The study was conducted via an online survey distributed through the Qualtrics system of the Data Center at the INSS. The surveys were distributed to participants of the iPanel online panel across two sub-samples: a Jewish sample and an Arab sample, constituting—after sectoral weighting—a representative sample of the adult Israeli population (aged 18 and older). The survey conducted in July included 968 respondents. The fieldwork for the survey was conducted between June 28 and July 2, 2026. As part of the quality control process, 32 respondents were removed due to completing the survey too quickly or because they were identified as duplicates. The survey conducted in February 2026 included 921 respondents from the general public (following the application of the same quality control measures), alongside the expert survey (N = 147) used in this report for point-in-time comparison. Findings for the closed-ended questions were analyzed using SPSS software.

The survey examined public perceptions regarding 14 threat scenarios. Each respondent was asked to select the three scenarios they considered most threatening from the list, and subsequently answered follow-up questions (likelihood of occurrence in the next five years, anticipated level of damage and, in the current survey, anticipated impact on various domains and level of preparedness) regarding the three scenarios they selected. Additionally, to enable an estimation of the entire sample's attitudes toward each scenario (and not just among those who selected it), each respondent was presented with follow-up questions for additional scenarios randomly selected from those they did not choose: three scenarios in the current survey, and seven in the previous survey.

Naturally, the responses of those who chose a scenario as one of the top three most threatening differ from the responses of those who did not select it. In order for the findings to reflect a reliable estimate for the general population (incorporating both groups at their correct weight), the data were normalized using an exposure-weights method: for each scenario, the responses of participants to whom the scenario was randomly assigned also represent the remaining non-selectors who were not assigned that scenario at random. Their weight was calculated as the ratio between the total number of non-selectors and the number of non-selectors actually questioned about that specific scenario, based on the observed data in each survey. This reconstructs the full composition of the sample for each scenario—selectors and non-selectors alike—in their true proportions.

Additionally, the general public sample was weighted to match the proportions of the adult population in Israel—80% Jewish and others, and 20% Arabs. The final weight for the follow-up questions is the product of the exposure weight and the demographic weight. Following the calculation of the final weight, normalization was performed: each respondent's weight was divided by the average of the weights within that specific scenario, such that the average weight after normalization equals 1 and the sum of the weights equals the actual number of respondents. This normalization does not alter the estimates themselves, as dividing by a constant maintains the proportions; however, it prevents a situation in which a sum of weights greater than the sample size artificially inflates the statistical significance tests. The weights were calculated and normalized separately for each of the two sets of comparisons presented in the report—current public versus previous public, and experts versus previous public. The expert sample was not demographically weighted (as it lacks defined population targets), but exposure normalization was calculated for it separately, within the expert group.