Drug and Alcohol Review (i992) ii, 363-37 °

The standard drink and alcohol consumption

SUSAN j. CARRUTHERS* & COLIN W. BINNS t

* National Centre for Research into the Prevention of Drug Abuse and t Department of Public Health, Curtin University, Perth, Australia Abstract

The term 'standard drink' is commonly used when researchers collect alcohol consumption data and when educators create campaigns to encourage people to drink responsibly. However, little is known about community knowledge of the term 'standard drink' or what it represents in terms of what people are drinking in their own homes or on unlicensed premises. This study measured the amounts of a variety of alcoholic beverages that a sample of Perth metropolitan people use when drinking on unlicensed premises. The level of knowledge of the alcohol content of a variety of beverages and the knowledge of the term standard drink and what it represents in terms of commonly consumed beverages was also measured. The results showed that knowledge of the alcohol content of beverages was very poor. Knowledge of the term standard drink and what it represents in terms of absolute alcohol was also poor. Regarding the amounts of beverages poured, red and white wine, champagne and spirits are likely to be poured in amounts well in excess of a standard drink when people are drinking in their own homes. The results of this study have implications for the efficacy of educational campaigns designed to encourage safe and responsible drinking practices through the monitoring of personal intake. Before these can be effective, the terminology used must be familiar and well understood by the community towards which they are aimed. The results also discussed with respect to the design and administration of alcohol consumption questionnaires.

Key words: alcohol; consumption; knowledge; standard drink.

Introduction

In x988 , the National Health and Medical Re- search Council (NH & MRC) recommended the use of the term 'standard drink' to describe the consumption of alcohol for educational and re- search purposes [r]. Those in the field of health education will be familiar with the term 'standard drink' through its use in alcohol education programmes and alcohol consumption studies. The 'Drink Safe' campaign, launched in Western Australia in i988 , relies on the term 'standard

drink' when encouraging drinkers not to get 'stuck into alcohol' or to 'drink safe' and illustra- tions of 'standard drinks' are shown in the print media. This means that the success of this and similar education programmes is dependent on community knowledge of the term 'standard drink' and what it represents in terms of the wide variety of alcoholic beverages available to the public.

When measuring alcohol consumption by

Susan J. Carruthers, MPH, Research and Evaluation Officer, Healthway, PO Box ~u84, West Perth, Western Australia, 6872 , Australia, and Colin W. Binns, MBBS, MPH, FRACGP, FACOM, Head, Department of Public Health, Curtin University of Technology, Perth, Western Australia. Correspondence to Ms Carruthers.

363

364 Susan Z Carruthers ~ Colin IIC. Binns

means of community surveys, researchers in the past have often asssumed that consumers are using standard drink sizes when consuming alcohol. Even now researchers assume that consumers are aware of the quantity contained in glassware used to serve alcohol. Mulford and Miller stated that "it seems reasonable to assume that amongst drinkers there is consensus about the size of a bottle or glass of beer, a glass of wine or a drink of liquor" [2]. With the introduction of 'boutique beers' it is doubtful whether there is consensus about the meaning of a bottle of beer and glass sizes for wine and beer vary greatly. Beer is served in glass sizes ranging from I4o to 425 ml. The glass size most commonly served on licensed premises is the 285 ml 'middy' which gives the drinker approximately 250 ml of beer and is considered to be a 'standard serve'. Although a 'standard drink' of wine is io 5 ml, wine glass sizes on licensed premises vary from i5o to 250 ml and no particular size appears to be used regularly. There is obviously considerable variation in serving sizes on licensed premises. To add to the confusion there is not always agreement internationally when it comes to what a standard drink represents in terms of grams of absolute alcohol. Turner reviewed ia 5 published articles and found that the average number of grams of alcohol in a standard drink ranged from 9.2 to 23.0

g [3]" Alcoholic beverages are frequently consumed in

places other than licensed premises. In i985, the Australian Bureau of Statistics (ABS) reported that 64% of drinkers consume the majority of alcohol in private residences and it is likely this figure will have increased with the introduction of random breath testing in Western Australia in i989 [4]. At the time of this survey (x987-88) there was no published literature concerning standard drinks of alcohol, in Australia or overseas. However, since i989 a number of studies have been published. Stockwell et al. studied the effect of standard drink labelling on the ability of subjects to pour standard drinks or to estimate the number of standard drinks in alcohol containers I5,6]. Turner re- viewed I25 studies relating various indices of physical harm to different levels of alcohol consumption and stated " . . . it is necessary to take care in interpreting published data about levels of alcohol consumption related to harm" and "Re- searchers should define what they mean by a 'standard drink' in their c o u n t r y . . . " [3]" Miller et

al. made a 'plea for consistency' since variability in the reporting of alcohol content has led to considerable confusion in the field of alcohol studies [7]. Although these recently published articles deal with 'standard drinks' none actually measure the amounts of each alcoholic beverage the community are pouring in their own homes.

The aims of this study were to measure the amounts people pour in their own homes when consuming a variety of alcoholic beverages and to assess community knowledge of the term standard drink and the alcohol content of beverages commonly consumed in Australia.

Method

Sample

A randomly selected cluster sample technique was used. Eligible subjects were those aged between i8 and 45 years who had had an alcoholic drink in the past month. From a sample of 1866 households, contact was made with 1159 (62%). There were no eligible subjects in 528 households and 278 refusals. The adjusted response rate was 56% resulting in a sample of 356 (200 females). The data were collected between August and October I987.

Questionnaire

A questionnaire was designed to enable the quantities of a variety of alcoholic beverages to be measured. The questionnaire was pre-tested on a group of first-year students from Curtin University and modifications were made so it could be used in the subjects' homes with the minimum amount of inconvenience to participants and interviewers. The questionnaire consisted of 26 questions designed to collect demographic information and to assess subjects' knowledge of the alcohol content of a variety of alcoholic beverages. Both multiple-choice and open-ended questions were included.

Pourino~ test

To measure pouring habits respondents were requested to pour samples of a list of x2 alcoholic beverages. In the pre-test, actual beverages or imitations were used and glasses were provided by

the interviewers. Water was used as a substitute in the community study due to problems involved with interviewers being required to carry r2 sample bottles from house to house. Participants were requested to use their own glasses so that they would not be influenced by the type and size of glass provided by the interviewer. Amounts poured were measured using a graduated measuring cylinder.

.,.L/na/.yses

Data analysis was carried out using the Statistical Package for Social Sciences (SPSS-x). Statistical procedures used including chi-square, Student t- tests, Pearson's correlation and analysis of variance (one- and two-way).

Results

The 'standard drink'

Approximately 50% of all males and females stated

The standard drink 365

that they had heard of the term 'standard drink' (Table i). The remainder were either not sure or had definitely not heard of the term. Television, newspapers and magazines were the most com- monly cited sources o f information about 'stan- dard drinks' , followed by the Police and Heal th Departments. Eighty-five per cent of respondents who had heard o f the term cited these agencies as their source of knowledge. Those that had heard of the term 'standard drink' were asked how many grams of absolute alcohol were contained in one drink. Six per cent of males and i% of females gave the correct response of between 8 and io g (mean = 60.6 g, range i - 3oo g).

Percentage alcohol in alcoholic beverages

Subjects were asked to state the percentage alcohol in a range of alcoholic beverages. Re- sponses were accepted as correct i f they were within lO% of the correct answer. A 'don ' t know' category was included for this question to elimi-

Table I, Male and female knowledge of the term 'standard drink' by age group

Age group Males (%0) Females (%) (n = I56) (n = ~oo)

Heard of the Not heard Heard of the Not heard term or not sure term or not sure

I8-24 years 44.7 55.3 35.9 64"I 25-34 years 52.x 47.9 47 .I 52.9 35-44 years 4L4 47-9 39 .6 60"4 45 years 66.7 33.3 42.9 57 'I All ages 46.0 54-0 42-5 49.5

Table 2. Distribution of knowledge of the percentage alcohol (% v/v) of alcoholic beverages

Drink Correct Incorrect Don't know

Males Females Males Females Males Females

Beer (3.5 %) 55 9 I4 io 31 8i Beer (4.9 %) 76 22 5 8 19 7 ° Wine (I 2 0/0) 28 7 15 13 56 8I Spirits (38.5 %) 3 ° 5 22 15 47 80 Fortified wines (i8%) x 9 6 17 6 64 88 Wine cooler (3.7 %) 24 II 15 ii 6i 77 Liqueur (280/6) 4 3 28 I2 68 84

366 Susan Z Carruthers ~f Colin l~. Binns

nate guessing. Males were significantly less likely to tick the 'don' t know' category (p < o.oox) and significantly more likely to give the correct response (p < o.ooI). However, the overall knowl- edge of the percentage alcohol in alcoholic beverages was poor (Table 2).

Alcoholic beverages at home

Sixteen subjects (seven males and nine females) stated they did not consume alcohol at home. For the remainder, beer was the most frequently consumed beverage for males and wine the most frequent for females. Subjects who drank alcohol at home were asked to pour samples of commonly consumed alcoholic beverages. Subjects were asked to select the glass they would usually use for red wine. Having selected a glass they were passed a carafe of 'red wine' (water) and instructed to "pour a glass of red wine please". The same procedure was followed for all beverages. The beverages were red and white wine, champagne, fortified wine (port), full strength beer, spirits (gin) and a liqueur. Results are presented in Table 3 as 95 % confidence intervals for males and females in three age groups. Males aged between i8 and 24 years and between 25 and 34 years poured significantly more full strength beer than their females counterparts ( p < o . o 2 and p = 0 . 0 4 ,

respectively). Males in all age groups poured significantly more spirits than the females (p = o.o~).

Discussion

The results of this study support the hypothesis that alcohol consumers will not be familiar with the term 'standard drink'. Approximately half of the subjects had heard of the term 'standard drink' while the remainder were not sure or had never heard of the term. The most commonly quoted sources of knowledge were newspapers and televi- sion, suggesting that these media avenues should be used to promote health messages. The 'Drink Safe' campaign, launched in Western Australia in early i988 , has used these media sources to promote messages about standard drinks and safe drinking levels. This campaign is a long-term state-wide alcohol education programme funded by the West Australian Government and orga- nized by the Health Promotions Branch of the Health Department of Western Australia. One of

the aims of the campaign is to "encourage responsible use of alcohol by those who choose to drink it". The campaign uses a variety of strategies including the use of mass media and a community- based information approach. While this study has highlighted the fact that newspapers and television are useful for promoting messages, it also high- lights the fact that if only half of the community are aware of the term standard drink the message is probably not being understood.

The Royal Automobile Club of WA, the WA Police Force and the Health Department also use the term 'standard drink' in some of their educational material [8-xi]. A survey completed by the Police Department in i988 found that 37 % of drivers over-estimated the number of drinks they could consume before reaching the 0.08 blood alcohol level (legal blood alcohol driving limit in Western Australia) Ix2]. The standard drink sizes used by the Police Department at that time were not the same as those used by the 'Drink Safe' campaign. Taking the discrepancies in drink sizes into account the number of those under-estimating the amount of alcohol they consume may well have been more than 37 % .

In retrospect, it was unrealistic to expect members of the community to know how much absolute alcohol was contained in a 'standard drink'. This was evident by the response to this question. Only I4 (4%) respondents gave a correct response to this question. In future studies of this type it would be more appropriate to ask subjects how much of each beverage is equal to a 'standard serve'. However, Stockwell et al. asked subjects if they had heard of the term 'standard drink' and if they had, could they explain what it meant in terms of the quantity of their usual beverage or in terms of all alcoholic beverages. While 67% of subjects had heard of the term only x8% were able to give a correct definition [6].

In i978 , the Australian Brewers' Association stated "a source of great confusion in the community at large is the variety of measures used to express the alcohol contents of beverages" [i3]. Some of the measures then used were:

percentage alcohol by volume (%v/v) percentage alcohol by weight (%w/v) grams of alcohol per ioo ml of

beverage (%w/v) percentage of proof spirit (o) The Association then recommended that the

The standard drink

Table 3- 95% Gonfldence intervals for amounts of beverages poured by males and females (proportion of io g standard drink)

367

Beverage Sex 1 8 - 2 4 years 25-34 ),ears 35-44 years

R e d w i n e M 1.1-1. 4 1 .2-I . 4 1 .2-1 .4

F 1.1-1. 4 1.2-1. 3 1.o-1.2

W h i t e w ine M i . I - i . 5 1.2-1. 5 1 . I - I . 3

F 1.2-1. 5 1.2-1. 3 1 . o - i . 2

C h a m p a g n e M 1.1-1.4 1.2-1. 5 I . I - I . 4

F 1 .2-I . 5 1 .2-1.4 L I - I . 3

For t i f i ed w i n e M 1.o-1.2 o . 9 - 1 . o o . 9 - 1 . o

F o.7-1.o o.7-o.8 0.7-0. 9 Beer (4.9 %) M 1.1-I. 3 1.2-1.4 1.2-I. 3

F o.8- i . i i .o- i .2 1.o-1.2

Spirits M I.I--I.8 1.1--1. 5 I.O--1. 3 F 1.I-1. 9 1.o-1. 3 1.o-I.I

Liqueur s M o . 7 - 1 . o o . 6 - I . o o.7-1.1 F o.7-1.o o.7-o. 9 o.6-o.8

alcohol content of beverages should be expressed by one simply understood measure, namely per- centage of alcohol by volume (°/0v/v). This form of labelling is printed on all containers of alcohol but from the results of this study it is obvious that most of the community is still confused. With the exception of the alcohol content of full strength beer, subjects' knowledge of the alcohol content of a range of alcoholic beverages was extremely poor. This is illustrated in Table 2. Since the term 'standard drink' has been adopted by the NH & MRC and is frequently used in alcohol education programmes, it would be of more benefit to the public if the percentage alcohol by volume statement on alcohol containers was replaced with the quantity of a particular beverage representing a 'standard drink'. This suggestion is supported by the findings of Stockwell et aL who found that wine and beer drinkers expressed a preference for the alcohol content of beverages to be illustrated in terms of standard drinks rather than grams [5]" Suggested messages for each beverage are dis- played in Table 4-

Pouring alcoholic beverages in private residences

The majority of subjects poured standard serves or less of the less frequently consumed alcoholic

b¢verages, such as fortified wines and liqueurs and the lower alcohol beverages, wine coolers and low alcohol beers. However, the more commonly consumed beverages such as full strength beer, wine and spirits were frequently poured in amounts exceeding a standard serve (see Table 3). Before these results can be extrapolated to the general population a number of issues need to be addressed. First is the use of water as a pouring medium in place of the actual alcoholic beverage or imitation beverage. It is not known whether using water as a pouring medium has any effect on the amount poured by respondents. The decision to use water as a pouring medium came after consultation with colleagues, interviewers and students involved in the pilot study. The students did not feel that using water in place of the alcoholic beverages would have any effect on the amount of each beverage they poured. It was also thought that using the participants' own glasses would cancel any bias introduced by using water as a pouring medium. The second is the effect of being under observation. This may have had an inhibitory effect on respondents. Future studies of this type will have to address these issues. Comparisons between the amount of red and white wine poured by students in the pilot study and respondents in the community study were not

368 Susan Z Carruthers & Colin 1~. Binns

Table 4. Suggested labelling using standard serving sizes as an alternative to the percentage alcohol by volume

Beverage Container size Suggested message

Beer (5%) 375 ml can or stubbie 75 ° ml bottle

Wine 75 ° ml bottle

Wine 2 or 4 litre casks

Spirits All sizes

Fortified wines All sizes

This can contains I~ standard drinks This bottle contains 3 standard drinks This bottle contains approximately 7½ standard drinks ioo ml (4 oz) of this wine is equal to one standard drink 3 ° ml (x oz or x nip) of this spirit is equal to one standard drink 7 ° m l (2 oz) of this wine is equal to one standard drink

significantly different. Hence, for the purposes of this exploratory study we have assumed that using water did not have significant effect on respon- dents' 'usual' pouring practices.

Most articles reporting the results of alcohol consumption studies give examples of the question- naire design but not all give examples of the method used to quantify the amount of each beverage contained in a serve. Of the three reports found that do give details of quantification, two assumed that the respondent was aware of the size glass he or she uses and the third included a visual prompt card showing a variety of glasses compared to a can containing 375 ml [I4-I6J. Unfortunately the latter study only collected data on women so no comparison using this and other methods could be made. The last Australian Bureau of Statistics alcohol consumption study in Western Australian in x985 also assumed that respondents were aware of the quantity of beverages contained in each serve [4]. Although the accuracy and comparability of questionnaires and data collection techniques in alcohol consumption studies has improved, the method is still far from perfect. Comparisons of reported alcohol consumption figures and apparent consumption figures (compiled by the ABS) continue to show that reported figures only account for up to 50% of apparent figures (Table 5) [i6]. If consumers are not familiar with the concept of a 'standard drink' or what it represents in terms of a variety of alcoholic beverages it is highly likely that inappropriate serving sizes are being used by some. This practice may have a substantial impact on the accuracy of alcohol consumption studies where researchers are assuming that respondents are

aware of the amounts of each beverage contained in a 'standard serve' and are in fact pouring similar amounts in their own homes. A hypothetical case is used to illustrate the possible impact of the under- estimation of wine on reported alcohol consump- tion.

(x) A wine consumer reports the consumption of alcohol on 3 days per week with three drinks per occasion.

(2) A standard serving size of i2o ml is assumed and alcohol intake is recorded as nine drinks per week.

(3) The quantity per serve is actually i5o ml. Therefore consumption should have been recorded as i i drinks per week.

This represents a 250/o deficit in recorded alco- hol consumption. If this is the case, present recommended safe dr inking levels could be de- scribed as conservative. These levels are deter- mined by relating various indices of harm to reported alcohol intakes. If a significant propor- tion of alcohol intakes are 25% more than reported, then the dose/harm relationship will be altered. More studies examining the methodolo- gies of alcohol consumption studies are needed to ensure that alcohol data in the future are as accurate as possible.

When considering the health of an individual, it is the total amount of alcohol consumed that is important rather than the amount served per drink. The N H & MRC state that men should not exceed four standard drinks per day if dr inking regularly and women should not exceed two standard drinks. Therefore the community need to

The standard drink

Table 5. Comparison of apparent and reported consumption by type of alcoholic beverage

369

Beverage

Apparent Reported % of apparent consumption consumption accounted for

('ooo kg AA a) ('ooo kg AA a) by reported

Beer 72123 31378 44 Table wines 12198 5267 43 Fortified wines 7406 1743 24 Spirits 12 652 4II7 33

Total 1o4 289 42 505 41

a AA = absolute alcohol.

be aware of what a standard drink is before they can benefit from this information even if they do not pour the correct amount per serve.

Conclusions and recommendat ions

The subjects in this study were not familiar with the term 'standard drink' or the alcohol content of the beverages they are consuming. While it appears that some beverages are being poured in amounts similar to a 'standard serve', the more popular ones, beer, wine and spirits are frequently being poured in amounts well in excess of the standard. This may have an effect on the accuracy of alcohol consumption studies and be partially responsible for the discrepancy between reported and apparent alcohol consumption figures. While errors such as these will never be removed completely, researchers need to be aware that they exist and improve recording methods to increase the accuracy of alcohol consumption data.

The 'Drink Safe' campaign has now been running for 4.5 years in Western Australia. It would be interesting to reassess community knowledge of the term 'standard drink' as one method of measuring any increase in community knowledge as a result of the campaign. Because knowledge of the percentage alcohol in beverages was so poor it is recommended that it be replaced by a message concerning 'standard drink' serving sizes.

Rcfcrcnccs

[1] National Health and Medical Research Council. Is there a safe level of daily consumption of alcohol for men and women? Canberra: Australian Government Publishing Service, 1987.

[2] Mulford HA, Miller DE. Drinking in Iowa: the extent of the drinking in selected socio-cultural categories. J Stud Ale 196o;21:26-39.

[3] Turner C. How much alcohol is in a 'standard drink'? Analysis of 125 studies. Br J Addict I99o;I85(9):H7I-5.

[4] Australian Bureau of Statistics. Alcohol consump- tion patterns in Western Australia, October 1985, Cat. No. 43o2.5, October I986.

[5] Stockwell T, Blaze-Temple D, Walker C. The effect of 'standard drink' labelling on the ability of drinkers to pour a 'standard drink' Aust j Public Hlth 199i;i5(1):56-63.

[6] Stockwell T, Blaze-Temple D, Walker C. A test of the proposed to label containers of alcoholic drink with alcohol content in 'standard drinks'. Health Promotion International, November 199o. x99t~6:2o7-I 5.

[7] Miller W, Heather N, Hall W. Calculating standard drink units: an international comparison. Br J Addict I99t;86:43- 7.

[8] Royal Automobile Club of Western Australia. Plan alternatives to driving after drinking. The Road Patrol 1987;Dec/Jan:2-6.

[9] Police Department of Western Australia. Drive to stay alive. 1984:42.

[io] Health Promotion Services. Alcohol awareness: women and alcohol. Perth: Health Department of Western Australia, 1986.

[ii] Drug Education Centre. Alcohol: drug informa- tion series for health professionals. No. i. Perth: West Australia Drug and Alcohol Authority, 1986.

[12] Kennedy P. Many are ignorant of drink limits --survey. West Australian I988;Sept 22:12.

[13] The Australian Associated Brewers. Methods of measurement--an industry perspective. Commun Hlth Stud 1978;II(3):12o-2.

[I4] Miltwood JE, Mackay AM. Measurement of alcohol consumption in the Australian population. Commun Hlth Stud 1978;I1(3):123-32.

37 ° Susan Z Carruthers ~ Colin I~. Binns

[15] Polich JM, Armour DJ, Bariken HB. The course of alcoholism. New York: John Wiley, 1978.

[i6] Corti W. Women's alcohol consumption in West- ern Australia. National Centre for Research into the Prevention of Drug Abuse. Perth: Curtin University of Technology, September i988.

[17] Australian Bureau of Statistics. Apparent con- sumption of selected foodstuffs and nutrients, Australia i985-86 , Cat. No. 43o6.o, November 1987.