A nice visualisation of the data, done by the University of Lund, is here: https://lu-unevil.github.io/WintonReport/WintonReport-en.html Survey data from **US and UK** collected on 19th March 2020 using to the US Questionnaire saved in this project. Participants were recruited through the online platform Prolific.co and were representational of their respective countries on gender, age and ethnicity. Note that since participants were recruited online they are not representative of educational background and socioeconomic status. This was the best way for us to get fast and extensive data, and we are aware of its limitations. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. Rapid graphs that we have made from it, we wrote about here: https://medium.com/wintoncentre/coronavirus-and-public-trust-e156c89be5d4 Survey data from **Italy** collected on 21st-24th March 2020 using the Italian translation questionnaire saved in this project. Participants were recruited online through sampling company Respondi. Due to difficulties with the representational quotas, some oversampling occurred in some groups. We have removed excess participants per quota according to time stamp (i.e. as if the quota system had worked as programmed) and this sample is of 700 participants representative by age and gender. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. Survey data from **Spain** collected on 22nd-25th March 2020 using the Spanish translation questionnaire saved in this project. Participants were recruited online through sampling company Respondi. Due to difficulties with the representational quotas, some oversampling occurred in some groups. We have removed excess participants per quota according to time stamp (i.e. as if the quota system had worked as programmed) and this sample is of 700 participants representative by age and gender. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. Since data was collected over an extended time period, please check the end date time stamp for details of when each submission occurred. Survey data from **Australia** collected between 20th and 24th March 2020 using to the US Questionnaire saved in this project. Participants were recruited online through sampling company Dynata. Due to difficulties with the representational quotas, some oversampling occurred in some groups. We have removed excess participants per quota according to time stamp (i.e. as if the quota system had worked as programmed) and this sample is of 700 participants representative by age and gender. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. Please refer to the column showing the date and time of completion of the questionnaire since this data was collected over several days during which the political and pandemic situation was changing rapidly. Survey data from **Germany** collected between 23rd and 24th March 2020 using the German translated questionnaire saved in this project. Participants were recruited online through sampling company Respondi. Due to difficulties with the representational quotas, some oversampling occurred in some groups. We have removed excess participants per quota according to time stamp (i.e. as if the quota system had worked as programmed) and this sample is of 700 participants representative by age and gender. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. Survey data from **Mexico** collected mostly on 21st and 22nd March but with some continuing to 26th March 2020 using the Spanish translated questionnaire saved in this project. Participants were recruited online through sampling company Respondi. Due to difficulties with the representational quotas, some oversampling occurred in some groups. We have removed excess participants per quota according to time stamp (i.e. as if the quota system had worked as programmed) and this sample is of 700 participants representative by age and gender. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. Please also use the 'End Date' column to determine when each participants completed the survey. Survey data from **Sweden** collected on 28th-29th March 2020 using the Swedish translation questionnaire saved in this project. Participants were recruited online through sampling company Respondi. Due to difficulties with the representational quotas, some oversampling occurred in some groups. We have removed excess participants per quota according to time stamp (i.e. as if the quota system had worked as programmed) and this sample is of 700 participants representative by age and gender. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. Survey data from **Korea** collected on 9th-11th April 2020 using the Korean translation questionnaire saved in this project. Participants were recruited online through sampling company Respondi. Due to difficulties with the representational quotas, some oversampling occurred in some groups. We have removed excess participants per quota according to time stamp (i.e. as if the quota system had worked as programmed) and this sample is of 700 participants representative by age and gender. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. Survey data from **Japan** collected on 10th-12th April 2020 using the Japanese translation questionnaire saved in this project. Participants were recruited online through sampling company Respondi. Due to difficulties with the representational quotas, some oversampling occurred in some groups. We have removed excess participants per quota according to time stamp (i.e. as if the quota system had worked as programmed) and this sample is of 700 participants representative by age and gender. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. Survey data from **China** collected on 9th-10th April 2020 using the Simplified Chinese translation questionnaire saved in this project. Please note that some questions were not possible to ask in China and hence the questions and the response coding is different in this dataset. Please refer to the Questionnaire for full details. Participants were recruited online through sampling company Respondi. Due to difficulties with the representational quotas, some oversampling occurred in some groups. We have removed excess participants per quota according to time stamp (i.e. as if the quota system had worked as programmed) and this sample is of 700 participants representative by age and gender. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. A second survey was done in the **UK** with data being collected between 10th and 12th April. This was a larger dataset, with 1050 being collected through Prolific.co, representative on age, gender and ethnicity and 1050 being collected through Respondi (participants who were registered with both companies were excluded to prevent overlap). This survey had some different questions in. The data file provided is from both participant pools combined, but includes a variable to distinguish them. A third survey was done in the **UK** with data being collected between 8th and 9th May. This was another larger dataset, with 1150 being collected through Prolific.co, representative on age, gender and ethnicity and 1150 being collected through Respondi (participants who were registered with both companies were excluded to prevent overlap). This survey again had some different questions in. The data file provided is from both participant pools combined, but includes a variable to distinguish them. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. A second survey was done in the **US** with data being collected on 8th and 9th May. This was a sample of 700 participants through Respondi. It had some additional questions compared with the first survey and has a more balanced political make-up. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. A second survey was done in **Spain** on 6th-7th May, with 700 participants recruited through Respondi. Again, it has additional questions not asked in the first round in Spain. We have anonymised the data but not cleaned it extensively - please use the available questionnaires to check the coding of different answers and be aware that this is raw, unexplored data. Data on questions that we have pre-registered hypotheses about will be made publicly available as soon as we have completed analysis. We have made public anything that might be informative to those making decisions around communication. Please mention the Winton Centre for Risk & Evidence Communication at the University of Cambridge as the source if you cite this data. If you wish to publish academic papers using it, please contact us first (we are happy for others to publish, but we'd like to know what you plan!): alex.freeman@maths.cam.ac.uk