Mapping the loneliness measurement landscape

Many different approaches have been used to measure loneliness. These include both loneliness multi-item scales and single item measures. Some measures ask about loneliness directly, while others ask about emotions associated with loneliness from which loneliness is then inferred. There are advantages and disadvantages associated with each.

Office for National Statistics (ONS) undertook a programme of scoping work and consultation with experts on existing approaches to measuring loneliness. We developed a list of current measures of loneliness, which are either in use or have been used in the past. Members of the Loneliness Technical Advisory Group (TAG) provided invaluable support through meetings and correspondence to collate this information. Specifically, we gathered information on:

• findings from studies including loneliness measures (for example, surveys and evaluations)

• any cognitive question testing already carried out

• any results of pilot work

• types of surveys on which loneliness measures have been used (for example, longitudinal or cross-sectional)

• any adaptations made to the questions or response scales and the reasons why this was done

• any comparative data on findings using different single item measures, different scales for loneliness measurement and combinations of these

• any longitudinal data of loneliness using these measures

• mode effects (how the method of administering the survey may affect responses)

1.1 Direct measures of loneliness

As loneliness is a subjective emotional state, which we may each experience differently and which may vary over the life course, asking people directly is an important way of allowing them to express their own emotions and to capture self-perceived loneliness. Some existing surveys use a single item question on its own, while others include a single item question along with a loneliness scale comprising several questions exploring aspects of loneliness. There are several versions of single item questions currently in use, focusing on specific issues such as intensity or frequency of loneliness. Typically, respondents are required to define “loneliness” for themselves rather than being offered a definition. Table 1 provides examples of some direct measures of loneliness used in the UK.

Community Life Survey (CLS) and English Longitudinal Study of Ageing (ELSA) are both major surveys covering England and both include the question, “How often do you feel lonely?”. These are some of the largest surveys that currently collect data on loneliness in England and, as a result, there is more evidence for this specific wording than other versions of the single item questions.

Despite using the same question wording, they have different response categories, as shown in Table 1. Fewer response categories tend to be more beneficial for telephone data collection and for certain age groups as they are easier to remember. However, in terms of measuring the impact of interventions and changes over time, more response categories may be more useful in detecting changes over time.

Although a single-item measure might be beneficial in encouraging wider adoption and roll-out, posing minimal extra burden on respondents and survey costs, there are also some possible disadvantages. These include:

• the potential for under-reporting due to a perceived stigma attached to loneliness; this seems to be more evident for males, introducing a possible gender bias (Borys and Perlman, 1985)

• that respondents must define “loneliness” for themselves, which may lead to people describing different types of feelings and experiences as “loneliness”, some of which may not align with the definition of loneliness used for policy or analytical purposes

• the difficulty of capturing the severity of loneliness in a single item measure (as “severity” is defined as a combination of the frequency, intensity and duration of self-perceived loneliness)

1.2 Indirect measures of loneliness

Non-direct measurement of loneliness relies on researchers designing measures to capture specific aspects of the concept of loneliness and defining someone as more or less lonely depending on their answers to these questions. This means we are more likely to identify people who feel similar, but it is ultimately the designer of the measurement scale who decides whether what they are feeling is loneliness.

The review highlighted three indirect measures, which are either currently in use on an existing national survey in the UK, or which were assessed as meeting aspects of our ideal criteria well. These have been summarised in Table 2.

Loneliness scales are often used either in addition to or instead of single-item questions on loneliness. The loneliness scales vary in length, with longer and shorter versions available for some, for example, the UCLA scale and the De Jong Gierveld scale. The number of items in the scale is an important issue in this context, as the intention is for the loneliness measure to be used on national surveys. An important consideration has been to avoid over-burdening respondents, which could compromise response rates, and to keep costs and survey space within feasible limits.

After the initial scoping review of measures and further discussion with the Technical Advisory Group (TAG), the decision was taken to base our preliminary recommendations on existing measures rather than developing new ones. This was based on the following main reasons:

• the desire to see them widely and consistently used, which would be more likely if we could encourage more researchers to choose measures already established and in use

• the existence of measures that have produced helpful insights, are well-tested and have a track record in relation to how well they perform for different population groups and using different approaches to data collection

• the advantages associated with building on the existing evidence base to bring further comparable insights into loneliness quicker than would be the case if new measures were used

Based on these considerations, we decided to focus on the following measures for further testing:

• the direct question of loneliness currently in use on the Community Life Survey (CLS)

• the short form (three-item) of the UCLA loneliness scale currently in use on the English Longitudinal Study of Ageing (ELSA) and the Understanding Society study

The UCLA loneliness scale was designed to measure relational connectedness, social connectedness and self-perceived isolation. There are several versions including a 20-item and a three-item scale. Due to our requirement to use the measure on national surveys, only the three-item scale was considered. Although the UCLA scale uses negative wording (for example, focusing on a perceived lack of social connection), it is well-established internationally, aiding wider comparisons and suggesting translations are readily available if required. It has also been found to perform well both in self-completion questionnaires and in telephone interviews.

The use of both a direct question and a scale measure is the approach currently taken by ELSA and Understanding Society. This enables measurement of loneliness via a scale that has been assessed as valid and reliable, as well as allowing the respondent to report for themselves whether they feel lonely, providing further insight into the subjective feeling of loneliness for different people. Also, there is variation in how people understand the term “loneliness” and some people might be reluctant to admit to loneliness, and this might be particularly true of certain groups such as men. Those who are most lonely may find it upsetting to discuss their feelings and experiences of loneliness. A multi-item measure that does not mention loneliness directly can be helpful to address these issues.

We undertook a programme of work to test our preliminary recommended measures involving question testing on surveys across all age groups (from the age of 10 years) and cognitive question testing (see the Cognitive testing of loneliness questions and response options and Testing of loneliness questions in surveys chapters). All of the existing and new evidence was brought together to inform the recommended measures for loneliness, which have now been proposed as interim harmonised principles for use across the Government Statistical Service (GSS).

We will work with colleagues at the Department for Digital, Culture, Media and Sport (DCMS) and the GSS Harmonisation Team to encourage the roll-out of these indicators and suggest taking stock and making any refinements necessary within two years, after surveys using these measures have data available for analysis. After this, we will consider any further refinements needed to the indicators or guidance for their use before proposing the final Harmonised Principle (see the Recommended national indicators of loneliness chapter). To accompany the loneliness measurement recommendations, we have also developed a guidance report with suggestions for how to incorporate the measures in relevant surveys and interpret the results.