Impact of Bottle Size on In‐Home Consumption of Wine: A Randomized Controlled Cross‐Over Trial
To assess the impact of purchasing wine in 50 cl bottles compared with 75 cl bottles on the amount of wine consumed at home.
Cross‐over randomized controlled trial with a ‘usual behaviour’ period of a maximum of 3 weeks between conditions.
Households in the United Kingdom.
One hundred and eighty‐six households that consumed between two and eight 75 cl bottles of wine each week.
Households were randomized to the order in which they purchased wine in two bottle sizes. During two 14‐day intervention periods, households purchased a pre‐set volume of wine—based on their baseline self‐reported weekly consumption—in either 75 cl bottles or 50 cl bottles. On days 7 and 14 of each study period, participating households sent photographs of each purchased wine bottle.
The primary outcome was the volume of study wine in millilitres (ml) consumed during each study period estimated through returned photographs. The secondary outcome was the rate of consumption measured by the mean number of days taken to consume 1.5 litres from each bottle size.
One hundred and sixty‐six of 186 enrolled households satisfactorily completed the study. After accounting for pre‐specified covariates, 191.1 ml [95% confidence interval (CI) = 42.03–339.2] or 4.5% (95% CI = 1.0–7.9%) more wine was consumed per 14‐day period from 75‐cl bottles than from 50‐cl bottles. Consumption was 5.8% faster (95% CI = –10.9 to –0.4%) from 75 cl bottles than from 50 cl bottles.
— Theresa Marteau (@MarteauTM) April 9, 2020
Consuming wine at home from 50 cl bottles, compared with 75 cl bottles, may reduce both amount consumed and rate of consumption.
The results of this study suggest that, among households consuming between 1.5 and 6 litres of wine a week, purchasing wine in 50 cl bottles, compared with 75 cl bottles, reduces the volume of wine consumed at home by approximately 4.5%. If sustained and replicated, the effect size of this intervention has the potential to make a meaningful contribution to reducing wine consumption at home, where the majority of wine is consumed.
The uncertainty around this estimate is, however, considerable, ranging from 1.0 to 7.9%.
Consumption was also approximately 5.8% slower from the smaller bottles, although again with considerable uncertainty (0.4–10.9%).
Noteworthy, the size of wine glasses used with different-sized bottles may enhance or diminish any effect of bottle size. For example, larger wine glasses increase the volume of wine sold in restaurants but not in bars, which may reflect more free-pouring of wine in the former, given that in restaurants more wine is typically sold by the bottle rather than by the glass. Given that wine glasses have dramatically increased in size during the last three decades, it seems plausible that participating households were frequently using glasses that maximized rather than minimized consumption. Consuming wine from smaller bottles with smaller glasses may therefore enhance the effect size observed in the current study.
The “portion size effect”
The results of this study are in keeping with the robust evidence in relation to food and non-alcoholic drinks showing that people consistently consume more from larger portions or packages.
There are both social and personal norms at play affecting how to eat, drink or consume alcohol.
- Smaller bottles might decrease consumption by signalling a completed episode of consumption when empty, reflecting a tendency for people to consume in ‘units’ regardless of portion or package size
- Smaller bottle size may also have an effect by making additional intake of wine effortful.
Internationally, 75 cl is the most common size in which wine is sold, thereby setting a norm against which smaller and larger sizes are judged.
By constituting two-thirds of a 75 cl bottle, a 50 cl bottle is probably not too small to be considered a large deviation from the norm and thus provoke resistance, but small enough to reduce consumption. Wider availability of 50 cl bottles of non-premium wines has the potential to shift this norm.
Availability of wine in 50 cl bottles, however, is currently very limited. For example, less than 2% of wine stocked by the retailer used in the current study is in 50 cl bottles.
Applicable to other alcoholic products?
The current study compared the impact of 75 cl with 50 cl bottles and provides evidence of the effectiveness of smaller bottles to reduce wine consumption. The results are potentially applicable to beer and spirits when consumed from larger compared to smaller bottles and cans.