OUP user menu

Alcohol Consumption Behaviours and Social Mobility in Men and Women of the Midspan Family Study

Carole L. Hart, George Davey Smith, Mark N. Upton, Graham C. M. Watt
DOI: http://dx.doi.org/10.1093/alcalc/agn125 332-336 First published online: 23 January 2009

Abstract

Aims: The aim of this study was to investigate relationships between alcohol consumption and social mobility in a cohort study in Scotland. Methods: 1040 sons and 1298 daughters aged 30–59 from 1477 families reported their alcohol consumption from which was derived: weekly units (1 UK unit being 8 g ethanol), exceeding daily or weekly limits, binge drinking and consuming alcohol on 5+ days per week. Own and father's social class were available enabling social mobility to be investigated. Results: More downwardly mobile men exceeded the weekly limit, the daily limit, were defined as binge drinkers and drank the most units per week of the four social mobility groups. Stable non-manual women were more likely to consume alcohol on 5+ days a week but very few were binge drinkers. Stable non-manual and upwardly mobile men and women were more likely to drink wine, and downwardly mobile men to drink beer. Conclusions: Downward mobility was associated with less favourable alcohol behaviours, especially in men. Wine consumption was more closely related to the social mobility groups than beer and spirits consumption. Drinking patterns could both influence and be influenced by social mobility.

Introduction

The effects of excessive alcohol consumption are of current concern with hospital admissions related specifically to alcohol more than doubling between 1995–1996 and 2006–2007 in England (Mayor, 2008). Deaths from alcohol-related causes have also increased and calls have been made to hasten measures to reduce the harm done by alcohol (Gilmore and Sheron, 2007). Recommendations for sensible drinking levels exist for men and women although older cohort studies may not have collected sufficient data to ascertain these recommendations. The current study asked detailed questions on alcohol consumption and also had information on social class at screening and prerecorded information about social class in childhood. As well as increases in alcohol consumption, changes from heavy industry to service industry have led to changes in society with large amounts of upward social mobility, partly due to the type of jobs available (Hanlon et al., 2006). However, some members of society have gone down the social scale. In this study, social mobility is examined in relation to exceeding recommended limits of alcohol consumption. Recent studies on socioeconomic position across the lifecourse and measures of alcohol consumption on hangovers (Batty et al., 2006), binge drinking (Yang et al., 2007) and exceeding daily and weekly limits and problem drinking (Batty et al., 2008) have not investigated as many measures of alcohol consumption as the current paper, nor looked at social mobility.

Methods

The Midspan Family study of adult offspring aged 30–59 took place between March and December 1996 (Upton et al., 2000; Hart et al., 2008). Both parents of participants had been part of the Renfrew/Paisley prospective cohort study between 1972 and 1976 (Hawthorne et al., 1995). The Renfrew/Paisley study consisted of 7049 men and 8353 women aged between 45 and 64 living in the towns of Renfrew and Paisley and included 4064 known married couples.

1040 sons and 1298 daughters from 1477 families participated in the 1996 study (response prevalence of 73% for individuals, 84% for families). They completed a questionnaire and attended a screening examination. The questionnaire included a comprehensive set of questions on alcohol consumption. Participants were first asked ‘Have you ever taken alcoholic drinks?’, and next ‘Do you take alcoholic drinks at present?’. If they answered ‘No’ to either question, they were directed to bypass the remaining alcohol questions. Drinkers were then asked to complete a 7 × 5 grid of alcoholic drinks consumed in the previous 7 days. For each day (Monday to Sunday), they entered the number of pints of (1) non-alcoholic beer, lager, etc., (2) low-alcohol beer, lager, etc., (3) beer, lager, shandy, cider, stout, etc., (4) the number of single glasses of whisky, vodka, gin, rum, etc. and (5) the number of single glasses of wine, sherry, Martini, port, etc. They were then asked if the last week was fairly typical of what they usually drank in a week and if not typical, whether they would normally drink more or less. Units of alcohol per day (and per week) (1 UK unit being 8 g ethanol) were calculated from the answers in the grid, assuming 0.3 units per pint of low alcohol beer, 2 units per pint of beer and 1 unit per glass of spirits or wine. Definitions of alcohol consumption were then calculated. These were the recommended daily limits of 4 units for men or 3 units for women and the recommended weekly limits of 21 units for men and 14 units for women (Department of Health, 1995). Binge drinking has been variously defined in studies (Herring et al., 2008) and in this paper the definition used was exceeding 8 units for men and 6 units for women on a day (double the recommended daily amount in one day) (InfoScotland, 2008). The final definition was whether alcohol was consumed on 5 or more days per week, in line with advice on having 2–3 alcohol-free days per week to allow the liver to repair itself (InfoScotland, 2008). The percentages consuming wine, beer and spirits and combinations were also calculated. There were no questions on alcohol consumption in the parental study.

Own social class for offspring was coded from the occupation given in the offspring questionnaire, using the Registrar General's Classification of Occupations (General Register Office, 1990). Women's own occupation was used unless they reported being housewives/homemakers and did not give a previous occupation, in which case their husband's (or father’s) occupation was used. Father's social class was obtained from reported occupation in the parental study and coded using the contemporaneous Registrar General's Classification of Occupations (General Register Office, 1966). Father's social class was missing for 20 offspring, 15 of whom had mother's social class substituted and the remaining five used offspring reports of father's occupation in the 1970s from the 1996 survey. Father's and own social class were also defined as manual (social class III manual, IV or V) or non-manual (social class I, II or III non-manual).

Social mobility was defined as stable non-manual (father's and own social class non-manual), upwardly mobile (father's social class manual and own social class non-manual), downwardly mobile (father's social class non-manual and own social class manual) and stable manual (father's and own social class manual).

Analyses were carried out using Stata release 9, adjusting for clustering of offspring within families. Means and proportions of alcohol behaviours by father's social class, own social class and social mobility were standardized by 5-year age groups. Men and women were analysed separately due to their differing alcohol consumption. The 27 men (2.6%) and 62 women (4.8%) who reported never having consumed alcohol and 64 men (6.2%) and 87 women (6.7%) who reported not currently drinking were included in the analyses with no units of consumption. Data on weekly units by social class, father's social class and social mobility were previously published in a paper on general risk factors (Hart et al., 2008) and have also been included here for completeness.

Results

Men with fathers in manual social classes generally exceeded the recommended limits more than men with fathers in non-manual social classes (Table 1). More women with fathers in non-manual social classes than in manual social classes consumed alcohol on 5 or more days of the week (21.7% versus 14.2%, P = 0.002) and women with fathers in non-manual social classes consumed more units per week than women with fathers in manual social classes. More men whose own occupations were manual exceeded the weekly limit, exceeded the daily limit and were binge drinkers compared with non-manual men (Table 2). Manual men also consumed more units of alcohol than non-manual men. More non-manual than manual social class women consumed alcohol on 5 or more days (19.2% versus 7.5%, P < 0.0001) and non-manual women consumed more units per week.

View this table:
Table 1

Alcohol behaviour in men and women from the Midspan Family study by father's social class

Father's social class
Non-manualManual
Men
N327713
 % consumed alcohol on 5 or more   days in a week35.0 (29.9–40.1)35.3 (31.7–38.9)
 % exceeded 21 units in a week33.8 (28.6–39.1)40.5 (36.7–44.2)
 % exceeded 4 units on at least     1 day56.7 (51.0–62.4)63.1 (59.3–66.9)
 % exceeded 8 units on at least     1 day32.2 (27.0–37.3)39.3 (35.5–43.1)
 Mean weekly units18.3 (16.2–20.4)20.3 (18.8–21.8)
Women
N398900
 % consumed alcohol on 5 or more     days in a week21.7 (17.6–25.8)14.2 (11.8–16.5)
 % exceeded 14 units in a week12.3 (9.2–15.4)9.5 (7.4–11.5)
 % exceeded 3 units on at least     1 day35.5 (30.7–40.2)34.3 (31.2–37.5)
 % exceeded 6 units on at least     1 day4.8 (2.7–7.0)6.6 (4.9–8.3)
 Mean weekly units7.3 (6.5–8.0)6.0 (5.5–6.4)
  • Percentages or means with 95% confidence intervals.

View this table:
Table 2

Alcohol behaviour in men and women from the Midspan Family study by own social class

Own social class
Non-manualManual
Men
N608432
 % consumed alcohol on 5 or more     days in a week37.7 (33.8–41.6)31.7 (27.2–36.1)
 % exceeded 21 units in a week34.0 (30.1–37.8)44.1 (39.4–48.8)
 % exceeded 4 units on at least     1 day57.1 (53.0–61.1)66.2 (61.6–70.9)
 % exceeded 8 units on at least      1 day31.0 (27.2–34.8)45.2 (40.4–50.0)
 Mean weekly units17.9 (16.5–19.4)22.0 (20.0–24.0)
Women
N1000298
 % consumed alcohol on 5 or more      days in a week19.2 (16.7–21.7)7.5 (4.5–10.5)
 % exceeded 14 units in a week10.1 (8.1–12.0)11.4 (7.8–15.0)
 % exceeded 3 units on at least     1 day34.1 (31.1–37.0)37.3 (31.8–42.8)
 % exceeded 6 units on at least     1 day5.4 (4.0–6.9)7.9 (4.8–11.1)
 Mean weekly units6.6 (6.2–7.0)5.6 (4.9–6.3)
  • Percentages or means with 95% confidence intervals.

There was a large amount of upward social mobility, with 35% of men and 50% of women being upwardly mobile (Table 3). Only 8% of men and 4% of women were downwardly mobile. Half of the downwardly mobile men exceeded the weekly limit, 71% exceeded the daily limit and half were defined as binge drinkers. Downwardly mobile men drank the most units per week of the four groups. Stable manual men had slightly lower proportions than the downwardly mobile men for these variables but higher proportions than the stable non-manual and upwardly mobile groups. Men in the stable non-manual and upwardly mobile groups had similar alcohol behaviours, although more men in the upwardly mobile group exceeded the weekly limit. All four groups had similar percentages of men who consumed alcohol on five or more days, with the upwardly mobile group being the highest.

View this table:
Table 3

Alcohol behaviour in men and women from the Midspan Family study by social mobility

Stable non-manualUpwardly mobileDownwardly mobileStable manual
Men
N24436483349
 % consumed alcohol on 5 or more days in a week35.8 (29.8–41.7)39.0 (33.9–44.2)33.2 (23.6–42.7)30.8 (25.8–35.8)
 % exceeded 21 units in a week28.8 (23.1–34.5)37.5 (32.4–42.6)51.0 (41.2–60.7)42.4 (37.1–47.7)
 % exceeded 4 units on at least 1 day52.7 (46.4–59.1)60.1 (54.8–65.3)70.8 (61.2–80.4)65.4 (60.2–70.6)
 % exceeded 8 units on at least 1 day26.9 (21.3–32.4)33.7 (28.7–38.8)49.6 (39.3–59.8)44.6 (39.2–50.0)
 Mean weekly units17.0 (14.6–19.4)18.5 (16.8–20.3)23.0 (19.4–26.5)21.7 (19.4–24.0)
Women
N34665452246
 % consumed alcohol on 5 or more days in a week23.3 (18.8–27.8)17.0 (14.0–19.9)9.9 (2.2–17.6)6.7 (3.6–9.7)
 % exceeded 14 units in a week11.2 (8.0–14.4)9.3 (6.9–11.8)17.0 (7.8–26.2)9.9 (6.2–13.7)
 % exceeded 3 units on at least 1 day35.2 (30.1–40.2)33.4 (29.7–37.0)39.0 (25.4–52.6)37.4 (31.4–43.4)
 % exceeded 6 units on at least 1 day4.0 (1.8–6.1)6.2 (4.3–8.1)7.3 (1.2 –13.3)8.0 (4.4–11.5)
 Mean weekly units7.3 (6.5–8.0)6.2 (5.7–6.7)7.3 (5.5–9.1)5.3 (4.5–6.1)
  • Percentages or means with 95% confidence intervals.

Nearly a quarter of the stable non-manual women consumed alcohol on 5 or more days, with progressively smaller proportions in the upwardly mobile, downwardly mobile and stable manual groups. Only 6.7% of women in the stable manual group drank on 5 or more days. In contrast, few stable non-manual women were binge drinkers (4%), with progressively larger proportions in the other three groups. Downwardly mobile women had the highest proportion exceeding the weekly limit and exceeding the daily limit. Stable non-manual and downwardly mobile women consumed the most alcohol per week and stable manual women consumed the least.

There were large differences between social mobility groups in the percentage of men who consumed wine (Table 4). Whilst 61% of stable non-manual men consumed wine, only 37% of downwardly mobile men and 26% of stable manual men did so. About half of the upwardly mobile men drank wine. More downwardly mobile than stable non-manual men drank beer. There were no differences between the four groups in consumption of spirits. Combining types of alcohol consumed, there were more stable non-manual men (and upwardly mobile men) than stable manual men who drank both wine and beer. This was influenced by wine consumption in the four social mobility groups. The same patterns (as for wine) were seen for the other groups involving wine. No differences were seen in the spirit and beer drinking category.

View this table:
Table 4

Type of alcohol consumed by men and women from the Midspan Family study by social mobility

Stable non-manualUpwardly mobileDownwardly mobileStable manual
Men
 % consumed wine60.9 (55.0–66.9)50.7 (45.5–55.9)36.8 (26.8–46.9)26.0 (21.2–30.8)
 % consumed beer68.6 (62.9–74.4)75.0 (70.5–79.5)83.4 (75.7–91.1)73.7 (69.0–78.4)
 % consumed spirits48.0 (41.9–54.1)45.0 (39.9–50.1)41.6 (30.7–52.4)48.4 (43.2–53.5)
 % consumed wine & beer44.1 (37.9–50.2)41.6 (36.5–46.6)32.0 (22.4–41.5)21.8 (17.3–26.3)
 % consumed spirits & beer38.3 (32.2–44.4)38.9 (33.8–44.0)37.6 (26.9–48.4)40.6 (35.4–45.8)
 % consumed wine & spirits31.4 (25.6–37.2)26.8 (22.1–31.5)17.1 (8.8–25.4)15.2 (11.3–19.1)
 % consumed wine, beer & spirits25.1 (19.6–30.5)23.8 (19.2–28.3)14.7 (7.0–22.4)13.5 (9.8–17.2)
Women
 % consumed wine67.2 (62.2–72.2)62.5 (58.7–66.3)47.4 (32.1–62.6)30.4 (24.8–36.1)
 % consumed beer16.7 (12.7–20.8)13.8 (11.0–16.6)15.8 (5.7–25.9)16.3 (11.5–21.2)
 % consumed spirits45.9 (40.5–51.2)46.6 (42.6–50.6)49.3 (35.7–62.8)44.9 (38.1–51.7)
 % consumed wine & beer12.1 (8.6–15.7)9.5 (7.2–11.8)6.7 (0.9–12.4)4.9 (2.1–7.7)
 % consumed spirits & beer8.6 (5.7–11.5)6.8 (4.8–8.8)12.3 (2.9–21.7)8.6 (4.8–12.3)
 % consumed wine & spirits32.9 (28.0–37.9)29.4 (25.8–33.0)23.0 (10.6–35.3)14.6 (10.1–19.0)
 % consumed wine, beer & spirits5.7 (3.3–8.1)4.7 (3.0–6.3)5.1 (0.1–10.1)2.2 (0.3–4.0)
  • Percentages with 95% confidence intervals.

More stable non-manual women and more upwardly mobile women consumed wine than stable manual women. There were no differences between the four social mobility groups in beer and spirits consumption. The combined groupings were strongly influenced by wine consumption. There were no differences for women between social mobility groups for percent consuming all three types of alcohol but the numbers were small.

The analyses were repeated using only the 1836 participants who reported the week to be typical of their usual consumption. The results had the same pattern for men and only slight differences for women for daily and binge drinking in the social mobility analysis (not shown).

Discussion

Many participants exceeded the recommended limits, particularly men. The effects of own social class were stronger than father's social class for alcohol behaviours. The direction was different for men and women, with men from manual classes but women from non-manual classes consuming more alcohol. Downwardly mobile men generally had the less favourable alcohol profile and consumed beer rather than wine. Stable non-manual women were more likely to drink on 5 or more days but less likely to binge drink. A strength of the study was that father's social class was reported accurately, as it was given by the parents rather than by recall of the offspring (Hart et al., 2008).

Differences in Behaviour Between Social Classes

The finding that the women from higher social classes drank more alcohol than women from lower social classes has also been seen in the Scottish Heart Health Study (Woodward et al., 1992), the Scottish Health Survey (Bromley et al., 2005) and in the Whitehall II study (by employment grade) (Britton and Marmot, 2004). This could be due to work-related pressures leading to professional women consuming alcohol like their male peers, or it could be due to financial considerations. However, a study of young adults in New Zealand showed that the frequency of drinking and amount drunk in lower occupation groups was higher than in higher occupation groups for women (Casswell et al., 2003).

Binge drinking was higher in manual than in non-manual men and women. There is similar evidence from other studies. In the 1958 British birth cohort study, there was a social gradient in binge drinking for men, with lower social class men having greater odds of binge drinking at ages 23, 33 and 42 years (Jefferis et al., 2007). For women, lower social class women were less likely to binge drink at age 23, but more likely to binge drink at ages 33 and 42. In a large prospective cohort study of adult men and women living in the Netherlands, the risk of very excessive drinking was higher for those with fathers from unskilled manual and lower professional/routine non-manual workers, compared to fathers who were high professionals (van de Mheen et al., 1998). A study of middle-aged men and women born in Aberdeen, Scotland, showed that higher adult social class was associated with a reduced risk of hangovers (Batty et al., 2006). A study of Russian men of similar age to the current study showed that hazardous drinking was related to socioeconomic variables, being especially high in men with least education and unemployed men (Leon et al., 2007; Tomkins et al., 2007). Disadvantaged circumstances throughout the lifecourse were associated with exceeding daily and weekly alcohol guidelines in a small cohort of men living in Scotland at age ∼59 (Batty et al., 2008). That study also looked at problem drinking as assessed by the CAGE questionnaire (Ewing, 1984) and found it to be associated with social disadvantage, with adult socioeconomic circumstances having stronger relationships than early life circumstances (Batty et al., 2008).

For both men and women, drinking on 5 or more days was higher in non-manual than in manual participants. It could be due to income or type of consumption (such as wine taken with meals). This was in contrast to the other measures that were less favourable in manual class men.

Social Mobility

Men in the stable non-manual group and the upwardly mobile group had similar alcohol behaviours and men in downwardly mobile and stable manual groups had similar alcohol behaviours. This suggests that adult social class is more important than childhood social class with respect to alcohol behaviours. In a study of Finnish men, Yang found that both early and later socioeconomic conditions contributed to binge drinking, but the effects of adult socioeconomic circumstances were stronger (Yang et al., 2007). In a French study of men (GAZEL), excessive alcohol consumption was defined as a daily consumption of at least four glasses in the previous week. Men who were not excessive alcohol drinkers in 1992 experienced more mobility in the time to 1999 from employee to middle executive than excessive drinkers (Ribet et al., 2003).

In a study of young Swedish men of mobility from father's to own social class, risky use of alcohol was highest in the stable manual group and next highest in the downwardly mobile group, although these two categories had similar proportions (Hemmingsson et al., 1999). The upwardly mobile and stable non-manual groups had similar, lower proportions. These patterns were similar to the current study for exceeding weekly amount and binge drinking.

Wine drinking is generally associated with affluence. Studies have shown that wine drinkers were better educated and wealthier compared with beer and spirits drinkers (Nielsen et al., 2004) and that wine drinking was associated with higher parental and own social status and more education in younger adults (Mortensen et al., 2001). In an older study of Scottish men screened in the 1970s, the percentage of wine drinkers was highest in the stable non-manual, was lower for the upwardly mobile, lower still for the downwardly mobile and the lowest for stable manual men (Hart et al., 1998). The direction was similar to the current study although the actual percentages were far lower in the older study (12% in stable non-manual, 6% in upwardly mobile, 5% in downwardly mobile, 3% in stable manual compared with 61%, 51%, 37% and 26%, respectively, for men in the current study), reflecting recent increases in wine consumption. Although consumption of spirits was similar across all social mobility groups, it was of interest that more downwardly men consumed beer than the other groups.

Alcohol consumption is a major public health issue (Gilmore and Sheron, 2007). This paper has identified groups with the most serious problems, especially downwardly mobile men, which should aid targeting of action. However, outcomes for alcohol consumption may parallel what has been seen with smoking, where the more affluent respond to health advice, causing relative health inequalities to grow rather than to shrink (Marmot and Wilkinson, 1999).

Limitations of the Study

No information was asked on place of drinking so it was not possible to tell if alcohol was consumed in drinking establishments or with meals at home. Information was not collected on alcohol addiction.

Generalizability

The participants were not representative of the general population as their inclusion was based on their parents’ participation in the Renfrew/Paisley study 20 years before. However, results were generally similar to those from the Scottish Health Survey 2003 and these were similar to those in England (Bromley et al., 2005). Women in the current study appeared to have lower prevalences of binge drinking than women of similar ages in the Scottish Health Survey.

Conclusions

There were high levels of excessive drinking in this population, as in others. Women from non-manual classes drank more than women from manual classes in contrast to men. Downwardly mobile men in particular had less favourable behaviours and consumed beer rather than wine. Drinking patterns could both influence and be influenced by social mobility.

Acknowledgments

Victor Hawthorne conducted the original Renfrew/Paisley study in the 1970s. Alex McConnachie was the statistician for the Midspan Family study in 1996. Ethical approval was obtained for the Midspan Family study from the Local Research Ethics Committees of the relevant Health Boards. The subjects gave informed consent. This study was funded by NHS Health Scotland.

References

View Abstract