The apples and oranges of social relationships

I was watching two well-known politicians ‘argue’ on film the other day. When asked fairly specific questions, they took a very different approach to me. They had to. They were politicians after all, not scientists. They needed to answer swiftly, broadly and in a manner that carefully balanced truth with mass appeal.

Some people seem to have a natural ability to argue just about anything with confidence and zeal, without wasting time on the nitty-gritty. The reason I struggle to enjoy the company of these types over prolonged periods, entertaining as they are, is that I have a tendency to really need to break down exactly what it is that is being argued about (apples and oranges spring to mind) even if that takes a while. It’s not the most exciting approach I’ll accept, but it is the most scientific.

Our job as scientists, and in my case also a clinician, is not just to have a ‘broad stab’ at what the answer to a loosely framed query may be, but to ask precise questions and test hypotheses in the most robust, appropriate and meaningful way possible, whether it yields a sound bite or not.

Scientific discourse often struggles to fit into our modern sound bite culture.

Social relationships

My particular research interest is in social relationships and their influence on our mental health. I am currently focused on the experience of loneliness, a subjective unpleasant mismatch between the quality and quantity of social connectedness one desires and has. There is a strong, growing evidence base showing social relationships have a significant impact on health. But the literature on social relationships spans a wide range of concepts, some of which relate very closely to loneliness and others that do not: apples, oranges and the occasional avocado for good measure.

Example of social relationship concept	Brief definition
Loneliness	Subjective, distressing feeling when a person experiences a mismatch between the quality/quantity of relationships they have and what they would like.
Social network	Considered by many as a measure of the structural component of relationships including density, duration and reciprocity. The ‘how many? whom? when?’ questions.
Social isolation	Can refer to an objective measure of whether or not someone is living alone for example.
Social capital	A measure of how connected or involved one is with their local environment/ community.

Listed above are some examples of concepts one is likely to encounter in the social relationship literature….there are many many more!

The safest way to approach this complex and at times confusing area of research, is to step back and ask: what exactly are we measuring, and how do we do it? Dr Valtorta and colleagues attempt to answer this in their paper entitled ‘Loneliness, social isolation and social relationships: what are we measuring? A novel framework for classifying and comparing tools’ published in BMJ Open.

The authors aimed to:

Provide a summary of what measures people are using to study social relationships
Provide a ‘transparent’ classification system for these tools, to highlight what particular aspects of social relationships they covered.

This study classified measures in a way that “transcends disciplinary and conceptual boundaries, [so that they could] compare tools developed from different theoretical perspectives”.

Methods

From reading this, it appears they were conducting two distinct literature reviews:

One on how social relationships predict health and social care service use in the elderly,
and a second on how social relationships influence onset of strokes and heart disease.

The reviews were not conducted with the aim of identifying an exhaustive list of social relationship measures used across physical and mental health, but rather focused on the tools that happened to come up in the papers identified when conducting the two reviews above.

The search strategy for the first review included eight databases, and no language or time restrictions. The search terms could have been broader, to include social networks and capital, but the second literature search does provide what looks like a more thorough list of search terms (appendices 1 and 2 in paper) and covers 16 databases.

As a psychiatrist, I make note these were not mental health-focused reviews and, whilst likely to capture many relevant measures, the searches are limited to the outcomes they set out to study in their original review papers.

Results

The authors found 54 distinct instruments in total, measuring various aspects of social relationships.

After going through the questionnaires identified in the first review, they proposed classifying the tools based on two dimensions:

Tools measuring structure versus function of one’s relationships
The degree of subjectivity asked of respondents (ranging from a subjective quantitative estimate of people in your network through to a subjective appraisal of your feelings relating to social relationships)

Their summary shows, perhaps unsurprisingly, loneliness measures (e.g. UCLA Loneliness Scale, De Jong Gierveld Loneliness Scale) are at the most subjective, functional end, whereas specific social network scales are at the less subjective, structural end of the spectrum. However, they also demonstrate a number of measures fall in between the two, for example the Duke Social Support Index and Multidimensional Scale of Perceived Support (references all provided in the paper).

Conclusions

This is a helpful summary of a number of different tools that have been used in the study of social relationships (as they relate to health/social service use and cardiovascular outcomes). The reader can get an overview of how many items a tool has, and broadly what sort of questions they cover (structure versus function of one’s relationships, and ‘how subjective’ the questioning is).

Careful choice of measures is essential to further our understanding of the links between social relationships and health.

Strengths and limitations

This is a helpful paper, in that it contributes to the important discussion on which components of social relationships influence health in different ways. There are a few points to note, however:

This is not an exhaustive list of all tools available. As mentioned above, it only covers tools people have used in the studies included in their reviews. Studies focusing on mental health outcomes for example, are not considered.
Following on from the above, there are certainly relevant scales that have not come up in their searches, e.g. the Health and Lifestyles Survey Social Capital Questionnaire, or the 8- or 3-item UCLA scales that are fairly widely used. Again, researchers should not rely entirely on this list, as important appropriate tools may be missing, depending on their area of interest.
The study does not include tools used with young people or adults under 65, but we know there is a peak in loneliness during adolescence. The study focuses more on the elderly.
There is no mention of the psychometric properties of any of these tools, which should play a very important role in deciding whether or not one would use a particular measure (data on validity, reliability etc.), as well as which populations the measures have been developed or used in.
What would be even more helpful than identifying relevant measures, would be working towards specific definitions of key terms, drawing from the literature, aiming for a more universal social relationship ‘language’.

Summary

On the whole, this is a useful contribution to the social relationship literature; serving as a starting point for interested researchers looking for potential tools to use. It is not an exhaustive list, and does not focus particularly on mental illness, nor can it tell you anything about the quality of the individual measures. This is a growing area of research, so these are important issues to consider. If we are going to intervene in peoples’ social relationships, we need to know what exactly to intervene in.

To conclude with a fruit analogy: this paper goes some way to sorting your oranges and apples, but doesn’t account for the whole fruit basket and won’t tell you which are best to eat (no quality assessment).