Design, content, and fieldwork procedures of the COVID‐19 Psychological Research Consortium (C19PRC) Study – Wave 4

Abstract Objectives This paper outlines fieldwork procedures for Wave 4 of the COVID‐19 Psychological Research Consortium (C19PRC) Study in the UK during November–December 2020. Methods Respondents provided data on socio‐political attitudes, beliefs, and behaviours, and mental health disorders (anxiety, depression, and posttraumatic stress). In Phase 1, adults (N = 2878) were reinvited to participate. At Phase 2, new recruitment: (i) replenished the longitudinal strand to account for attrition; and (ii) oversampled from the devolved UK nations to facilitate robust between‐country analyses for core study outcomes. Weights were calculated using a survey raking algorithm to ensure the longitudinal panel was representative of the baseline sample characteristics. Results In Phase 1, 1796 adults were successfully recontacted and provided full interviews at Wave 4 (62.4% retention rate). In Phase 2, 292 new respondents were recruited to replenish the panel, as well as 1779 adults from Wales, Scotland, and Northern Ireland, who were representative of the socio‐political composition of the adult populations in these nations. The raking procedure successfully re‐balanced the longitudinal panel to within 1% of population estimates for selected socio‐demographic characteristics. Conclusion The C19PRC Study offers a unique opportunity to facilitate and stimulate interdisciplinary research addressing important public health questions relating to the COVID‐19 pandemic.


| INTRODUCTION
The 'first wave' of the COVID-19 pandemic in the UK was abating by the summer of 2020 (Kontis et al., 2020), and citizens were experiencing respite from the government-imposed restrictions on social, economic, and educational related activities that had been in place since March 2020 to control the spread of the virus. For example, in August 2020, 'shielding' initiatives requiring ∼2 million elderly and/or medically vulnerable individuals to self-isolate to avoid contracting COVID-19 were paused across most of the UK (UK Government, 2020a). A month-long 'Eat Out to Help Out' scheme offering discounted meals at indoor venues was launched to support the reopening of businesses (UK Government, 2020c). Employers were actively encouraged to reassure employees that it was safe to return to office workplaces (UK Government, 2020e). Also, the majority of primary and secondary schools re-opened for face-to-face teaching for all students (UK Government, 2020b).
During this time, concerns were raised by public health experts about the potential impact of an equally, if not more, devastating 'second wave', evidence of which was already being reported in parts of Europe, which was predicted to hit the UK by autumn 2020 (Academy of Medical Sciences, 2020; Looi, 2020;Mahase, 2020;Middleton et al., 2020). By October 2020, as the UK COVID-19 reproduction rate was estimated to be between 1.3 and 1.5, signalling high levels of infection transmissibility in communities, onequarter of the UK population (∼16.8 million citizens) was forced back into lockdown (UK Government, 2020d). The Office for National Statistics (2020) reported that COVID-19 cases were increasing rapidly, and that COVID-19 related hospital admissions were close to the peak of the 'first wave' in spring 2020 (see Figure 1). On 5 November 2020, the UK Prime Minister announced a second national lockdown for England, initially for 4 weeks (UK Government, 2020g), and the government 'furlough scheme', which provides up to 80% income support for unemployed workers, was extended until March 2021 (UK Government, 2020f). On 2 December 2020, the lockdown was replaced with a revised regional COVID-19 tiersystem, which re-imposed regulations on social gatherings (UK Government, 2020h). On this same date, the UK became the first country in the world to approve the Pfizer-BioNTech vaccine (Ledford et al., 2020) and the vaccination rollout commenced 6 days later.
The Oxford-AstraZeneca vaccine was anticipated to be approved and subsequently deployed in January 2021  in the UK, 2021).
In late 2020, whilst grappling with preparations for an upcoming COVID-19 surge, as well as planning the population-wide COVID-19 vaccination rollout, the UK government was facing a major historical political event: the end of the Brexit transition period on 31 December 2020 (Wright & Etherington, 2020). The crippling economic impacts of the protracted pandemic, compounded by uncertainty surrounding the short-to-medium term impacts of Brexit, was predicted to widen existing regional inequalities in the UK (Bhattacharjee et al., 2020;Petrie & Norman, 2020). The decision to hold the 2016 European Union referendum, and the voting outcome of that referendum, had already resulted in societal division across the UK and Europe (Hobolt, 2018;Outhwaite, 2017), particularly in relation to views on culture, migration, racism, national identity, and political ideologies (Bachmann & Sidaway, 2016;Corbett, 2016).
Moreover, the pandemic itself, coupled with states' initiatives to control the spread of the disease across populations, may have contributed substantially to the rise of nationalism and its social relevance on a global scale (Bieber, 2020;Woods et al., 2020). Methodological reports for the C19PRC Study in the UK (Wave 1, March 2020; Wave 2, April 2020; Wave 3, July/August 2020) and other countries are available elsewhere (Bruno et al., 2021;McBride et al., 2020McBride et al., , 2021Valiente et al., 2020Valiente et al., , 2021.

| Brief overview of C19PRC Study methodological framework for the C19PRC-UKW4
Collecting data on adults' mental health difficulties (e.g. anxiety, depression, and traumatic stress) at each survey wave, using standardised and validated measures, is a fundamental design feature of the C19PRC Study. Our study has spearheaded COVID-19-related mental health research through the production of timely, highquality research articles, the findings of which have already demonstrated relatively stable prevalence estimates for depression, anxiety, and COVID-19 traumatic stress during the first 6 months of the pandemic in the UK and Ireland (Hyland, Vallières, Daly, et al., 2021;Shevlin et al., 2021 Wave 3 (C19PRC-UKW3) so that the cross-sectional sample at each wave would continue to be (1) representative of the UK adult population (i.e. new participants were recruited to 'top-up' baseline quotas, determined by age, gender, and household income, due to modest levels of attrition) and (2) large enough (n > 2K adults) to conduct sub-group analyses for core study outcomes.
This paper describes the fieldwork procedures for C19PRC-UKW4. Two key decisions were made during the planning phase: (1) to prioritise collection of data on respondents' socio-political views, attitudes, and behaviours to assess the combined impact of Brexit and the COVID-19 pandemic on adults' national identity, and how this might in turn shape their responses to, and experiences of, the pandemic; and (2) to recruit new respondents into the study by oversampling adults from the devolved UK nations (Wales, Scotland, and Northern Ireland), which would facilitate robust betweencountry comparisons for a range of important socio-political outcomes, in addition to the core mental health outcomes. Here, we (i) examine patterns of attrition in the C19PRC Study by this fourth wave and whether these could be predicted by baseline mentalhealth attributes, psychological characteristics, as well as sociodemographic factors; (ii) conduct and assess weighting procedures to manage attrition in the longitudinal panel; (iii) determine the success of sample replenishment and oversampling procedures conducted at C19PRC-UKW4; and (iv) describe the prevalence of common mental disorders among participants in the C19PRC-UKW4 sample, as well as their socio-demographic characteristics and political-related beliefs and behaviours. Phase 1 comprised two strands. Qualtrics re-contacted all adults who participated in any previous wave(s) (N = 2878) via email, SMS, or F I G U R E 1 Graphical presentation of the number of daily COVID-19 cases and deaths in the UK, sourced from Our World in Data, 2020, aligned to the COVID-19 Psychological Research Consortium (C19PRC) Study survey waves. Note: New daily deaths and cases depicted as 7day rolling average in-app notifications and invited them to participate further in this survey (invitations were tailored to remind adults of their participation in previous survey waves). Only 2025 of these eligible respondents participated at baseline and were being invited to participate in a fourth survey; the remaining 853 first entered the panel at C19PRC-UKW3 and were being invited to participate in their first follow-up survey. Phase 1 fieldwork was completed on 22 December 2020.
Fieldwork for Phase 2 (sample replenishment and oversampling) was conducted between 25 November and 19 December 2020.
Similar to the process of recruitment at baseline, new participants for Phase 2 were sampled from Qualtrics' partners' existing survey panels and were alerted to the C19PRC-UKW4 by Qualtrics in one of two ways: (1) they opted to enter studies they were eligible for by signing up to a panel platform; or (2) they received automatic notification through a partner router which alerted/directed them to studies for which they were eligible. To avoid self-selection bias, survey invitations provided only general information and did not include specific details about the contents of the survey. Participants were required to be adults, able to read and write in English, and resident in the UK. No other exclusion criteria were applied. All panel members routinely received an incentive for survey participation (e.g. gift cards), based on the length of the survey, their specific panellist profile, and target acquisition difficulty, amongst other factors.
Qualtrics' partners released invitations in batches and, after the initial invitation was received, respondents who had not completed the survey were sent two reminders to encourage them to participate. The first reminder was sent approximately 36-48 h after the initial survey invite, with the second reminder sent another 36-48 h after this first reminder.

| Informed consent process
Participants were informed about the purpose of the C19PRC Study, that their data would be treated in confidence, that geolocating would be used to determine the area in which they lived (in conjunction with their residential postcode stem), and of their right to terminate participation at any time. Participants were also informed that some topics might be sensitive or distressing. Information about how their data would be stored and analysed by the research team was also provided. Participants were also informed that they would be re-contacted at a later date to invite them to participate in subsequent survey waves. Participants provided informed electronic consent prior to completing the survey and were directed to contact the NHS website upon completion if they had any concerns about COVID-19.

| Compliance with General Data Protection Regulation (GDPR)
C19PRC data will be stored confidentially in line with GDPR. When the study data is deposited with the UK Data Service, location data will be removed and replaced with relevant socioeconomic summary data (e.g. area-level deprivation and population density data). All other personal data will also be removed.

| Quality control
Qualtrics are committed to delivering high-quality survey data from online survey panels and multiple validation checks are conducted on the C19PRC-UK data to ensure this target is met. First, the survey is piloted ('soft launch'; n = 100) prior to the fieldwork going live ('full launch') to rectify sequencing/coding errors and omissions prior to the full launch. The soft launch also calculate the median survey completion time, which provides an opportunity to tailor the content to ensure the median survey time does not exceed 30 min; this is important to minimise respondent burden and maximise participation over time.
For C19PRC-UKW4, a soft launch was conducted (comprising ∼50 respondents) for each phase. The median survey completion times were 23 min 17 s for Phase 1 and 23 min 7 s for Phase 2. These respondents were excluded from the final sample for that Phase. Given the median times were under the 30-min threshold, additional measures were included in the survey prior to the full launch. Second, each participant must achieve 'legitimate respondent status' upon entry into the survey. This means that the respondent must spend a minimum amount of time completing the survey (i.e. half the median soft launch completion time for that wave) the first time they participate.
Respondents who do not achieve this status are flagged as 'speeders' and removed from the study. And third, any respondent who does not meet the inclusion criteria, or who does not complete the survey in full, is removed from the final sample for that Phase.

| Ethical approval
Ethical approval for the project was provided by the University of Sheffield (Reference number 033759).  Third, the outcome of recruitment at Phase 2 for the replenishment or 'top-up' strand was assessed by comparing the characteristics of adults in the combined Phase 1 and Phase 2 samples (excluding the oversample) with respect to gender, age, and household income, compared to the target sampling quotas specified at baseline to obtain a nationally representative sample of UK adults. The percentage differences between the baseline and C19PRC-UKW4 quota bands for gender, age, and household income were calculated.
And fourth, the socio-demographic, mental health, and political characteristics of the C19PRC-UKW4 sample were assessed using counts and frequencies (weighted, where appropriate) and comparisons across the sample strands using chi-square tests. Figure 2 illustrates the outcome of recruitment of C19PRC-UKW4, Phase 1 and Phase 2.

| Outcome of Phase 1 recruitment and attrition analyses
The As presented in Table 2, compared to respondents who were lost to follow-up after baseline, respondents who participated in all four survey waves were characterised by being older in age (i.e. lower   Table 2).
Of the 853 additional respondents who were also eligible to be recontacted at C19PRC-UKW4 having only entered the panel at the previous wave (C19PRC-UKW3), 525 (61.5% recontact rate) were successfully followed up.

| Weight procedure Phase 1 longitudinal panel from baseline
The raking procedure successfully re-balanced the characteristics of responders at this fourth wave (N = 1271) to the baseline T A B L E 2 (Continued)

| Recruitment of new respondents: Phase 2
The median survey completion time for Phase 2 was 34 min 48 s. At Phase 2, 3073 adults were successfully engaged by Qualtrics partners and, following quality control checks, 1002 respondents were removed due to a failure to (1) complete the survey in full (n = 344); (2) satisfy the inclusion criteria (n = 185); (3) fulfil the legitimate respondent status (n = 50), or (4) satisfy country of residence sampling quotas (n = 420), or due to other minor technical errors (n = 3). This resulted in a Phase 2 sample of2071 1 , of which 292 respondents were recruited to 'top-up' quotas due to attrition in Phase 1, and the remaining 1779 constituted the UK-nation oversample. The 'top-up' quotas successfully re-balanced the C19PRC-UKW4 cross-sectional sample to be presentative of the UK adult population aged 18 years and older, with respect of age, gender, and household income (see Table S2).  In a final set of supplementary analyses (see Table S3), summary statistics for core political variables in the C19PRC-UIKW4 study stratified by country are presented to further highlight the potential for robust between-country socio-political analyses in the context of the COVID-19 pandemic using the C19PRC-UKW4 study data.

| DISCUSSION
By December 2020, four survey waves had been conducted for the C19PRC Study in the UK since its inception at the start of the COVID-19 pandemic in March 2020. The C19PRC Study comprises a diverse sample and contains a huge array of mental health, psychological, socio-economic, and political measures. The major objective of the C19PRC Study is to explain changes in UK adults' attitudes, experiences, and behaviours throughout the course of the COVID-19 pandemic using a range of innovative measures and approaches.
Our Consortium has previously contributed to the on-going debate about the strengths and limitations of probability versus non-probability survey designs during the pandemic (McBride et al., 2020. Here, we focus our efforts on unpacking the core credentials of this ongoing longitudinal panel study by winter 2020: (1) the successful re-interviewing approximately six-in-ten (62.8%) of baseline respondents at this fourth wave; (2) the low levels of attrition -only 15% of baseline respondents were completely lost to follow-up by this stage in the study; and (3) (Czeisler et al., 2021b). Moreover, complete dropout from the survey was high; 57.6% of baseline respondents completed only that single survey wave (Czeisler et al., 2021b). We concur with concerns raised by (Czeisler et al., 2021a) that problems relating to retention of respondents over time tempers optimism and confidence in findings emerging from some longitudinal mental health surveys conducted during the pandemic. Evidence indicates that attempts to re-engage 'temporary dropouts' is important to increase sample variability with respects to life changes (Müller & Castiglioni, 2020). Our Consortium has worked proactively at each survey wave post-baseline to minimise study limitations due to sample attrition. Given the unpredictable nature of the COVID-19 pandemic, and the likely impact that the pandemic has had on respondents' ability to participate at different waves, we continue to attempt to re-engage all baseline respondents at each planned survey wave (of which there are two due to take place before the current ESRC funding grant ends in November 2021) so that they continue to have an opportunity to reenter the panel at a suitable time for them.
Analyses presented here revealed that attrition in the C19PRC Study longitudinal panel has mostly been influenced by baseline socio-demographic characteristics as opposed to baseline experiences of mental health problems; that is, more women, younger adults, lower income earners, and those with dependent children have been lost-to-follow-up over the four waves of data collection.
Again, using the COPE Initiative as a comparison, attrition analyses in that study revealed that respondents who completed two or more of the four surveys administered between April and September 2020 had significantly lower prevalence estimates of adverse mental health  Consortium also has an auxiliary two-wave study of young people aged 13-24 years (N = 2002) which assess the impact of the pandemic on the health and relationships of this segment of the UK population (Levita et al., 2021a(Levita et al., , 2021b. Analyses of this important study data is on-going. In summary, and in recognition of these recruitment outcomes and attrition analyses, we seek to offer guidance as to how the dynamic C19PRC Study data can be used to address specific COVID-19 related research questions: (1) the amalgamation of data from specific strands of the survey (e.g. returning panels and 'topup' strand) produces a large, cross-sectional survey (N = 2088), which is nationally representative of UK adults aged 18 years and older with respect to gender, age, and household income, and can be used to address point-in-time COVID-19 related research questions; (2) analysis of data produced by the longitudinal panel returning from baseline (N = 1271) applying the C19PRC-UKW4 weighting variable provides an optimal vehicle for the pursuit of research questions relating to changes in multiple aspects of health, wellbeing, and life experiences over the first 9 months of the pandemic; and (3)   14 of 17to respond to, and cope with, the on-going psychological demands and the existential threat of COVID-19. The availability of geospatial data also facilitates the enrichment of individual survey responses via linkage to country-specific external data resources (e.g. measures of area-level deprivation; population density; availability of green spaces; area-level rates of COVID-19 testing/infection/ death, etc.).
Consistent with UKRI ESRC funding regulations, the C19PRC-UK study will be lodged with the UK Data Service before the end of 2021. In the interim, the baseline (C19PRC-UKW1; March 2020) and first and second follow-up (C19PRC-UKW2, April-May 2020; C19PRC-UKW3, July-August 2020) surveys are publicly available on the Open Science Framework (see https://osf.io/9emvp/). This will facilitate the public sharing of this rich data resource with scholars, academics, researchers, and stakeholders across a wide range of fields and disciplines.