What we found

We wanted to better understand how well the statistics on loneliness are meeting the needs of those that need to use them. To do this, we spoke with a broad range of key stakeholders and users of loneliness statistics. A full list can be found in Annex C. The following sections share what we found.

New statistics have been developed which are valuable to users

There is a lot that is positive to report, with each of the countries of the UK now including a measure of loneliness in at least one of their national statistical surveys.

The emergence and development of data and statistics on loneliness is a credit to the hard work of the responsible statisticians. Over the course of the COVID-19 pandemic, we have been impressed how statistics producers have adapted to meet new and increased societal needs for information.

The development of national loneliness strategies and the increased information demands arising from the COVID-19 pandemic has highlighted the importance of insightful statistics on loneliness. To meet this need producers of loneliness statistics have changed some of their data collections and outputs and invested resources into improving what data is already available.

The Opinions and Lifestyle Survey (OPN) was adapted by the ONS in March 2020, to become a weekly survey used to collect data on the impact of the coronavirus pandemic on day-to-day life in Great Britain. The survey includes a loneliness measure for England, Wales and Scotland.

Public Health England also produce the Public Health Outcomes framework, which draws data from the Active Lives Adult Survey, showing local level data on loneliness in England.

Recent ONS research, Mapping loneliness during the coronavirus pandemic, has linked OPN data with contextual data and statistics such as local crime, areas with ‘strong local business’ and employment rates at lower levels of geography to provide insight into the societal factors that may affect loneliness levels in different demographic groups during the COVID-19 pandemic. The ONS have been able to produce a map (Figure 1) of loneliness in England and some parts of Scotland and Wales to local authority level. The map represents the percentages of adults aged 16 years and over across Great Britain who were asked how often they felt lonely and responded with “often or always”. This graphic gives the user an interactive view into different levels of loneliness among different local authorities.

Figure 1: Loneliness rates by local authority

Source: Mapping loneliness during the coronavirus pandemic, (ONS)

The Scottish Government have commissioned a telephone survey to better understand the impact of the COVID-19 pandemic on the Scottish population. The impact of COVID-19 wellbeing in Scotland report includes the results of this survey, where respondents were asked how much time during the last week they felt lonely. This survey has been compared to the 2020 wellbeing survey and the 2018 Scottish Household survey. This release also breaks down loneliness into sub-groups, such as age, gender, and disability status.

Welsh Government have responded to changing data needs during the pandemic by adapting the National Survey for Wales, an annual face-to-face survey, to a monthly telephone survey from May 2020 onwards using a random selection of those who had taken part in the annual survey. Users can now access the National Health Survey for Wales on a monthly, quarterly, and annual basis. As these statistics include a question on loneliness, there are now more timely data for users on loneliness in Wales, reported to a national level. Additionally, the Public Health Wales Public Engagement Survey produced fourteen weekly reports during the pandemic that captured loneliness and isolation during the pandemic.

In Northern Ireland the NISRA Coronavirus (Covid-19) Opinion Survey measures the prevalence of loneliness in Northern Ireland and allows for comparisons with pre-pandemic data.  The survey is based on ONS’s Opinions and Lifestyle Survey.

Each producer has identified barriers in producing statistics during the pandemic. Due to lockdown rules, producers have had to amend surveys that were previously face-to-face and instead conduct them as telephone or web-based surveys. This means that the mode effect of their surveys has changed, with some producers reporting this has caused a drop in sample size. This has meant that, for some individual categories, the sample can now fall below the disclosure limits, resulting in a need to merge categories for future releases. To address these issues in their COVID-19 opinions survey, NISRA told us they will be producing a supplementary newsletter that includes high-level breakdowns in addition to their original survey release. Despite the issues faced by producers during this time, we are encouraged by the solutions presented and the commitment to still publishing statistics to meet user needs.

There are expert users who enhance the wider evidence base for loneliness

Where their needs are not currently being met by statistics, some users of loneliness statistics are carrying out their own research or turning to other sources of data.

Much of the non-official data on loneliness are being produced by charities, who are not therefore only users of loneliness statistics, but also important data producers. We heard that for some charities, collecting data can subsume a significant amount of their funding and resources.  Examples of data collections we heard about in our review include:

• The Royal British Legion who, after identifying a gap in data showing loneliness among veteran groups, undertook their own survey in 2005 and 2014 to understand levels of loneliness among these groups.
• The British Red Cross have carried out numerous studies into loneliness across the nation and during the pandemic.
• The Campaign to End loneliness founded the Research and Policy Hub in 2011 – an international network of research into loneliness. This group identifies current gaps and debates in current loneliness research and seeks to create new research opportunities to better understand loneliness sin society.
• The Campaign to End Loneliness commissioned Age UK to survey loneliness throughout the UK. This research concluded that, through better linkage across the nations, using current surveys could provide much of the information users currently need.

There are two distinct worlds in loneliness – a world of official statistics, and a world of un-official data, where charitable organisations are contributing to a wider, largely untapped and often disparate, evidence base on loneliness.

We heard that it is often the gaps in loneliness statistics which are driving the research carried out by expert users. These gaps serve to highlight where greater insight into who might be lonely, where they are lonely, and why they might be lonely is needed to identify lonely groups and to target interventions.

Users of the official figures told us that they believe the official statistics on loneliness are more trustworthy and of higher quality, meaning they can place more weight on them when decision-making. However, they also told us that when and where official statistics do not exist, they have no choice but to make decisions based off potentially less trustworthy and potentially lower quality data.

There is a need for more demographic information at lower levels of geography

Users we spoke to expressed frustrations at the current lack of statistics on loneliness at local levels.

We heard that local level data is needed to understand loneliness across communities, local authorities and by demographics, so loneliness can be tackled through both local policy implementation and grassroot, bottom-up initiatives.

Charities require data on their local community for a number of reasons – to understand which groups are at higher risk of loneliness, allocate their resources and target interventions where they are most needed. More widely, such data helps support a broader understanding of how levels of loneliness differ depending on locations.

Without these additional data, loneliness statistics are largely providing a count of who is lonely within the population.

We also spoke with users who highlighted a need for more data and statistics on the relationship between loneliness and various protected characteristics. This ranged from general demographic information, and more in-depth analysis of characteristics like ethnicity and disability, to more breakdowns within the statistical age bands.

One example of this need comes from Age Scotland, who told us that they need local level data to target their services where there is greater need, support funding applications and improve their understanding of the scale of loneliness throughout Scotland. Age Scotland currently use official statistics where these are available, but find they must supplement this with polling data and their own research. They told us that they would like to see official statistics available at local authority level and for data to be published in a more user-friendly format.

There will always be a need for local level data that cannot be fully met by official statistics and that needs to be collected by alternative methods, such as ad hoc telephone surveys and data sharing with other charities. These collection methods give organisations some of the richness they require in their data, but those we spoke to were aware that the coverage can be ‘patchy’ and felt that there was a role for official statistics to fill some these gaps currently met by locally collected data.

Users are not able to make UK comparisons using the current statistics

Currently there are no official UK wide statistics on loneliness, and the official statistics produced by each country are not directly comparable.

Different strategies on tackling loneliness have been developed in each UK nation which has resulted in variation in the way that loneliness is measured by nation. Furthermore, each country has developed different performance indicators, national targets, and have different reasons for requiring official statistics. This has meant that the statistical landscape is varied, with different frequencies for publications, different demographic groups covered, and varied disaggregation.

We heard that such comparisons are important for charities who operate on a UK scale. For example, Age UK operate across each devolved country, as well as within each country. To allocate their funding and resources equitably, they need to be able to compare which countries have higher levels of loneliness. Other charities, such as the British Red Cross, and academics, are interested in UK-wide comparisons for research purposes. They want to understand the different levels of loneliness in each country to better understand the differential impacts on levels of loneliness in a population. Some users are concerned about the harmonisation and standardisation between countries. They are generally unsure of the work each nation is doing to make their statistics more comparable.

We recognise the relevant and in depth work that the GSS Harmonisation Team have undertaken to work towards UK comparability, and the reasons that this is currently not possible. More information on this work can be found in Annex B. We would encourage the GSS Harmonisation Team to continue to consider how the barriers to harmonisation could be overcome in future.

As an alternative to national headlines measures, users are turning to different sources of information for UK comparisons. These included Understanding Society, a longitudinal household survey that captures life across the UK and the UK wide Millennium Cohort study which last ran in 2017 and asked whether respondents felt lonely in the past two weeks.

Key producers have, however, worked together to form an informal group on loneliness. This group includes England, Wales and Northern Ireland and aims to identify and fill information gaps on loneliness. In future the group will feed these findings into decisions over funding, linking with policy colleagues. We are encouraged by these efforts to work collaboratively and to identify and fill gaps in statistics. By sharing ideas and resources, statistics producers can work to both produce more valuable statistics and aid efforts to tackle loneliness.

More contextual information would make statistics most valuable to users

Users told us of the importance of looking at loneliness alongside other factors. As, without wider contextual information – such as experiences of poverty, mental health illnesses, and social isolation – the statistics were often seen by users as a headcount, offering little insight into the causes of loneliness.

It is acknowledged in the respective governments’ loneliness strategies that to effectively tackle loneliness requires it to be considered alongside other wider societal factors. Official statistics however are not currently fully meeting this need. In research, loneliness is often not considered in isolation but as part of a broader suite of factors around community cohesion, social capital and well-being. For example, The Carnegie Trust’s work on Gross Domestic Well Being (GDWe) puts loneliness in this broader context. In doing so, it can help users and researchers investigate correlations between loneliness and other factors. This may help their preventative measures become more effective if they know what groups are more likely to require their interventions.

The COVID-19 pandemic has highlighted a new need for this type of linked information. For example, academics and charities are expressing concern about ‘Covid loneliness,’ a new phrase used to describe those who have become lonely due to the unique situation the pandemic has caused. They told us the difference between loneliness and ‘Covid-loneliness’ is nuanced and may not be well reflected in official statistics. The pandemic has caused more people to become lonely, but once restrictions ease, their feelings of loneliness may disappear. Understanding these trends may help charities and service providers in focus their resources at similar times of need.

The differences between someone being lonely and being socially isolated is another important factor. Social isolation refers to when someone has no physical contact with others or society. A person may be socially isolated, but not lonely. As such, social isolation must be considered separately to loneliness by service providers and charities when they carry out interventions. Users found they could not explore this difference using current statistics. Measures to tackle social isolation are included in the England, Scotland and Welsh Strategies, but there are few official statistics on those who experience social isolation. The Scottish Household Survey does include a measure on social contact and isolation as part of their social capital indicator. This indicator can be broken down by geography and by personal characteristics and contextual information.

In addition, official statistics do not currently capture loneliness in a longitudinal way, for example how an individual’s experience of loneliness changes across a person’s lifespan. There are alternative studies, including the millennium cohort study and ELSA that do capture loneliness longitudinally, but these are not official statistics.