Living Costs and Food Survey technical report: financial year ending March 2021

This report provides an update on the sampling, fieldwork and data processing for the Living Costs and Food Survey (LCF) for the financial year ending (FYE) 2021. The survey is undertaken by Social Survey Operations, which is part of the Surveys Directorate of the Office for National Statistics (ONS).

This report contains response information, questionnaire changes and new or changed methodology for FYE 2021. It does not describe methodology that has changed prior to FYE 2021. For changes prior to FYE 2021, users should refer to our Living Costs and Food Survey technical report: financial year ending March 2020 methodology. For a more in-depth explanation of LCF processes and methodology, users should refer to the Living Costs and Food Survey methodology. 

The purpose of this report is to update the FYE 2020 technical report and accompanies our Family spending in the UK: April 2020 to March 2021 bulletin. Section 4 of this report also provides an update to the LCF National Statistics Quality Review (NSQR). 

Alongside this report, we are publishing updated Living Costs and Food Survey: technical report data tables that provide information on response, characteristics of the sample, confidence intervals and interview metrics.

As shown in Table 4 of the accompanying dataset, the overall response rate for the Living Costs and Food Survey (LCF) in Great Britain was 24% in the financial year ending (FYE) 2021. Though this is a 16% decrease in comparison with FYE 2020, the number of households is comparable. This is because of the increase in sample size to accommodate the change to telephone data collection during the coronavirus (COVID-19) pandemic. A total of 22,516 addresses were sampled for the LCF in Great Britain. Of these, 4% did not contain a private household and were therefore classified as ineligible.

The number of returned cases for the survey was higher than the number of cases represented in the publication and technical tables. In total 6,842 UK responses were collected for the survey, of these, 5,400 cases were processed for the FYE 2021. This was because of significant delays in processing the data as a result of the changes made to collection during the pandemic. This caused a backlog of 1,442 responding cases that were not included in the final dataset.

A decision was taken, in consultation with stakeholders, to cease processing the FYE 2021 data because of the impact on the processing of the FYE 2022 data. From the remaining backlog, we prioritised coverage of underrepresented demographics for household composition and regional coverage before processing stopped.

Of the eligible sample, it was not possible to contact 46% of addresses, a further 21% refused to take part and 10% had another reason for non-response. Of the 5,170 responding households in Great Britain, 5,125 cooperated fully, meaning they completed both interview and diary sections of the survey.

In FYE 2021, partial responses accounted for 1% of all co-operating households. Of these 45 partials, 43 occurred because one or more adults in the household refused to keep the diary but were happy to take part in the interview (Table 6). 

Interviewers record the main reason why people refuse before or during an interview from a list of pre-coded answers. In FYE 2021, the two most commonly cited reasons for refusing to take part in the survey were (Table 10): 

• cannot be bothered (29%), which was the top reason cited, as in the previous year
• refusal to headquarters after interviewer's visit (13%), which is a level similar to the previous year

Preparatory work takes place in September 2019 for the April 2020 questionnaire changes. This allows us to plan and map questionnaire changes and seek stakeholder approval. However, the onset of the coronavirus (COVID-19) pandemic in early 2020 resulted in urgent changes being made to the Living Costs and Food Survey (LCF) questionnaire to accommodate income and expenditure pattern changes from respondents. Therefore, there were two types of changes to the LCF questionnaire: pre-planned changes for April 2020 (pre-coronavirus pandemic) and changes made in response to the pandemic.

Pre-planned questionnaire changes for April 2020 (pre-coronavirus pandemic)

The aims of these changes to reduce respondent and interviewer burden and the overall questionnaire length, and to improve data collection by refining the questionnaire’s flow and signals. This was achieved through removing questions or question blocks, adding certain questions, and other changes.

a) Blocks or questions removed

• A block of questions on pregnancy were removed. The Department for Environment, Food and Rural Affairs (DEFRA) originally requested these questions be included for women aged 16 to 55 years, in order to gain nutritional estimates. On further investigation, DEFRA found that they no longer required these data.

• Single telecommunication service provider name variables were removed. Questions asking respondents who provides their internet, television and telephone services were added last year to aid editors during imputation. Consultations around the usefulness of these questions revealed that they are not used.

• Home improvement and expenditure questions were removed. Both blocks had previously asked how an item was purchased 16 times. For example, for those who had recently renovated their home, they would have repeatedly been asked "How did you pay for x? 1. Loan, 2. Hire purchase, 3. Credit card, 4. Debit card, 5. Standing order/direct debit, 6. Bank transfer, 7. Cheque, 8. Cash, 9. Other method". This had been flagged as burdensome for both the interviewer and respondent, so for this reason, 16 examples of this question were removed.

• To ensure items purchased by loan and hire purchase are not double counted, an instruction has been added to the block's introduction for the interviewer.

• There were 12 instances within the LCF questionnaire where respondents were asked who is best able to answer questions about a payment. A focus group was held with interviewers who confirmed these questions were not needed and were rarely asked. These questions have been removed to reduce the interviewer and respondent burden.

b) Addition of questions

• Where questions had multiple answer options, including the option "other" and a follow up question to determine what "other" was, the "other" response options have been removed, as responses fitted into one of the other pre-determined answer options. This is found most notably at the furniture purchases question.

• "Other types of holidays" were made "spontaneous only" to encourage respondents to put answers into an appropriate category.

• A specification box was added to the computer ownership variable because of the number of "other" responses.

• In the absence of a known date and month of birth, interviewers were instructed to enter 15 June. If this date is entered, then an additional question is asked to clarify if this is the actual date of birth or an estimate. This question was added to aid data linkage.

• A question was added asking respondents if they have previously done any casual or holiday work. This was only asked of respondents who reported they had never worked. The question was added for user requirements.

c) Other changes

• Historically, expenditure questions asked an amount first and then the period the amount referred to, for example "How much did you pay?" and then "What period did this payment cover?". Testing has suggested switching the order results in a faster response time. This was rolled out across more blocks for further testing, including several income blocks.

• The voucher block was amended to remove all references to postal vouchers, as the default delivery method is now via email. Respondents can, however, receive postal vouchers if explicitly requested.

• Several changes were made to interviewer instructions and signals. A signal was altered to inform interviewers why loans need to be fully accounted for. Instruction was added to the signal at gross income, asking interviewers to collect unknown amounts during diary completion period. An erroneous signal was corrected that allowed respondents to mention they are making loan payments for a car but later claim they have no loans.

• The number of pay bands has decreased from 34 to 13. The previous pay bands used for respondents who do not give information for gross income were found to be outdated, so the bands were updated and aligned to modern income levels following consultation with the ONS Methods, Data and Research Division.

Changes made to the questionnaire in response to the pandemic

It was necessary to switch from face-to-face to telephone interviewing to ensure data collection activities were carried out in line with coronavirus restrictions. To make the telephone interview as user-friendly as possible and a suitable length, many blocks of questions and selected questions were removed. Before doing so, each block and question was assessed to check the overall impact of removing the block and question on downstream processes. As usual, feedback from interviewers and the edit team, analysis of questionnaire timings, analysis of data quality, and user requirements all informed the decisions made about the changes.

When converting to telephone interviews, the aims were to reduce the length of questionnaire to make the interview shorter, and to remove references to showcards from question wording to improve the flow. To aid this, three standard prompts were introduced to inform how these questions should be asked:

• prompt as necessary: allow respondent time to answer, then prompt with responses if required

• running prompt: pause after reading out all coding options with no pause between each option

• individual prompt: pause after each coding option to give respondent opportunity to answer

The main changes were:

a) Removal of blocks for telephone mode

The following blocks of questions were removed for telephone mode: country of birth, nationality, ethnicity (added back in mid-year), well-being, health, accommodation, economic status, second job hours, consumer durables, material deprivation, what the respondent can afford.

b) Addition of pandemic questions

Because of the pandemic, new questions were added to the questionnaire to capture:

• whether the household income changed compared with the start of the pandemic in the UK

• the reasons for the decrease or increase in household income

• whether people have had any time off work because of coronavirus, self-isolation, or any other illnesses, and whether they were paid during that time

• if people have received support from the various government schemes made available because of the pandemic, for example, the Coronavirus Job Retention Scheme (Furlough) or the Self-Employment Income Support Scheme

c) Topics reduced in length for telephone mode

Question blocks on education, length of residency, and rooms or business rooms were reduced in length to reduce respondent burden.

d) Diary changes

Prior to the pandemic, respondents completed a two-week paper expenditure diary recording all spending during that period. Since interviewers were unable to visit respondents' homes to collect paper diaries as a result of coronavirus restrictions imposed from April 2020, respondents were asked to provide copies of receipts (electronic or paper) for the two-week diary period. Interviewers recorded non-receipt-based expenditure via additional regular telephone calls during the two-week diary period.

e) Incentives

Prior to the pandemic, a trial was conducted to understand whether increasing the value of the conditional incentive would result in a higher response rate without negatively impacting data quality. The trial concluded that offering a £40 incentive increased response rates by 6 percentage points, compared with £20 or £30. In April 2020, an ongoing split sample trial was planned and implemented to test the impact on response rates of offering £40 versus £50 per full interview and diary completed. Given the unplanned mode change implemented in April 2020, the decision was made to move to a £50 voucher for all responding adults.

In response to the shift to telephone interviewing and the lower response rates achieved, unconditional incentives were increased from £5 to £10 from June to September 2020. However, this change did not positively affect response rates so in October 2020 the unconditional incentive reverted to £5.

The Living Costs and Food Survey (LCF) National Statistics Quality Review (NSQR), published in May 2016, identified 30 recommendations for the LCF Survey to ensure the survey remains fit for purpose. The Office for National Statistics (ONS) response to the NSQR was published in January 2017, describing at a high level how the recommendations will be addressed. This has been followed by annual updates published as part of our LCF technical report. Three approaches were identified for taking the work forward:

• the ONS Data Collection Transformation Programme (primarily through the transformation of ONS' household financial surveys)
• research projects taken forward by the LCF team
• projects led by other areas of ONS, in particular methodology and prices

Progress against each of the 30 NSQR recommendations is provided in Table 17 of the accompanying dataset. Where appropriate, recommendations have been combined, resulting in 25 recommendations to take forward. Of the recommendations, 20 have been completed. The remaining recommendations will be taken forward as part of the ONS Household Financial Statistics Transformation project. As such, this will be the final update on progress against the NSQR actions within the LCF technical report.