Table 1 An overview of the studies included in the review

(Blackwell, Drax, Attwood, Munafò, & Maynard, 2018)

(UK [19])

Between-subjects design

To examine the impact of alcohol labels on (i) knowledge of weekly guidelines, and (ii) motivation to drink

1,884

Study 1)

a) basic ABV label

b) responsibility deal label

c) food label equivalent

d) pie chart

Study 2)

One of eight general or specific, positive or negatively framed warnings

Study 1) Accuracy of estimating weekly servings

Study 2) Motivation to drink

Strong

(Jongenelis et al., 2018)

(Australia [22])

Between-subjects design

To assess whether exposing drinkers to warning statements on alcohol products can increase their capacity to make healthier choices

364

One of five warnings:

a) warning: alcohol increases your risk of cancer

b) warning: alcohol increases your risk of diabetes

c) warning: alcohol increases your risk of liver damage

d) warning: alcohol increases your risk of mental illness

e) warning: alcohol increases your risk of heart disease

The extent to which participants believed alcohol is a risk factor for:

a) cancer

b) diabetes

c) liver damage

d) mental illness

e) heart disease

Moderate

(Pham, Rundle-Thiele, Parkinson, & Li, 2017)

(Australia [28])

Study 1) within-subjects design

Study 2) between-subjects design

To investigate the levels of attention paid to alcohol warning labels

Study 1) 559

Study 2) 87

Study 1)

a) moderate size black written warning (control)

b) red text instead of black (colour condition)

c) original label size increased by 50% (size condition)

d) red text and label size increased by 50% (colour and size condition)

Study 2)

a) healthy eating poster

b) organ donation poster

c) one of four randomly assigned wine warning labels (as in study 1)

Study 1)

a) “How much attention did you pay to [labels]”^{Footnote 2}

b) “How much did you concentrate on [labels]”^{Footnote 3}

Study 2)

a) Fixation count

b) Fixation duration

c) Time to first fixation

Study 1) Moderate

Study 2) Moderate

(Hobin et al., 2017)

(Canada [20])

Between-subjects design

To test the efficacy of alcohol labels with SD information and Canada’s LRDG compared to ABV% labels on consumers ability to estimate alcohol intake

2,016

Control label (current Canadian label regulations including ABV%) compared to one of five conditions:

a) pictogram

b) chart

c) label listing number of SD per container

d) SD information and LRDG as a pictogram

e) SD information and LRDG as a chart

Participants were asked to estimate:

a) the amount in a SD

b) the number of SD in an alcohol container

c) the number of SD to consume to reach the LRDG

Moderate

(Wigg & Stafford, 2016)

(UK [23])

Between-subjects design

To test the effectiveness of a range of health warnings

60

a) no health warning (control)

b) text only warning

c) pictorial warning

a) level of fear arousal

b) perceptions of health risk of consuming alcohol

c) intentions to reduce and quit drinking

Low

(Miller, Ramsey, Baratiny, & Olver, 2016)

(Australia [24])

Within-subjects design

To investigate the impact of cancer warning statements on consumer’s level of agreement, prompting conversation, influencing drinking behaviour, and educating others about cancer risk

1,547

a) three drinks a day increases your chance of cancer by 20%

b) alcohol causes cancer

c) two or more drinks a day can increase your risk of mouth and throat cancer by over 50%

d) 1 in 5 breast cancers are caused by alcohol

Likert scale from “strongly agree” to “strongly disagree”:

a) raise awareness about the link between regular alcohol consumption and cancer

b) prompt conversations about the cancer risk involved in drinking alcohol regularly

c) prompt me to drink alcohol less often

d) prompt my friends to drink alcohol less often

e) prompt me to talk to my family and/or friends about the cancer risk associated with alcohol

f) prompt me to educate my children about the cancer risk associated with alcohol

Moderate

(Chen & Yang, 2015)

(USA [25])

Within-subjects design

To examine whether risk perceptions of alcohol-attributable cancer influence alcohol consumption among students

127

a) text warning

b) table warning

c) graph warning

a) perceived susceptibility

b) perceived severity

Moderate

(Krischler & Glock, 2015)

(Luxemburg and Germany [27])

Between-subjects design

To investigate the effectiveness of tailored pictorial warning labels formulated as questions or statements.

122

a) Warning pictures expressed alongside questions

b) Warning pictures expressed alongside statements

c) Bottles with no warning statements

a) Depression-related outcome expectancies

b) Socially-related outcome expectancies

c) Tension-related outcome expectancies

Low

(Osiowy et al., 2015)

(Canada [21])

Between-subjects design

To investigate whether standard drink labels would

improve drinkers’ accuracy when estimating personal alcohol consumption.

301

2 × 3 × 3 experimental design:

two label designs (%ABV, standard drink labels); three beverages (beer, wine and spirits); three beverage strengths (low, regular, high)

(a) relative and (b) absolute percent errors in their estimations of own drinking in comparison with correct answers

Strong

(Pettigrew et al., 2014)

(Australia [26])

Between-subjects design

To investigate the acceptability of cancer warning statements for alcoholic beverages

2,168

Control health statement “Warning: alcohol harms your health” compared to three randomly shown statements from the following:

a) Warning: alcohol increases your risk of cancer

b) Alcohol causes cancer: reduce your intake to reduce your risk

c) Reduce your drinking to reduce your risk of cancer

d) Alcohol increases your risk of bowel cancer

e) Alcohol increases your risk of breast cancer

f) Alcohol increases your risk of breast, bowel, throat and mouth cancer

g) Alcohol increases your risk of cancer

h) Alcohol can cause breast cancer

i) Alcohol can cause bowel cancer

j) Alcohol causes around 5,000 new cases of cancer each year

k) Alcohol causes 1 in 2 cancer deaths

Participants asked to report the extent to which they found the message:

a) believable

b) convincing

c) personally relevant

Moderate