A

BouMIwwaRa(cCe©

K

I

eaet

sT

(

0d

Available online at www.sciencedirect.com

International Journal of Drug Policy 20 (2009) 183–187

Short report

Low alcohol alternatives: A promising strategyfor reducing alcohol related harm

David S. Segal a,∗, Tim Stockwell a,b,c,1

a Centre for Addictions Research of BC, University of Victoria, Canadab Department of Psychology, University of Victoria, Canada

c BC Mental Health and Addictions Research Network, Canada

Received 13 January 2008; received in revised form 29 May 2008; accepted 2 June 2008

bstract

ackground: Less than 1% of the beer market in British Columbia comprises beers with an alcohol content below 4%, despite the successf low alcohol beers in other countries, e.g. Australia. A small experimental study is described in which male students were given eithernmarked low alcohol beer (3.8%) or regular strength beer (5.3%) to investigate their enjoyment and subjective intoxication.ethods: Thirty-four male students who reported drinking 5 or more beers in 1 day at least once in the last month volunteered for the study.

n each drinking session, small groups of between 6 and 10 students consumed two servings of beer while playing dominoes. Each subjectas his own control in the experiment by attending two group-drinking sessions, drinking a different beverage each time. The different beersere given in balanced order with half the subjects in each group drinking each type of beer. Standard measures of subjective intoxication

nd enjoyment were used. Blood alcohol levels were tested before, during and after drinking.esults: Although significantly higher blood alcohol levels were obtained with the higher strength beer (means of 0.026 versus 0.033 mg/100 mlt the end of the study, p < 0.001), (i) most participants reported enjoying the two sessions equally or preferred the low alcohol beer session,ii) most did not report feeling different between the two sessions and (iii) only about half correctly guessed which was the higher alcoholontent beer. There was a preference, however, for the taste of the stronger beer.

onclusion: We conclude beer drinkers cannot readily distinguish low and regular strength beers and can enjoy socializing equally withither. We recommend taxation strategies to create incentives for the manufacture, marketing and consumption of low alcohol alternatives.

2008 Elsevier B.V. All rights reserved.

eywords: Low alcohol beer; Alcohol taxation; Preferences for & consumption of low alcohol beer

epB

ntroduction

Alcohol is a leading cause of preventable death, dis-

ase, injury and disability, contributing to 3.2% of deathsnd 4% of disability adjusted life years globally (Rehmt al., 2004). This almost equals the global burden fromobacco and surpasses that from illicit drugs to a consid-

∗ Corresponding author at: Centre for Addictions Research of BC, Univer-ity of Victoria, PO Box 1700 STN CSC, Victoria BC, Canada V8W 2Y2.el.: +1 250 472 5445.

E-mail addresses: dsegal@uvic.ca (D.S. Segal), timstock@uvic.caT. Stockwell).

1 Tel.: +1 250 472 5445; fax: +1 250 472 5321.

locctoclec&

955-3959/$ – see front matter © 2008 Elsevier B.V. All rights reserved.oi:10.1016/j.drugpo.2008.06.001

rable degree (Rehm et al., 2004). In Canada, over 5000eople die from hazardous alcohol use each year (Rehm,aliunas, et al., 2006), and are estimated to cost $14.6 bil-

ion (Rehm, Patra, & Popova, 2006). Beer constitutes 52.5%f the Canadian alcohol market in terms of pure ethanolontent (Statistics Canada, 2007) and there is evidence thatonsumption of lower alcohol content beers are less likelyo contribute to harm. However, in British Columbia, 81%f the beer market comprises beer of 5% or greater alcoholontent (Stockwell et al., 2007) while only 0.22% comprises

ow alcohol beer (2.5–4.0%), and the latter tends to be morexpensive (Stockwell et al., 2007). This pattern of prices runsounter to that in Sweden (Ponicki, Holder, Gruenewald,

Romelsjo, 1997) and Australia (Stockwell & Crosbie,

1 l Journ

2c

schipbC

Ip

hw(MfibwnbGtiabobhKewsllhr

ii2lf

edpstt(ct

ii

M

S

wg(b

P

HVrbyuttipf

w1(snsipsatcpotwuapta

M

84 D.S. Segal, T. Stockwell / Internationa

001) where they reflect alcoholic strength of beers morelosely.

This paper reports an experimental investigation ofubjective levels of enjoyment and intoxication followingonsumption of moderate amounts of unmarked low alco-ol beer (3.8%) compared with regular strength beer (5.3%)n a sample of young Canadian beer drinkers. The pur-ose is to examine the potential acceptability of low alcoholeers as a strategy for reducing alcohol related harm inanada.

s alcohol content a mediating factor for brandreference?

Several laboratory studies have investigated whether alco-ol content plays a significant part in brand preferencehen participants are blind to the beverage being consumed

e.g. Milner, 1979; Fowler, 1982; Cox & Klinger, 1983;cLaughlin, 1988). A common finding is that beer drinkers

nd it hard to discriminate between regular and low strengtheer. Recognizing the compromised social validity associatedith laboratory studies, Geller and his associates conducted aumber of naturalistic studies in the U.S. to investigate howeer brand labelling effects alcohol consumption (Russ &eller, 1998; Kalsher & Geller, 1990). Results from blind

aste tests with college students prior to a fraternity partyndicated that students did not show any preference betweenseries of low alcohol beers with alcohol contents varying

etween 1.7 and 3.7%. When the beers were later labelled andffered at the party, students selected the two higher strengtheers. However, students continued to drink the lowest alco-ol beer when other options ran out. A later study by Geller,alsher, & Clarke (1991) compared drinking behaviour and

njoyment at fraternity parties in which free unmarked beeras provided under two conditions, low strength 3% and high

trength 7%. Subjects consumed similar quantities regard-ess of beer strength and appeared to demonstrate similarevels of enjoyment. However, consumers of the 3% beerad significantly lower blood alcohol concentration (BAC)eadings.

Although some people drink with the intention of gettingntoxicated, many others drink for a variety of reasons, includ-ng social pleasure, taste, and to help relieve thirst (Nutt,006). Clearly, lower alcohol content beers can help with theatter types of reasons and there may be an untapped marketor good tasting low alcohol beers in Canada.

To our knowledge no published study has systematicallyxamined the degree of subjective enjoyment from drinkingifferent alcohol content beers. Geller et al. (1991) “observedractically no complaints about weak drinks” (p. 202), anduggests future research should asses the social validity ofheir intervention. The present study addresses three ques-

ions regarding beer drinkers experiences of drinking low3.8%) versus regular (5.3%) strength beers: (1) Can theyorrectly identify them when consumed unmarked? (2) Dohey equally enjoy the activities engaged in while drink-

B

ts

al of Drug Policy 20 (2009) 183–187

ng? (3) Do they experience differing levels of subjectiventoxication?

ethods

tudy design

A within-subject and within-group experimental designas employed in which 34 participants each attended tworoup sessions, drinking a different strength unmarked beer3.8 or 5.3% alcohol by volume) on each occasion and in aalanced order.

articipants and setting

Ethical approval for the study was obtained from theuman Research Ethics Committee of the University ofictoria. Forty male students, under 30 years of age, were

ecruited by poster advertising and classroom presentationsetween August and October 2006 (mean age was 25.5ears). They were told only that they would receive twonmarked servings of beer on two separate occasions, thatheir BAC may exceed .05%, and that they would be askedo report on their subjective levels of enjoyment and intox-cation. Given the small size of the sample and the greaterreference for beer among males it was decided to excludeemale drinkers from the study.

The Alcohol Use Disorders Identification Test (AUDIT)as administered (Babor, De La Fuente, Saunders, & Grant,989) and persons scoring higher than 8 were excludedn = 3). All participants had to report that they also had con-umed at least five beers at least once in the past month buto more than twice in the last week. This was to ensure theample had some experience of beer drinking and due to eth-cal restrictions minimize the likelihood of serving alcohol toarticipants who may have drinking problems. Six potentialubjects were excluded: Five failed to attend both sessions,nd one was turned away at the beginning of the study dueo prior consumption of alcohol leaving 34 participants. Thelassroom had two tables set up for groups to sit around andlay a group game of dominoes. A bar was set up at the frontf the room, which was used to conceal the beer so that par-icipants could not determine the percentage or brand while itas being poured and served. The same selection of low vol-me music was played during the sessions. Individual deskst the back of the classroom were used for participants torivately fill out their questionnaires. All participants eitherook public transport or arranged someone to drive them tond from the study.

easures

ACThe Alco-sensor IV (Intoximeters, Inc., USA) was used

o assess BACs before, during, and at the end of experimentalessions.

l Journ

S

Efi

(

(

(

B

a(bsc

P

pmsooatawrroiibttsmr

twl

srswapi

S

chtaowc

R

hi

B

rta

S

sdd

D

ub

Enjoyment of Experience Rating (n = 34)

D.S. Segal, T. Stockwell / Internationa

ubjective measuresA combined Subjective Intoxication Scale and Subjective

njoyment Scale was completed by the participants uponnishing two beverages during both experimental sessions.

A) The Sensation Scale (Maisto, Conners, Tucker,McCollam, & Adesso, 1980) consisted of 31 items cap-turing sedative and stimulant effects of alcohol. Itemswere scored using a likert scale ranging from 0 (not atall) to 9 (extremely), with higher scores indicating higherlevels of perceived intoxication.

B) The Subjective Enjoyment Scale was adapted fromPerkins, Ciccocioppo, Jacobs, & Doyle (2003) DrinkEvaluation Questionnaire (DEQ) and consisted of sixitems scored on a likert scale ranging from 1 (not at all)to 7 (extremely): (1) “How do you LIKE the beer youjust drank?” (2)“How good did the BEER taste?” (3)“How SIMILAR is this beer to your preferred brand ofbeer?” (4) “How much do you want ANOTHER beer?”(5) “How APPEALING does the beer look?” Scoreswere totalled and higher scores indicated higher levelsof enjoyment of that particular beer.

C) A single item was added to the DEQ: “How much didyou ENJOY your game of dominos?”

everagesThe low strength beer was Doc Hadfield’s Pale Ale (3.8%

lcohol by volume), a locally brewed product. Kokanee Gold5.3% alcohol by volume) was chosen as the regular strengtheer after conducting blind taste tests where a number ofimilar beers to the Doc Ale were compared based on theriteria of taste, body, and visual similarity.

rocedure

Once the final sample of participants was confirmed,articipants were randomly assigned into groups of approxi-ately six subjects each required to attend two experimental

essions together. All participants were instructed not to eatr drink for 2 h prior or take any medication within 24 hf the study. Upon arrival, participants were asked to readnd sign an informed consent form. Prior to sitting at theables, participants BAC levels were taken and a numeric codessigned which corresponded with a particular seat. Beersere poured into cups behind the bar and then served by the

esearch assistant to ensure blind conditions. Each participanteceived the equivalent of one 341 ml serving of beer in a 9-z, non-transparent plastic cup with their ID on it and werenstructed to begin playing the game of dominos while drink-ng their beverage at a ‘relaxed pace’. After 15 min the secondeer was served and another 15-min window was providedo drink their beers and play dominos. Upon completion, par-

icipants moved to individual desks with their ID on it andilently completed their questionnaires. This took approxi-ately 15 min. After completing the forms, BAC scores were

ecorded. Food was then provided and a final BAC level wasrt

al of Drug Policy 20 (2009) 183–187 185

aken. Those participants whose BAC levels exceeded .04%ere asked to remain in the classroom until it fell below that

evel.All participants completed a debrief process following the

econd session involving questions pertaining to their expe-ience and systematically were revealed the details of thetudy. After debriefing the initial group of 12 participants, itas decided to formalize this process and record responses

s the purpose of the experiment was revealed gradually. Thisermitted only 22 participant responses to be recorded andncluded.

tatistical analysis

One-tailed paired T-tests were applied to confirm or dis-onfirm the explicit hypothesis that drinkers would reportigher levels of enjoyment and intoxication when exposedo the stronger beer. Data obtained from the debriefingt the end of the second session were evaluated usingne-tailed chi-square tests to confirm whether participantsere able to identify which beer had the higher alcohol

ontent.

esults

A summary of all results comparing the low and high alco-ol beer conditions are shown in Table 1 and then describedn more detail below.

lood alcohol concentration

BAC levels were significantly higher after consuming theegular strength beers, both 15 and 30 min after consump-ion confirming that experimental procedures successfullychieved different BACs in the participants.

ubjective Intoxication Scale (SIS, n = 29)

Participants’ subjective intoxication, as measured by totalcore on the SIS did not differ significantly by beverage con-ition. Scores on the SIS for five participants were unusableue to missing data.

rink Evaluation Questionnaire (DEQ, n = 34)

Scores on the DEQ suggest that overall appeal of the reg-lar strength beer was reliably greater than the low strengtheer.

There was no significant difference between subjectiveatings of enjoyment of their group game experience betweenhe low condition and the regular condition.

186 D.S. Segal, T. Stockwell / International Journal of Drug Policy 20 (2009) 183–187

Table 1Mean values of objective and subjective measures and significance of differences between the two beverage conditions

Measure 3.8% beer mean (S.D.) 5.3% beer mean (S.D.) Paired T-test, p-value

BAC (mg/100 ml) at 15 min 0.030 (S.D. = 0.011) 0.037 (S.D. = 0.012) t (33) = −3.212, p < .03BAC (mg/100 ml) at 30 min 0.026 (S.D. = .009) 0.033 (S.D. = 0.010) t (33) = −3.96, p < .001SDE

D

norrpfχ

wMb(

D

taeofibpesofafemmitacw

L

wit

blalmtoinaotattoes

C

baltdusbasns

A

adts

ubjective Intoxication Scale 46.5 (S.D. = 38.67)rink Evaluation Questionnaire 17.71 (S.D. = 6.44)njoyment of Drinking Game 4.67 (S.D. = 1.45)

ebrief Questions (n = 22)

Results from one-tailed chi-square tests revealed no sig-ificant differences in preference of one session over thether, with 10 participants reporting a preference and 12eporting no preference (χ2 = .09, p > .05). Of the 10 thateported yes, 4 preferred the low condition. Furthermore,articipants reported no significant difference in how theyelt emotionally between the two sessions (yes = 9, no = 13,2 = .367, p > .05), and only 12 of 22 guessed correctlyhich session they received the low beer (χ2 = .09, p > .05).ost participants (17/22) reported noticing a difference

etween the beers provided in the two experimental sessionsχ2 = 3.54, p < .05).

iscussion

This study of young male beer drinkers in a naturalis-ic group drinking situation suggests that an unmarked lowlcohol beer was highly acceptable and that the participantsnjoyed the group experience equally with either the 3.8%r the 5.3% beverage. This replicates Geller et al. (1991)ndings that student drinkers could not tell the differenceetween low and regular strength beer. The results also sup-ort Geller et al.’s (1991) hypothesis that similar levels ofnjoyment would be experienced despite differing alcoholtrengths. Differences in ratings of overall enjoyment of theccasion were actually inclined towards more positive scoresor the lower alcohol content beverage, though ratings evalu-ting the overall appeal of the beers were significantly higheror the 5.3% beer. More than half of the participants did notnjoy one session more than the other, and of those that did asany as 40% preferred the low session. While significantlyore participants could tell the difference between the qual-

ty of the two beers and preferred the regular content brand,his appeared to have no connection with alcohol content,s many guessed incorrectly when asked to identify the lowondition. Furthermore, enjoyment of the drinking occasionas not different across the two beverage conditions.

imitations

While the sample size of subjects completing the studyas relatively low, the within-subject experimental design

s a powerful method for detecting differences. The rela-ively small difference in mean BAC after drinking each

dMad

50.97 (S.D. = 34.54) t (28) = −.551, ns21.52 (S.D. = 5.93) t (33) = −2.271, p < .054.26 (S.D. = 1.35) t (33) = 1.281, ns

everage was detected at the less than 1% significanceevel. Due to ethical restrictions, providing larger doses oflcohol were not permitted, nor was the possibility to col-ect data in a purely naturalistic setting. Hence the results

ay have less applicability to higher consumption situa-ions and suffer some of the same social validity limitationsf earlier laboratory based studies. Noteworthy, however,s that the Geller et al. (1991) study was conducted at aaturalistic party setting with no limits on consumptionmounts. Second, the relatively small convenience samplef 34 male university students recruited may not be represen-ative of the general population of male university studentsnd males between the ages of 19–29 years in Canada. Fur-hermore, the study employed only single-blind methods, i.e.he researchers were aware of the study conditions through-ut. This could have potentially caused some experimenterffect and double-blind conditions should be utilized in futuretudies.

onclusions

Despite the extremely small market share of low alcoholeers in Canada, this experimental study suggests that lowlcohol beers may in fact be acceptable alternatives to regu-ar strength beers under some circumstances. Therefore, withhe right incentives and with effective promotions young beerrinking males might be encouraged to consume these prod-cts more often. A shift towards the consumption of lowertrength beers would have many public health and safetyenefits and this need not come at the expense of either thelcohol industry or potential consumer satisfaction. Furthertudies investigating the acceptability of low alcohol alter-atives are recommended using larger general populationamples.

cknowledgements

The authors of this report gratefully acknowledge thessistance and advice received from many individuals atifferent stages of this project. In particular, they wisho acknowledge Melanie Prince and Leila Scannell, bothtudents at the University of Victoria for assistance with

ata collection; Drs. Scott Macdonald, Zinhui Zhao andichael Hunter for advice on statistical and research design

spects, and Ajay Puri for helpful comments on an earlyraft. Lastly, we would like to thank Spinnakers Gastro

l Journ

Bs

Foi

R

B

C

F

G

K

M

M

M

N

P

P

R

R

R

R

S

S

and the marketplace. International Journal of Drug Policy, 12(2),

D.S. Segal, T. Stockwell / Internationa

rewpub for providing a portion of the beer used in thistudy.

Funding for this study was provided from the Endowmentund of the Centre for Addictions Research of BC, Universityf Victoria and was free of any influence from the alcoholndustry and its stakeholders.

eferences

abor, T. F., De La Fuente, M. F., Saunders, J. B., & Grant, M. (1989).AUDIT—The alcohol use disorders identification test: guidelines for usein primary health care. Geneva: World Health Organization.

ox, W. M., & Klinger, E. (1983). Discriminability of regular, light andalcoholic and non-alcoholic near beer. Journal of Studies on Alcohol,44, 494–498.

owler, R. L. (1982). Detection of differences in beer consumption by brandand level of calorie content. Perceptual and Motor Skills, 55, 967–970.

eller, E. S., Kalsher, M. J., & Clarke, S. W. (1991). Beer versus mixed drinkconsumption at fraternity parties: A time and a place for low-alcoholalternatives. Journal of Studies on Alcohol, 52, 197–203.

alsher, M. J., & Geller, E. S. (1990). Beer consumption at university parties:Stimulus control of brand labels. Blacksburg, VA: Virginia PolytechnicInstitute and State University.

aisto, S. A., Conners, G. J., Tucker, J. A., McCollam, J. B., & Adesso,V. J. (1980). Validation of the Sensation Scale: A measure of subjectivephysiological responses to alcohol. Behavior Research and Therapy, 18,

37–43.

cLaughlin, K. (1988). An investigation of the ability of young maleand female social drinkers to discriminate between regular, calo-rie reduced and low alcohol beer. British Journal of Addiction, 83,183–187.

S

al of Drug Policy 20 (2009) 183–187 187

ilner, G. (1979). Light alcohol and standard beers. Medical Journal ofAustralia, (October 6), 383.

utt, D. J. (2006). Alcohol alternatives: A goal for psychopharmacology?Journal of Psychopharmacology, 20, 318–320.

erkins, K. A., Ciccocioppo, M., Jacobs, L., & Doyle, T. (2003). Thesubjective and reinforcing effects of visual and olfactory stimuli in alco-hol drinking. Experimental and Clinical Psychopharmacology, 11(4),269–275.

onicki, W., Holder, H. D., Gruenewald, P., & Romelsjo, A. (1997). Alteringalcohol price by ethanol content: Results from a Swedish tax policy in1992. Addiction, 92(7), 859–870.

ehm, J., Room, R., Monteiro, R., Gmel, G., Graham, K., & Rehn, T.(2004). Alcohol. In Comparative quantification of health risks: Globaland regional burden of disease attributable to selected major risk factors,Vol. 1. World Health Organization., pp. 959–1108.

ehm, J., Baliunas, D., Brochu, S., Fischer, B., Gnam, W., Patra, J., et al.(2006). The Costs of Substance Abuse in Canada 2002. Ottawa: CanadianCentre on Substance Abuse.

ehm, J., Patra, J., & Popova, S. (2006). Alcohol-attributable mortality andpotential years of life lost in Canada 2001: Implications for preventionand policy. Addiction, 101(3), 373–384.

uss, N. W., & Geller, E. S. (1998). Exploring low alcohol beer consump-tion among college students: Implications for drunk driving. Journal ofAlcohol and Drug Education, 33(2), 1–5.

tatistics Canada. (2007). Sale and Control of Alcoholic beverages inCanada, 2006. Ottawa: Statistics Canada.

tockwell, T., & Crosbie, D. (2001). Supply and demand for alcoholin Australia: Relationships between industry structures, regulation

139–152.tockwell, T., Pakula, B., Macdonald, S., Buxton, J., Zhao, J., Tu, A., et al.

(2007). Alcohol in British Columbia and Canada: A case for liquor taxesthat reduce harm. Center for Addictions Research of BC: Victoria.