A new approach to measuring drinking cultures in Britain

Ally, A., Lovatt, M., Meier, P., Brennan, A. and Holmes, J. (2016), Developing a social practice-based typology of British drinking culture in 2009-2011: Implications for alcohol policy analysis. Addiction (early view)

Holmes, J. British drinking culture mixes moderation and excess, our new study shows. The Conversation. 20th April 2016.

When governments propose changes to alcohol policies, the announcement is often followed by public debate on the potential for the policy to change the country’s drinking culture. The UK Government’s 2012 Alcohol Strategy made changing the drinking culture a strategic policy aim (HM Government, 2012). However, specifying what the drinking culture is, what is problematic about it, what it should be changed to, what interventions might trigger such a change and whether success has been achieved have all been problematic topics in alcohol policy discourse and the research literature.

Existing literature has typologised drinking cultures along a series of key dimensions (Room and Mäkelä, 2000, Gordon et al., 2012) which risk conflating varied drinking behaviours into homogenous ‘national drinking cultures’. An alternative literature which characterises drinking cultures of different societal groups through detailed description, using qualitative methods, fails to adequately address the variety of drinking cultures which exist within and across nations.

We aimed to address this by focusing on one key manifestation of a nation’s drinking culture: drinking occasions. We developed typological models of drinking occasions and supplemented these with focus group research in order to gain greater understanding of how drinking occasions relate to drinkers’ broader social and cultural lives.

We used an embedded mixed methods approach with the qualitative components informing and validating the main quantitative study.

Quantitative component: This research uses data from the Alcovision survey, a commercial product collected by the market research company, Kantar Worldpanel. Alcovision is a monthly retrospective online diary survey with an annual representative sample of 30,000 individuals aged 18+ in Great Britain. The data used was collected in 2009-2011 and comprised 187,871 drinking occasions nested within 60,125 individuals (after excluding those not drinking in the diary week). We used the market segmentation technique Latent Class Analysis (LCA) to develop an eight-class typological model for the general population and also for demographic sub-groups defined by age, sex and socioeconomic status.

Qualitative component: we used focus groups to elicit descriptions of drinking occasions in drinkers’ own words. The focus groups were conducted in two stages. In Stage One we asked participants to describe their typical drinking occasions, the characteristics of these (for instance location, who they drank with) and how and why the characteristics differed across occasions. In Stage Two we used personas (personalised narratives about fictitious drinking occasions) derived from our typological analyses and asked participants whether the personas seemed realistic to them, and how they compared to their own drinking occasions. We also asked participants to describe the different factors which influenced their drinking occasions.

Population-level typology

A summary of the eight-class population level typology with illustrative quotations from focus group participants is shown in Table 1 below (a more detailed version can be found in the main report). The most common occasion type is light drinking at home with a partner which accounts for almost a fifth of occasions. Taking this occasion alongside light drinking at home with family and drinking at home alone mean this generally low risk, everyday kind of home drinking accounts for almost half (46%) of all occasions. Home drinking is, however, not always low risk. A heavy drinking version of these occasions (heavy drinking at home with a partner) accounts for 9.4% of occasions and always involves drinking at increasing or high levels. Similarly, the second most common occasion is get togethers at someone’s house which accounts for 14.4% of occasions. Focus group findings suggest that normative values associated with being invited to a friend’s house made bringing alcohol ‘a requirement’ and that the relative cheapness of drinking in a friend’s house compared to going out could result in greater quantities of alcohol being consumed. This was supported by the results of the typological analysis which showed almost half of these occasions (46%) involved drinking at increasing or high risk levels.

The greatest likelihood of high risk drinking (p=0.34) was found in a diverse set of occasions described as mixed location heavy drinking. These occasions are difficult to characterise but appear to encompass a range of drinking behaviours with evolving locations and participants. These may include nights out with pre-loading or drinking throughout the day in different locations. The ‘big night out’ with associated binge drinking may be split between this type of occasion and the going out with friends which does not include any off-trade drinking (i.e. no pre-loading) and accounts for 11.1% of all occasions

The final occasion type is classified as going out for a meal(?) with the question mark indicating that, as no data were available on whether the drinking occasion was in a restaurant or whether food was consumed, we have had to infer the nature of this type from probabilities suggesting it was particularly likely to occur at meal times.

Table 1: Summary of the population-level typology

The proportion of occasions within each type which involved low, increasing and high risk alcohol consumption is shown in Table 2, which shows that increasing and high risk drinking are mainly concentrated within four occasion types:

• Heavy drinking at home with a partner;
• Get togethers at someone’s house;
• Mixed location heavy drinking;
• Going out with friends.

Table 2: Proportion of occasions by consumption level and occasions type

Given the observed differences between sociodemographic groups in the distribution of their occasions across types, we analysed whether substantially different occasion typologies would be derived for each sociodemographic group. These were difficult to derive because the high volume of data made a systematic attempt to test how different the typologies may be for different sociodemographic groups impossible. Across all sociodemographic groups, three broad types of occasion were identified: Domestic get togethers with friends and family; Nights out; and Drinking at home. These varied to greater or lesser degrees by sociodemographic group.

Policy-related factors such as price, availability and health considerations influenced participants’ drinking occasions, but did so in ways that intersected with participants’ personal and social circumstances and values. Some participants explicitly stated that the relatively cheap cost of shop-bought alcohol compared to on-trade prices influenced them to pre-load at home before going out, but they also emphasised how such occasions provided valued opportunities to catch up with friends in a quiet home environment. Some participants also attributed the increase in home drinking by themselves and others to the widespread availability of alcohol in supermarkets and other shops, and the closure of local pubs. In the context of health, while some participants had heard of the UK lower risk drinking guidelines, few adhered to them. Many participants demonstrated an awareness of the potential health harms caused by alcohol, but this awareness was filtered through their own circumstances and health considerations.

The complexity of the interrelationship between different influencing factors has implications for policy interventions designed to influence the characteristics of drinking occasions and suggests that, although robust evidence exists on the likely effects of policies at the population level and, in some cases, on groups of the population, the effects on specific drinking occasions by specific groups are likely to be more complex to predict. If specific changes in drinking cultures are required, qualitative narratives such as those provided above may help inform understanding of which interventions are likely to achieve the desired effects and what barriers and facilitators may exist.

Policy implications

We argue that our typology of drinking occasions affords new opportunities for analysis of alcohol policy in the context of drinking cultures. Firstly, the typology provides an opportunity for more systematic consideration by policy makers and stakeholders in policy debate of what it is about the culture they wish to change. Secondly, the typology invites commentators to suggest what drinking culture they believe Britain should aspire to. If the problematic cultural features can be identified using the typology, a commentator should also be able to specify parameters for the model which would represent an acceptable drinking culture. Finally, the typology provides new opportunities for evidence-informed policy making and policy evaluation. While alcohol policy decisions are generally subject to evaluation against metrics of alcohol consumption and related harm among various groups within society, these metrics rarely take account of the complexity of drinking behaviours which policies are seeking to address. By segmenting the drinking occasions of different societal groups and the occasions on which different kinds of drinking take place, our typologies provide clear data to support understanding of the potential effectiveness of different policy options.

We argue that our typology of British drinking occasions substantially advances research on national drinking cultures, particularly in how culture manifests as behaviours with consequences for public health and social order. Along with Mustonen et al.’s typology of Finnish drinking occasions, (Mustonen et al., 2014) we demonstrate that national drinking cultures can be represented by a quantitative model with greater detail than has previously been achieved. However, as our model is of only one key and observable manifestation of drinking cultures – drinking occasions – future research supplementing our typology with data on the nature of intoxicated behaviour and social attitudes towards different occasion types may be beneficial for understanding and prioritising policy responses. Further research could focus on gaining a better understanding of each occasion type.

To further develop the typology for use in alcohol policy analysis, a number of steps would be beneficial. These include development of equivalent publicly available data, improved understanding of the relationship between policy and occasion types and between occasion types and alcohol-related harm and, finally, understanding of how occasion typologies vary across national contexts.

For further information please contact Dr John Holmes, Sheffield Alcohol Research Group, School of Health and Related Research, University of Sheffield.

GORDON, R., HEIM, D. & MACASKILL, S. 2012. Rethinking drinking cultures: A review of drinking cultures and a reconstructed dimensional approach. Public Health, 126, 3-11.

HM GOVERNMENT 2012. The Government’s Alcohol Strategy. London.

MUSTONEN, H., MÄKELÄ, P. & LINTONEN, T. 2014. Toward a typology of drinking occasions: Latent classes of an autumn week’s drinking occasions. Addiction Research and Theory, 22, 524-534.

ROOM, R. & MÄKELÄ, K. 2000. Typologies of the cultural position of drinking. Journal of Studies on Alcohol, 61, 475-483.