National Statistical

The Labour Force Survey has for many years been the cornerstone of ONS statistics on UK employment. While now complemented by other, often faster sources of information like real-time PAYE data, it remains the richest source of data on many aspects of the world of work and wider society. Darren Morgan discusses the challenges this survey faced during the pandemic and how we now plan to transform it. 

A major survey on a crucial topic 

The Labour Force Survey is our largest regular household survey outside of the decennial Census. It surveys thousands of households every quarter and tells us important information about whether people are in work (employed), out of work and looking for work (unemployed) or out of work and not looking for work (inactive). It also tells us much more about the people we speak to, such as the sectors they work in, their nationality and how many hours they work. Taken together they provide the important information needed to monitor changes in the labour market.  

Before the pandemic, we would write to addresses randomly selected from our database of all UK households, inviting people to take part and knocking on their doors to follow up if we didn’t hear back. During the pandemic, when face-to-face meetings became impossible, we had to rely on people responding to the letter. We saw pretty quickly that this was leading to bias in responses, with particular demographics such as the older population more likely to respond, while we were less likely to hear from people who rented their property. While we were able to take account of this new bias and swiftly reweight our figures, we knew we needed to speed up work already underway to improve the survey. 

Making the LFS fit for the future  

Today we have announced radical new plans to increase the sample size and improve the methods for collecting data, including the flexibility to quickly change the questions we ask so that we reflect the key needs of the day. We will also provide more robust insight on the detailed characteristics of those within and outside work.  

On collecting data, we will move to an on-line first approach, supported by telephone collection and ‘knock to nudge’ to ensure we can reach the widest possible pool of people, ensuring the Labour Force Survey can give the most accurate picture ever of changes in the UK’s labour market.  

More households, more inclusive coverage  

We are also increasing the sample size of the survey to 142,000 households per quarter as well as streamlining the questionnaire. This will vastly increase the number of those who complete the survey and make it more representative of the population as a whole. This, together with the addition of administrative data (another part of our labour market statistics transformation), will mean we can publish monthly estimates of labour market statistics in the future, providing users with more timely data. 

However, we are aware that such an ambitious set of improvements could lead to some discontinuities in results, although we won’t know this until the end of the year when we have data collected through all three methods. We will, of course, keep you informed and do everything we can to support users through the transition. 

What’s next?  

Today is just the announcement of the plans for this important new chapter in the Labour Force Survey’s long history. We will continue updating and engaging with our users throughout the development and begin to share analysis and insights from the transformed survey in early 2023. We understand these changes may have an impact on some of our users and we are carrying out a survey to better understand those impacts. 

These plans form one part of our wider transformation agenda across economic statistics that has been taking place over recent years, which has involved identifying and utilising new data sources (such as real time tax data) introducing new and transforming existing surveys and improving our methods to ensure we can produce the best estimates of our ever-changing economy and society.