answers were backed by science). In fact, we believe the above were received quite favorably. We adopted these principles/practices from the start, and the fact that they received a positive reaction shows our initial stance was not wrong.

After the initial stage, we stopped emphasizing that it was considered safe as long as the dose did not exceed several tens of mSv. Instead, we focused on providing accurate information based on actual measured data and scientific findings, so that the questioner may make the best decision reflecting his/her lifestyle choice and risk perception. Our answers started to reflect this tendency as our colleagues agreed on the need for a change. However, there was no noticeable response on Twitter regarding the change.

The issues for the future will be discussed later, but it should be stressed that our Q&A is an excellent archive of actual post-nuclear accident communication, showing what kind of topic/information was sought at each moment in time. Thus, it serves as a good source of information for someone providing information to the public after a future accident. At the same time, the questions, answers, and tweets show how the people’s interests/attitudes changed over time and can provide hints for post-crisis communication in general.

It should be added that analyzing Twitter activities was difficult. For example, what is the difference between “Retweets” and “Likes”? Or how should we interpret the large number of tweets in response to our tweet; whether the tweet reflected sympathy for the questioner, interest in the answers, or both?