Table of Contents
About the survey
What content does the COVID-19 survey cover?
The content of the questionnaires can be found in the content plan.
What questions were asked and how often?
Some core questions were asked in each wave of the survey, while some were asked every few waves, and others were asked only once. See Section 7 (Questionnaire Content) in the User Guide and the COVID-19 content plan.
Who can I contact if I have further questions about the survey?
Please email covid@understandingsociety.ac.uk if you have any questions or need help with the COVID-19 study. For help with the main survey data please contact user support.
How many waves of the survey were conducted and when?
The COVID-19 survey started in April 2020, with Wave 2 happening in May, Wave 3 in June, Wave 4 in July, Wave 5 in September, Wave 6 in November, Wave 7 in January 2021, Wave 8 in March and Wave 9 in September. Participants were also sent bio-kits containing Covid-19 antibody tests in April 2021.
A SDQ youth questionnaire was sent to young people aged 10-15 in Wave 4 in July 2020 and again in Wave 8 in March 2021 and a longer youth questionnaire was sent out in November 2020.
What survey data has been released so far?
We have released all nine waves of the COVID-19 survey. Data are available from the UK Data Service. We have released three special license access data files with Census 2011 Lower Layer Super Output Areas, Local Authority District identifiers, and School Codes. We have also released data from youth surveys in July, November 2020 and March 2021 and 2019 pre-pandemic data from the mainstage interviews.
What was the interview mode?
The interview mode for the regular surveys was web. In May and November 2020 we conducted a telephone interview for those households where there are no regular web users. In July and November 2020 and March 2021 we sent paper self-completion surveys to 10 to 15-year old sample members.
How long did the interview take?
The web surveys took approximately 20 minutes to complete.
Respondents
How were eligible sample members invited to participate?
Pre-notification letters were sent to the eligible sample members. These letters explained the purpose of the COVID-19 survey, what they would be asked to do and how they would be rewarded. They were also told that they would be sent invitation letters via email or SMS depending on the contact information held by the survey. See Section 6 (Fieldwork) of the User Guide for further details.
Does the COVID-19 survey include children?
All household members aged 16 and over were invited to take part in the COVID-19 survey. In July and November 2020 and in March 2021 10 -15 year olds were sent a short paper questionnaire. In July 2020 and March 2021 it included a SDQ (Strengths and Difficulties) questionnaire, while the November 2020 survey repeated the SDQ module along with questions about technology use and screen time, friendships and socialising, family, schoolwork, caring, health, nutrition, and exercise. We also asked parents for data on their children in the modules on home schooling, special educational needs and SDQ for under 10s, as well as parent-child relations for both within household families and separated families.
What was the incentive for completing this survey?
Adult sample members (16+) who completed the monthly survey were offered £2 for each monthly survey completed which they could exchange for a variety of gift cards, e-vouchers or a donation to charity. In the final wave all sample members were offered £10 for completing the survey.
Who was eligible for COVID-19 survey interviews?
All members aged 16 years and older, as of April 2020, of the main Understanding Society samples (GPS, EMBS, BHPS, IEMBS), who were in active households (i.e. households who have participated in at least one of the last two waves of data collection) but excluding those who were adamant refusals or mentally or physically unable to make an informed decision to take part, and those with unknown postal addresses or addresses abroad. See Section 5 (Sample) of the User Guide.
Who are considered to be non-respondents, partial respondents and full-respondents?
In the web surveys, those who did not participate in the interview or dropped out before completing the coronavirus illness module are coded as non-respondents. Those who dropped out before an important but not crucial module were coded as partial respondents (this cut-off point varies by the monthly survey). The remainder were full-respondents. See Table 6.2 in the User Guide where these cut-off points are specified. Note that the data for the few questions that the non-respondents completed before dropping out are not released. See Section 6 (Fieldwork) of the User Guide.
Using the COVID-19 data
How can I access these data?
These data can be accessed via the UK Data Service here.
Can I link the dataset to geographical data? How?
Datasets containing the Understanding Society COVID-19 geographical identifiers LSOA 2011 (SN 8663) and Local Authority District (SN 8664) are available from the UK Data Service (see Section 4 of the Covid-19 User Guide for access details and note that these files are classified as Special Licence). These datasets can be used to link the main Understanding Society COVID-19 dataset (SN 8644) to the LSOA 2011 or Local Authority District location for each respondent. Each dataset consists of a file for every wave. Each file contains PIDP (the linking variable) and the appropriate geographical identifier.
How do I read the questionnaire?
The questionnaire lists the questions in the order in which they were asked and for each question includes the question text, the source for the question and the universe which describes who were eligible for the question. See Section 8 (How to read the questionnaire) of the User Guide.
How can I find variables?
You can do a key word search for variables across all COVID-19 data files here. Alternatively, you can search by data files here. When you click on a variable from the search results, you will see a short description of the variable, the question text, the data file it is included in, and the frequency distribution across the different waves it was asked in (this will require you to click on the cascading wave tabs).
What is the naming convention for files?
There are two types of data files: those that include data pertaining to each monthly survey and those that include information that covers all monthly surveys. The wave specific filenames, have a prefix cW_ where W equals a, b, c,… depending on the monthly wave that the data file refers to. All data files that include data from the web surveys have a suffix _w. The cross wave file names start with x. For example, in the first monthly web survey in April 2020, one of the data file names was CA_INDRESP_W. Sections 10.2 (Naming convention of files) in the User Guide.
What is the naming convention for variables?
Variable names have a wave prefix cW_ where W equals a, b, c,… depending on the monthly wave survey it was asked in. Questions that are identical to the main annual survey questions retain the same variable names, but if the questions were modified then these variable names have a suffix _cv. For example, the variable CA_SCLONELY_CV in the datafile CA_INDRESP_W. The “Source” field in the questionnaire also explains the source for the question and whether it was modified. Some variables from the main annul surveys are included. These have the usual wave prefix reflecting the main annual survey they were taken from. E.g., J_IOUTCOME is the interview outcome from Wave 10.But if these variables are time-invariant or asked only once these do not have the wave prefix. E.g. PSU, STRATA, BIRTHY, RACEL_DV, BORNUK_DV. The variable PIDP, which is the unique cross-wave individual identifier, does not have a wave prefix and is included in all files. Sections 10.4 & 10.5 (Naming convention of variables, Naming of variables from multicode questions) in the User Guide.
What information do the cW_indresp_w and cW_indresp_t files contain?
These files include information asked in the individual web interviews (_w) of the telelphone interviews (_t). See Section11 (Contents of the INDRESP files) in the User Guide for further details.
What information does the XSAMPLE file contain?
The file XSAMPLE, includes information about every sample member who was eligible for the COVID-19 survey and invited to participate, (i.e. it includes information on non-respondents as well as partial and full-respondents). It includes some basic information about these sample members such as year of birth, sex, ethnic group, whether born in the UK, interview outcomes from Waves 9, 10 and 11 of the annual interviews. In addition, this file includes survey process information for the monthly surveys (and is updated every month to include new variables pertaining to that month) such as feed forward variables, interview outcomes, interview process variables. See Section 12 (contents of the sample file XSAMPLE) in the User Guide.
How are missing data coded?
All missing values are given a specific negative value to represent the reason for the data being missing. A value of -1 means the respondent reported “don’t know” while -2 means they chose “Prefer not to say”, but if they did not answer the question or either of the response options, “don’t know” and “Prefer not to say”, then the response is coded -9. If the respondent was not asked the question due to routing or because they broke off the survey prior to this question, then the variable is coded as -8. See sub-section 10.6 and Table 10.3 in the User Guide.
What are the income and earnings variables?
See sub-section 11.3 (Income and earnings variables) in the User Guide.
What variables are topcoded? How?
The End User License (EUL) version of the data includes top-coded versions of the income, earnings and household composition variables. Details of the topcoding rules are available in Section 11.5 (Top-coding of Variables) in the User Guide.
Can I link the data to the UKHLS main annual survey data? How?
Yes, the COVID-19 survey data can be linked to the main UKHLS annual survey data using the variable PIDP which is included in all data files. See sub-section 10.7 in the User Guide.
Can I link the data from the different COVID-19 monthly surveys? How?
Yes, the COVID-19 monthly survey data files can be linked to each other using the variable PIDP which is included in all data files. See sub-section 10.7 in the User Guide.
How can I identify the respondents in the same household?
As this survey was designed as an individual interview, there are no household identifiers available. But respondents from the same household can be approximately identified using information about whether they are living at the same address and if they were living in the same household at the last interview. The procedure explain in sub-section 11.8 in the User Guide.
What information is carried over from the main annual survey to the COVID-19 survey files?
Some basic background information about the sample member (such as year of birth, sex, ethnic group, whether born in UK), sample design variables (such as psu, strata) and interview outcomes and household identifiers at the last three main annual survey waves (i.e. 9 to 11) are included.
What weights are available?
See section 17 (Weighting and adjusting standard errors for sample design) in the User Guide.
How should I adjust for standard errors?
See section 17 (Weighting and adjusting standard errors for sample design) in the User Guide.
What does the file CA_SCHILD_W contain? How is this file structured?
Data on children were collected from parents on six occasions: in the April, July, September, November 2020, January and March 2021 web surveys. This information is contained in the SCHILD file. The April 2020 questionnaire was also used for the May 2020 telephone survey. The November 2020 questionnaire was used for both the web and phone surveys. Please note the telephone samples for children are small and should only be used in conjunction with the web survey for the same questionnaire. For more information on the SCHILD file please see the User Guide.
July and November 2020 and March 2021 Covid surveys included paper self-completion youth questionnaires sent out to sample members aged 10-15. The cd_youth_p and ch_youth_p files contain the answers to the paper self-completion Strengths and Difficulties Questionnaire (SDQ) that was sent out in July 2020 and in March 2021, while the cf_youth_p file contains the answers to the paper self-completion questionnaire that was sent out in November 2020 . The November survey repeated the SDQ module along with questions about technology use and screen time, friendships and socialising, family, schoolwork, caring, health, nutrition, and exercise. For more information on the YOUTH_P files see the User Guide.
Constructing a furlough related variable
Before you construct the furlough related variable that you are interested in please take look at the policy rules in place in a particular month and see if you can get the information you need to construct the variable that you are interested in from the information collected.
- In Waves 1-4 & 6 (telephone only),respondents in paid employment who have been never been furloughed are asked if they have been informed that they are on furlough: cW_furlough
- In Waves 7 & 8, respondents on paid employment are asked if they are on the Coronavirus Job Retention Scheme: cW_newfurlough
- In Waves 5 & 6, respondents who had said they had been on furlough and hadn’t said previously that they have stopped being on furlough were asked whether they were still on furlough: cW_stillfurl
- Options 1 & 2 for the cW_stillfurl question are only available in Wave 5 (Sep 2020) questionnaire and Wave 6 (Nov 2020) telephone questionnaire.
- In all waves, respondents on furlough can also be identified using these questions about reasons for hours change: cW_hrschange15 and cW_hrschange35.
- Some of these questions are not applicable for Welsh residents or Welsh residents who did not complete the interview in Welsh. Please refer to the questionnaire for the routing rules
Please note that the still furloughed question, cW_stillfurl, did not work for Wave 6 (Nov 2020). The reason is that we fielded it around the time a “Job Support Scheme” was going to be implemented to replace the JRS. But then that policy was scraped and so what we fielded doesn’t actually capture furloughing correctly in November (the response options in that month have a routing). The end result is that we don’t think the furloughing information can be easily constructed for Wave 6 (Nov 2020). However, there may be a creative way for data users to infer furloughing for the previously furloughed in November by looking at reported work hours. For example, those working for zero hours and employed are most likely on furlough, but people can work some small hours and still be furloughed.
Has Understanding Society published any findings from the COVID-19 survey?
Yes, you can find briefing notes and working papers produced by the Understanding Society team in the publications catalogue. The publications area also has information on research papers using the COVID-19 data.