This experiment was given ethical approval by the University of Cambridge’s Psychology Research Ethics Committee (PRE.2023.087)
UK-based Participants were recruited via the platform Prolific.com using their ‘representative sample’ option to give a participant pool representative of the UK’s population on the basis of age (5 brackets), sex (2 groups) and ethnicity (5 groups). Prolific state that they use the UK census to create the appropriately sized subgroups. The questionnaire itself was hosted on Qualtrics.
A total of 1497 participants provided informed consent and completed the survey. As preregistered in the Method, 28 participants (1.9%) failing an attention check ('please select 'slightly worried' for this question') were removed, resulting in a final sample of N = 1469. In this final sample, gender was relatively balanced (50.8% women, 47.3% men, and 1.9% identifying as non-binary or other. The mean age was 45.1 years (SD = 15.6), and 57.1% held a Bachelor's degree or higher.
Sample characteristics
Category | Response | n | % |
Gender | Women | 746 | 50.8 |
Men | 695 | 47.3 | |
Other | 28 | 1.9 | |
Education | No qualification | 9 | 0.6 |
Primary school | 2 | 0.1 | |
GCSE / O-Level / BTEC NVQ Level 2 | 222 | 15.1 | |
A-Level / International Baccalaureate / BTEC NVQ Level 3 | 246 | 16.8 | |
Higher National Certificates and Diplomas / Other vocational | 144 | 9.8 | |
Bachelor's degree or equivalent | 552 | 37.6 | |
Master's degree / Postgraduate qualification | 248 | 16.9 | |
Doctoral degree | 39 | 2.7 | |
Prefer not to say | 5 | 0.3 |
The median completion time for the experiment was 13.19 minutes, and participants were paid £7.16 per hour pro rata.
The survey contained 2 experiments. Participants were first given text explaining what R was and who SPI-M was (being made up of several independent experts). In the first experiment after that, participants were randomly assigned to one of five conditions in which they read three similar statements about the R number reported by SPI-M, presented as a range. The conditions are summarised in the table below.
Format | Content | n |
1. Control | No additional information | 291 |
2. Stated consensus | R range is presented as consensus of SPI-M | 291 |
3. Highly likely | R range described as ‘highly likely’ | 296 |
4. Little consensus | Notes there is ‘little consensus’ over the number | 293 |
5. High consensus | Notes there is ‘high consensus’ | 298 |
The control condition (1.) is shown with additions comprising other conditions in brackets:
The SPI-M group report that this week [2: their consensus is that] the overall reproduction number, R, of Covid-19 is [3: highly likely to be] between X and Y. [they state that there is [4: little] [5: high] consensus on this.].
Each participant rated three statements (random order), where the X-Y range is: 1.4-1.6, 0.9-1.1, and 0.4-0.6.
After each statement, participants were asked 10 questions, with responses captured on 0-100 sliding scale:
Variable label | Item text | Scale anchors 0-100 | ||
Perceived risk | How risky does this R number make the COVID-19 situation feel to you? | Very low risk | Very high risk | |
Perceived likelihood of upper limit | How likely do you think it is that the value of R is [upper limit]? | Not at all likely | Very likely | |
Perceived likelihood Covid-19 increasing | How likely do you think it is that Covid-19 is currently increasing, given this information? | Not at all likely | Very likely | |
Perceived R certainty | How certain or uncertain do you think the R number is? | Not at all certain | Very certain | |
Surprise if above range | How surprised would you be if the value of R actually turned out to be [0.2 above range]? | Not at all surprised | Very surprised | |
Surprise if range midpoint | How surprised would you be if the value of R actually turned out to be [range midpoint]? | Not at all surprised | Very surprised | |
Surprise if below range | How surprised would you be if the value of R actually turned out to be [0.2 below range]? | Not at all surprised | Very surprised | |
Trustworthiness | How trustworthy do you think the information is? | Not at all trustworthy | Very trustworthy | |
Effort required | How much effort did you have to put in to understand the information? | No effort at all | A lot of effort | |
Informedness | How well informed does the statement make you feel about the Covid-19 situation? | Not very informed | Very informed | |
In the second experiment, participants were randomly assigned to one of three conditions in which they were presented a graphical representation of the R number as a ‘combined consensus’ interval with range [X-Y](based on several different models).
Format | Content | n |
No Uncertainty | Only consensus interval displayed | 485 |
Low QoE | Consensus interval displayed along side intervals for each model, with high variability | 489 |
High QoE | Consensus interval displayed along side intervals for each model, with low variability | 495 |
Participants rated two R graphs each (in random order), where the X and Y range was 1.4-1.6 and 0.4-0.6. Participants were asked questions identical to Experiment 1.
The full dataset and questionnaire can be found in the OSF repository: https://osf.io/ypf6c/.