How the survey was conducted

The Russian Speakers in Finland 2026 survey (like the 2022 and 2024 surveys) was conducted on behalf of the Cultura Foundation by Bondata Research Oy (formerly Innolink Research Oy).

The study is based on 1,022 responses. The 2024 dataset contains 1,000 responses and the 2022 dataset 1,632. The margin of error for the study at the 95% confidence level is ±3.1 percentage points. The margin of error does not apply to the mean values presented in the report or, as such, to subgroup comparisons. No weighting was applied to the data.

The results were analysed using various background variables. In the subgroup analyses, some groups are very small — for example, there were 13 respondents who had lived in Finland for less than two years and 14 respondents who had arrived on the basis of temporary protection.

Two thresholds were applied in the report text when interpreting small groups. The results for groups with fewer than 50 respondents are interpreted as descriptive, and the number of respondents in the group is reported in connection with them. In groups of fewer than 30 respondents, results are presented primarily as numbers of respondents rather than percentages, and no conclusions are drawn beyond the group itself.

The target population of the study consisted of people aged 20 or over living in Finland whose native language was recorded as Russian in the Population Information System. For the study, Dun & Bradstreet provided an interview register of 6,587 people drawn from the Digital and Population Data Services Agency (DVV) register. Target numbers by major region, age group and gender were used in the sampling, and one person per household was included in the register. A condition for sampling was that a telephone number was found for the person. The youngest age group included those who had turned 20 by 30 June 2026. The target was 1,000 responses, with quotas set by age and gender within each major region in line with the distributions of the target population. The information available about the sampled people was their name, address, municipality, age group, gender and telephone number. This background information was therefore not asked about separately in the interview.

The interviews were conducted by telephone in Russian between April and June 2026. Each number was called up to five times on different days and at different times of day. For many of the questions, alternative responses were not read out to the respondent; instead, the respondent replied in their own words and the interviewer classified the response under the most appropriate alternative. The same practice was followed in 2022 and 2024. The interviews were conducted by professional Russian‑speaking interviewers living in Finland, and the lead interviewer responsible for implementation was the same across all three survey rounds in 2022, 2024 and 2026.

The respondents’ contact details and responses were handled securely, ensuring the anonymity of the respondents.

Although this is a follow‑up study, the 2022, 2024 and 2026 survey rounds are not fully identical in all respects: for example, in 2022 responses could also be given in Finnish, whereas in 2024 and 2026 the interviews were conducted in Russian, and there are changes in the wording or implementation of some individual questions.

The interview register comprised 6,587 people, all of whom had a telephone number available in accordance with the sampling criterion. During the telephone interviews, 1,683 people were reached. Of the remaining 4,904, some could not be reached because there was no answer or the number was not in use, and some were not contacted because the interview quota for that group had already been filled.

Of those reached, 1,022 engaged in the interview and 591 refused to respond. In addition, 70 people were screened out on the basis of language: they did not speak Russian at all, spoke it poorly or refused to speak it for other reasons. When those screened out on the basis of language are counted among refusals, as in previous survey rounds, there were 661 refusers in total, or 39% of all those reached (1,683). Respondents accounted for 61% of those reached. For comparison, in the 2024 study, refusals numbered 702 or 41%, and in the 2022 study, the proportion of refusals was 23%. When assessing refusal rates, it should be noted that in 2022, it was also possible to respond in Finnish, whereas the 2024 and 2026 interviews were conducted exclusively in Russian. The following examination of reasons for refusal relates to the 591 actual refusal situations and does not include those screened out on the basis of language.

Five main reasons for refusal were identified (Figure 1). The most common reason was that the respondent was not interested in responding or did not want to participate (240 people). The second most common was refusal without giving a reason (142 people), followed by hanging up (106 people), situational reasons and life circumstances (77 people), and distrust and matters of principle (26 people).

In addition to complete refusal, partial refusal is visible in the data. Interviewers recorded 168 observations from the interviews, and in 67 of these it was noted that the respondent had left a question or question group unanswered. By far the most commonly unanswered questions concerned the war or the eastern border (24 observations), and in 16 observations the respondent themselves described the questions as political. This should be taken into account particularly when interpreting questions concerning the war and Russia’s actions, as the proportion of respondents leaving them unanswered is slightly higher than for other questions.

Due to partial non‑response, the number of responses to the study’s questions and the constituent individual statements or items being rated varies, in some cases quite considerably. Differing Ns are indicated when discussing the results. In certain places, an examination was also made of what kinds of individuals had specifically left certain questions unanswered.
The relatively high proportion of refusals should be taken into account when interpreting the results, as non‑response may cause selection bias in the data. However, based on the available information, it is not possible to assess how those who refused differed from respondents or in which direction any potential selection bias affects the results.

Of the respondents in the core sample of the Russian Speakers in Finland 2026 follow‑up study, 165 were aged 20–29. The core sample was drawn from the Digital and Population Data Services Agency’s Population Information System, and its target population consisted of people aged 20 or over living in Finland whose registered native language is Russian.

The dataset was supplemented with an additional sample of 20–29‑year-olds, which yielded 50 interviews. Like the core sample, the additional sample was drawn from the Digital and Population Data Services Agency’s Population Information System and was conducted using the same telephone interview method and the same questions as the core sample.

Of the 165 people aged 20–29 in the core sample, four responses were excluded from the separate analysis of young Russian speakers, so 161 core sample respondents were included in the separate analysis.

With the additional sample, the dataset for the separate analysis of young people contains 211 respondents aged 20–29.

In this report, the responses of the additional sample are not taken into account when reviewing the survey results, but a separate subsection, which also draws on the additional sample, presents the key findings concerning young Russian speakers in Finland.