Date Published: October 23, 2018
Publisher: Public Library of Science
Author(s): David M. Troy, Angela S. Attwood, Olivia M. Maynard, Nicholas E. Scott-Samuel, Matthew Hickman, Andy Woods, Marcus R. Munafò, Lorenzo D. Stafford.
The shape of glassware may exacerbate or counteract biases in perceived volume, which may lead people to misjudge the pouring of alcoholic drinks. The aim of these studies was to investigate the effect of glass shape on the pouring accuracy of liquid volume.
In Study 1, using an online computerised task, participants (n = 211) were asked to pour liquid in glasses in a within-subjects design with factors of glass shape (straight, curved) and requested percentage fullness (10, 20, 25, 30, 40, 50, 60, 70, 75, 80, 90%). Curve estimations were carried out to determine if errors followed a linear or non-linear relationship. In Study 2, in a real world experimental study, participants (n = 96) were asked to pour water to the midpoint of pint glasses in a within-subjects design with one factor of glass shape (straight, curved, tulip, inverted). Differences between poured amounts were analysed using one-way repeated measures ANOVA.
In Study 1, participants under-poured in curved glasses compared to straight glasses at all requested amounts. In Study 2, participants under-poured in curved (p < 0.001, dz = 1.51) and tulip (p < 0.001, dz = 0.59) glasses compared to straight glasses. Findings were inconclusive as to whether or not a difference was present between pourings in inverted and straight glasses. Participants displayed a tendency to under-pour in all glasses relative to requested amounts in both studies. The shape of glassware appears to influence the pouring accuracy of liquid. Pouring in tulip and curved glasses was more inaccurate compared to straight glasses, possibly due to the height of liquid within the glass and volume changing in a non-linear relationship.
Alcohol consumption is a leading cause of ill health and premature death in the world . While effective interventions targeting price  and availability  exist, alternative approaches that do not directly restrict consumer choice may be more acceptable. One such approach is altering the presentation of objects in settings where alcohol consumption occurs . Changing certain properties of glassware such as the shape  and adding volume markings  to aid more accurate volume judgements may influence alcohol consumption and therefore could be a target for public health intervention.
Participants (n = 211; 49% female) were on average 33 years (standard deviation = 12, range = 18 to 65). Data from one participant were excluded for all trials as their responses suggested they did not carry out the experiment as instructed. Otherwise, outlying participant data points were removed at the trial level as the pattern of their data suggested they completed the task as instructed. Outliers removed comprised 0.01% of responses.
Participants under-poured in curved glasses at all points tested compared to straight glasses. This may be explained by individuals using the height of liquid in a glass as a proxy for volume. In straight glasses, the height of the liquid and the volume of the liquid change in a direct, linear relationship resulting in participants pouring more accurately by monitoring the height the liquid reaches in the glass. We next investigated the generalisability of these results in a real world environment using glasses of increased volume capacity of varying shapes.
Participants (n = 96; 65% female) were on average 23 years old (SD = 9, range = 18 to 63) and drank an average of 11 units of alcohol a week (SD = 13, range = 0 to 80), with 59% reporting that they consumed beer. No outlying data were detected.
Participants under-poured in tulip and curved glasses compared to straight and inverted glasses. Participants using the height of liquid within glasses as a proxy for volume estimations can broadly explain these findings. Height and volume of liquid changes in a direct, linear relationship in straight glasses while shaped glassware follows a non-linear relationship. Results suggest the more height and volume deviate from a direct, linear relationship in shaped glasses, the more inaccurate pouring becomes. This would explain why pourings in curved glasses were the most inaccurate because the difference in diameter from the narrowest to the widest point of curved glasses was greater than other glasses resulting in an increased overall deviation from a direct, linear relationship. There was no evidence to suggest that weekly alcohol consumption was a good predictor of pouring behaviour. This suggests that regardless of the level of alcohol use, individuals may make the same pouring errors in various glass shapes.
Pourings were closer to the requested amounts (eleven points in Study 1, midpoint in Study 2) in straight glasses compared to curved glasses (Studies 1 and 2) and tulip glasses (Study 2), consistent with our hypothesis. One potential explanation for this is that participants used the height of the liquid as the most salient dimension to estimate volume. Straight glasses may promote more accurate pourings because the height of liquid within the glass and the volume of the liquid changes in a direct, linear relationship.
The shape of glassware appears to influence the pouring accuracy of liquid volume. Straight and inverted glassware seem to assist participants in pouring more accurately than curved and tulip glassware in these studies. Further research is needed to disentangle the different perceptual biases present when pouring alcoholic beverages into differently shaped glassware to establish whether results observed here replicate in naturalistic environments, and to determine their effect on the pouring and consumption of multiple alcoholic beverages in a drinking session. These findings could inform viable interventions to reduce population alcohol use and corresponding alcohol-related harms.