Sufficient information on the methodology is included in each research report, and everyone is free to make their own value judgments.By reading our research methodology, a reader can assess any criticism leveled against the polling. We strive to be as objective as possible. That is why we provide information on the methods and timing of the study, the sample, the wording of the questions, and all response options.

Also, for the sake of transparency, we publish the raw data (all respondents' answers) for each round of polling. Anyone can check our statistical calculations and carry out their analysis based on these data.

We interview about a thousand people every two months. The composition of our sample corresponds to the general structure of the urban population of Belarus by gender, age, education, and size of the city of residence of the respondents. The statistical error in our research is 3%: our data can be wrong by that amount in either direction.

We survey an online panel using the computer-assisted web interviewing (CAWI) method, which in practice means it is limited only to internet users. This allows us to carry out these studies very quickly. Very importantly, survey participants feel more secure when they fill out the online survey themselves than they would if they were answering the questions over the phone.

Yes, it is. Several sampling models are used for Belarus and represent the entire population of the country with a specified error (“the sampling error”). The sizes of these samples range from about 600 to 2,500 people. For example, the Institute of Sociology of the Academy of Sciences conducts research with a sample of 600 people.

The choice of model used to obtain a representative sample population is determined by the study objectives, the resources available, and time constraints. In times of political instability, the time factor is critical: the situation is changing rapidly, so surveys need to be conducted in a short timeframe, otherwise there is a risk that information collected at the start of the study becomes out of date by the end.

For these reasons, a sample size of 1,000 people was chosen. This is sufficient for our research objectives.

Sociologists constantly change their methods of collecting data. Initially, sociologists collected data live (face-to-face), later telephone interviewing became the dominant method, and now sociologists most frequently use online surveys.

This is not a significant methodological problem since the number of internet users in Belarus is among the highest in the former republics of the USSR. According to a study by DataReportal, in 2020, Belarus had 7.8 million internet users, which is 83% of the total population. According to rough estimates, this is about 90-95% of the population over 15 years old.

CAWI is widely used both on its own and in combination with traditional face-to-face and/or telephone surveys. Its main advantages are the speed of data collection and the relatively low cost, without any loss of data quality. For example, online polls are used by reputable social science services such as Gallup and the Pew Research Center.

In addition to these features of CAWI methodology, we took into account the Covid epidemic. Remote polling allowed us to maintain social distancing to minimize the risk of spreading coronavirus. Another important motivation is the restrictions on sociological research in Belarus. We cannot risk the freedom and well-being of our interviewers. Accordingly, the CAWI technique seems optimal.

There are both organizational and methodological difficulties. Despite these, we are working on solutions and planning to expand the study to this segment of the population. It is not yet possible to give exact dates, but we will certainly include the rural population in the sample in the future.

At the same time, our preliminary analysis of the situation showed that, due to the level of urbanization in the country and the specifics of the information and communication space, there are no critical differences between answers given by rural and urban residents. Undoubtedly, Belarusian villagers differ from the inhabitants of Minsk, but their worldview is similar to that of residents in small towns.

We have developed and validated in the course of several tests a special tool that allows us to divide society into groups without asking a direct question about which side of the political conflict the respondent supports. This tool is a set of 32 statements with which we ask the respondents to agree or disagree. Consent or disagreement is expressed using a five-point scale, where 1 means ‘strongly disagree’ and 5 means ‘strongly agree’.

The original segmentation tool was Ward's hierarchical cluster analysis.

To assess the dynamics of changes in segments, we use cluster centres (average values for each of the clustering statements), which we set as starting parameters when conducting cluster analysis using the K-means method.

The results of our analysis are simple, logical, and are verified by analysing the differences between the identified segments of society on a number of issues that are presented in the study.

At the same time, we also ask a direct question about support for the parties in the political conflict. It is used to check the clustering results – most often, people who identified themselves as supporters of the authorities end up in the ‘Lukashenka’s base’ segment, just as almost all people who identified themselves as protesters are included in the ‘hardcore protesters’ segment.

At the same time, clusterization redistributes Lukashenka’s base and hardcore protesters to 35% of those who, when answering a direct question, said that they were closer to a neutral position.

To what extent do you agree or disagree with the following statements?
1. The results of the elections on 9 August 2020 were falsified.
2. I support peaceful protests against the official election results.
3. Even if the election on 9 August 2020 was not completely transparent, I am satisfied with the result.
4. I am a supporter of the current government.
5. Law enforcement actions to suppress peaceful protests were excessive.
6. I try to stay away from political issues.
7. The official results of the 2020 elections reflect the will of the Belarusian people.
8. I approve of the dismissal of people who do not support the current government.
9. I do not believe in the fairness of the Belarusian judicial system in cases of bringing to justice those who disagree with the political processes in the country.
10.Opponents of the current government are traitors to the Motherland.
11.I believe that our country is developing in the right direction.
12.I think that if everything remains as it is, the country's economic situation will
improve in the next year or two.
13.I do not approve of a close relationship with Russia.
14.In general, I have little interest in political events in the country.
15.I'm tired of constantly discussing political issues.
16.I trust Belarusian state television.
17.Stability is the basis for the development of the state and economy.
18. I feel confident about tomorrow.
19. I feel safe in Belarus.
20. A woman cannot be trusted to rule the country.
21. I am proud of Belarus’s position on the world stage.
22. I believe that the forces opposing Aliaksandr Lukashenka will win.
23. The police should be guided in their actions not only by the law, but also by political instinct.
24. I try not to pay attention to political events.
25. I do not believe that anything will change in Belarus in the near future.
26. Belarus has the democratic rule of law.
27. Human rights are sufficiently respected in Belarus.
28. Lukashenka will remain in power for a long time.
29. I am convinced that the protest movement was caused from outside by
unfriendly actions of foreign states.
30. Russia is the closest ally of Belarus.
31. Belarus should become part of Russia.
32. Those who disagree with official policy should not be able to engage in
teaching activities (teach in schools or universities).

Our preliminary conclusion is that fear does affect the results, but insignificantly. The results of this wave of research allowed us to assess the scale of distortion that the ‘fear factor’ introduces in the answers to certain questions in online research on sociopolitical topics.

We estimate the maximum possible distortion level to be around 9 per cent.

At the same time, a significant part of this distortion can be explained by the structural change in the sample of the online panel (caused by the departure from the country of a significant number of Belarusians opposed to the government) after the events of 2020, random sampling error, and other external factors.

This issue requires further study, and the Chatham House team will continue to work on the issue of the influence of fear and the measurement of this influence in the next waves of its research.

It is important to note that there are no grounds for the assertion that the share of support for the authorities and Lukashenka is lower than that observed in the Chatham House studies. People with ‘neutral’ views are most susceptible to the fear factor, and they give ‘neutral’ answers to questions. People who support the authorities, and Lukashenka, in general feel much less fear when answering questions.

This is the result of the high degree of polarization around the figure of Lukashenka in Belarusian society.

Analysis of the answer to the question about awareness by segment shows that ‘ignorance’ of Lukashenka’s figure is more characteristic of hardcore protesters. It can be assumed that in this way people express a protest against him personally.