Lecture 3 Model Answer
Attwood et al (2012) deploy the tools of psychophysics - developed to study perception - to address a health behaviour concern of contemporary relevance - alcohol consumption. Glass design is, they argue, a "currently unregulated marketing channel" (p.1) which could be influencing drinking speed.
Two possible mechanisms are offered. Glasses a) present branding information b) may affect perception of volume (for example, if the glass curves outwards towards the top the greater proportion of the contained liquid is in the top half of the glass). The authors cite experimental evidence that people rely on the vertical plane to make volume judgements (and can hence be misled by irregular shapes).
Participants were 159 social drinkers in Bristol, recruited from the University and city populations via a variety of methods, including word of mouth. Participants were social drinkers, defined as consuming 10-50 (males) or 5-35 (females) units of alcohol per week. Across two sessions (counterbalanced) participants consumed a beverage whilst watching a nature documentary (drinking task), and completed a task involving judging glass fullness (perceptual judgement task). There were 3 between-subjects factors, as part of a 2 (lager vs lemonade) x 2 (6 oz vs 12 oz amount) x 2 (straight vs curved glass) design. Equal males and females were included in each cell. Filler tasks were included to disguise the purpose of the study.
The main dependent variable in the drinking task was time to finish drink, but number of sips, duration of sips and sip interval were also measured, along with pre and post-consumption craving.
The main dependent variable in the perceptual judgement task was estimated midpoint of two glasses (i.e. "where is the glass half-full"). These were found using the staircase method, a classic method of psychophysics.
In the drinking task, subjects look longer to finish their drinks for lager vs lemonade, for straight glasses vs curved and for 12 fluid oz vs 6 fluid oz. A directly tested comparison (Figure 2), showed that in the case of 12 oz of lager vs of lemonade and curved vs straight glasses, the straight glass specifically slowed drinking time for lager only (from ~6 minutes in the other 3 conditions, to ~11 minutes).
Follow up analysis suggested this was due to more time spent between sips, rather than due to difference in the number or length of sips.
In the perceptual task participants, on average, perceived the glass midpoints to be below the true midpoint, with the curved glass midpoint significantly below that of the straight glass midpoint (28/60 for the straight glass, 21/60 for the curved glass).
There was no significant correlation between individual participant's perceptual judgement and their drinking speed.
== CONCLUSIONS, IMPLICATIONS AND CRITICISMS ==
The study provided weak evidence that glass shape may affect drinking speed, but the proposed mechanism of action (effect on midpoint perception) is not strongly supported. Although there is a group difference in midpoint perception there is no significant correlation between individual's perception and their drinking speed, and the group difference in drinking speed only applied for larger amounts and the alcoholic beverage. Whilst it is possible that participants consumed alcohol differently than lemonade, the most striking effect was that the straight glass slowed consumption, not that the curved glass speeded it.
The proposal that misperception of the midpoint causes people to rush their drink has an obvious direct test: use glasses which have the curves at the bottom so that the midpoint is perceived as higher rather than lower than the true. The idea that, only for alcohol, people regulate their consumption also has a direct test - people can be asked if they are savouring their drink. They could also have their enjoyment of consumption measured.
The authors suggest that clearly labelling the midpoint of glasses might be a policy target to reduce binge drinking. It is worth noting the large gap between the specific results of this experiment and the assumptions which need to hold for the result to generalise strongly enough to support a policy intervention. Not only do the results of the current study have to be broadly correct (curved glasses really do create faster drinking than straight), but their proposed mechanism of action also has to hold (i.e. that this is due to midpoint perception). This has to hold across different populations and between the experimental conditions and the normal, social, conditions of alcohol consumption. The subjects are all social drinkers, so this isn't unreasonable to assume. Also we have to assume that the effect will hold for larger amounts and that there will be no 'adaptation' effect (whereby people adjust their behaviour to somehow counter the intervention).
Reviews of this area have consistently concluded that psychological 'nudges' on drinking behaviour are less effective than economic and legal constraints such as minimum unit pricing or limiting sale hours (Booth et al, 2008). The prevalence of binge drinking before the introduction of branded curved glasses suggests that glass shape isn't a major factor in drinking alcohol consumption (although even minor factors can be important public health targets).
Attwood, A. S., Scott-Samuel, N. E., Stothart, G., & Munafo, M. R. (2012). Glass shape influences consumption rate for alcoholic beverages. PloS one, 7(8), e43007.
Booth, A. et al. (2008) Independent review of the effects of Alcohol Pricing and Promotion: Part A: Systematic Reviewsâ, Project Report for the Department of Health