Glassware design may affect drinking behaviour for alcoholic beverages, with glass shape and size influencing drinking speed and amount consumed. Uncertainty remains both about the extent to which these effects are restricted to alcohol and the underlying mechanisms. The primary aim of the current study was to examine the effect of differently shaped glasses on time taken to drink a soft drink. The secondary aim was to develop hypotheses about mechanisms concerning micro-drinking behaviours and perceptual effects.
In a single-session experiment, 162 participants were randomised to receive 330ml of carbonated apple juice in a glass that was either inward-sloped, straight-sided, or outward-sloped. The primary outcome measure was total drinking time. Secondary outcome measures included micro-drinking behaviours (sip size, sip duration, interval duration), and perceptual measures (midpoint bias, drink enjoyment).
Participants drank 21.4% faster from the outward-sloped glass than from the straight-sided glass [95%CI: 0.2%,38.0%] in adjusted models. They were also 18.2% faster from the inward-sloped glass than the straight-sided glass, but this did not reach statistical significance with wide confidence intervals also consistent with slower drinking [95%CI: -3.8%,35.6%]. Larger sips were associated with faster drinking times (Pearson's r(162) = -.45, p < .001). The direction of effects suggested sips were larger from the outward-sloped and inward-sloped glasses, compared to the straight-sided glass (15.1%, 95%CI: -4.3%,38.0%; 19.4%, 95%CI: -0.5%,43.6%, respectively). There were no significant differences between glasses in mean sip or interval duration. Bias in midpoint estimation was greater for the outward-sloped glass (12.9ml, 95%CI: 6.6ml,19.2ml) than for the straight-sided glass, although the degree of bias was not associated with total drinking time (Pearson's r(162) = 0.01, p = .87).
Individuals drank a soft drink more quickly from an outward-sloped glass, relative to a straight-sided glass. Micro-drinking behaviours, such as sip size, are promising candidates for underlying mechanisms.