Proposed pathways
1. Prior preferences.
Example: Adding apples to a display containing bananas and snack bars might increase the probability that a fruit option is customers’ most-preferred option, and therefore increase selections of fruit. In this example, if there were customers who would have preferred fruit, but do not like bananas, adding additional fruit options may lead them to swap their selection away from a snack bar.
1a. Increased availability → increased probability of product in target category being most preferred option
Individuals may select items in accordance with their prior preferences for products within a set of available options. As items are added or removed, the relative ranking of the target product or category with regard to an individual’s preferences may change, changing the likelihood of each option being selected.
1b. Increased probability of product in target category being most preferred option → increased likelihood of selecting target product
This assumption underlies random utility theory in economics, whereby individuals chose the option that yields the greatest utility (i.e. the option that provides them with the greatest satisfaction or gain). As such, if an individual’s relative preference for a target option changes following a change in availability (as in 1a above), and this leads to the target option becoming the favoured option, or dropping from this position, this would then increase the likelihood of a change in behaviour.
2. Social norms
Example: If a vending machine has few units of a particular type of product remaining, empty slots within the machine might imply greater popularity. The opposite pattern might be seen in cafeterias or supermarkets, where the greater presence of certain options – e.g. a greater number of types of chocolate bars vs. types of healthier snack bars – might imply greater popularity of these products in these contexts.
2a. Decreased availability [in the context of reduced quantity of stock] → Social norms updated to reflect greater consumption of target options by others
Increased availability [in the context of full stock] → Social norms updated to reflect greater consumption of target options by others
Availability may alter social norms regarding consumption if individuals infer levels of consumption of target options by others when they observe the available number of products in specific contexts. This may subsequently impact on behaviour, given that when individuals have knowledge of others’ diet-related behaviour in the same setting, they are more likely to make similar selections or consume similar amounts [
65,
66,
71].
Descriptive social norms – encompassing people’s beliefs about how common behaviours are in general or among individuals salient to them (e.g. people shopping in this supermarket are purchasing larger quantities of fruit and vegetables) – may be key. These are correlated with healthy eating behaviours [
7], whereas injunctive norms – beliefs about what other people expect or approve of with regard to these behaviours (e.g. people should buy larger quantities of fruit and vegetables) – are not [
43].
Studies manipulating social norm information often make others’ behaviour explicit – e.g., showing individuals lists that outline “previous participants’ selections”. Fewer studies have looked at the impact of implied popularity or implied social norms on behaviour. Importantly for their relevance to availability interventions, however, the influence of social norms does not have to involve seeing others. Behaviour can be changed through learning about the behaviour of others indirectly via environmental cues. For example individuals are more likely to select a healthier food option if they see empty wrappers suggesting that others have previously chosen this option [
66].
It is unclear whether the opposite pattern might also be seen in cafeterias or supermarkets. For example, individuals may infer increased consumption by others if they observe greater numbers of target products available in contexts where the layouts have been deliberately determined and stock is (close to) full - for example, given awareness of the idea of supply (availability) and demand in commercial enterprises such as supermarkets. This may in turn lead to greater selection and/or consumption on the part of the individuals observing increased availability if this increase alters the perceived social norms regarding purchasing or consumption of these products.
2b
i
. Social norms updated to reflect greater consumption of target options by others→ increased likelihood of selecting target option
Deutsch and Gerard [
19] proposed individuals follow social norms in order to: (a) enhance affiliation with social group – i.e. they want to be liked – and (b) to perform the ‘correct’ behaviour. Such modelling behaviour goes beyond mere imitation, involving an emotional component such as the desire to avoid social sanctions that may be imposed on those who do not follow such norms [
10]. Higgs [
28] proposes that following social norms is an adaptive behaviour, whereby following others in the context of diet makes people more likely to consume and share safe foods. For example, young children use social information to guide eating [
74]. Given these possible motivations, updating of social norms regarding the selection or consumption of particular products by others could lead to changes in behaviour. Indeed, if social norms actually do achieve enduring behaviour change, then this could create a positive feedback loop via continual updating of social norms to reflect this changing behaviour.
2b
ii
Social norms suggest greater consumption by others → increased liking for target option
Updating social norms relating to the selection and consumption of a particular option might also impact on behaviour by altering the liking of an option. Providing social normative information has been shown to change liking for targeted foods and beverages [
69]. This is reflected in reward-related brain activity [
37,
59].
Another consideration is whether the influence of social norms may be moderated by food type and/or individual differences. For example, Pliner and Mann [
65] found that social norms did not influence choices of unpalatable ‘healthy’ cookies over palatable ‘unhealthy’ cookies, while the results of Salmon et al. [
71] suggest a social norm intervention was only effective at encouraging healthier choices if an individual had low self-control. It is possible that social norms may not be able to change liking sufficiently to influence behaviour when choosing between products for which there is a large existing discrepancy in liking.
2c, 3c. Increased liking → increased likelihood of selecting target option
Taste preferences and liking are reported to be among the most important influences on dietary behaviour [
20,
24], including children’s food choices [
14,
50]. Beyond self-report, manipulating implicit attitudes [
26] – a measure of liking, which reflects the positive (vs. negative) associations of different food items to individuals – can result in altered food selection [
32,
34], suggesting that emotions associated with the product can drive behaviour.
3. Visual attention.
Example: If a display goes from containing 50% sugar-free beverages and 50% sugary beverages to 75% sugar-free and 25% sugary beverages, then there may be a greater likelihood that attention is drawn to the sugar-free beverages, given these take up a greater proportion of the visual field.
3a Increased availability → increased probability of target option attracting visual attention
Increasing the availability of a target option relative to other non-target options would lead to the target options taking up greater space in the visual field. If an initial fixation point when first viewing a scene is random, then increased availability would increase the likelihood that a target option is the one initially observed. Evidence suggests, however, that initial fixation is likely to be on the most salient stimulus [
36]. As such, the likelihood of initial fixation being on an option with increased availability would depend on its relative salience compared to the other options available. Following initial fixation, the attended-to location is transiently inhibited allowing attention to be redirected. When attention is redeployed, the increased proportion of the visual field dedicated to the option with increased availability may make this option more likely to attract attention. However, even with an increased visual presence due to increased availability, if individuals are actively searching for a particular option, they may not attend to these additionally available options, a phenomenon described as inattentional blindness [
75].
3b. Increased visual attention → increased liking for target option
Increased visual attention may increase liking for a target option in line with the “mere exposure” effect [
86]. This suggests that the appeal of a stimulus – such as Chinese ideographs, geometric shapes, or human faces – is increased after repeated image-based and physical exposures [
18,
55,
85]. Moreover, gaze duration may both reflect and influence preference; in one study participants’ gazes began evenly distributed between two options, but then shifted to predominantly focus on the option they subsequently selected, whereas manipulating gaze duration biased decisions towards the more viewed option [
73]. However, the mere exposure effect may require attention to be directed towards the stimulus [
85], and consistency of context between exposure and testing [
18]. In the context of research on food, the visual appeal (but not expected tastiness) of a food to children has been found to increase after viewing these items [
11]. Therefore if increased availability draws visual attention to a set of options, this may in turn increase liking for these options – although the extent to which liking might extend beyond the visual domain is uncertain.
Evidence that exposure or increased visual attention can change behaviour is limited, and the handful of available studies provide equivocal evidence. Rangel and colleagues have conducted a series of laboratory studies looking at selecting between two or three food options, and developed models to predict selection [
41,
42] (N.B. these models assume that initial gaze is random, contrary to possible influences of visual salience). They find increased visual attention increases the probability of selecting items self-reported as liked by participants, but decreases the probability of selection for disliked items [
5].
In terms of field studies, the extent to which visual attention might be involved in manipulations is generally unclear. Some studies suggest increased sales when healthier options are placed at eye level [
44,
80], but this positioning was also designed to increase the accessibility of these items, so the mechanism for an effect remains unclear. In contrast, Van Kleef et al. [
82] found no effect of top vs. bottom shelf placement. Other studies suggest individuals prefer the middle option in an array [
39,
53], perhaps because the centre of horizontal arrays receives more visual attention [
6]. But these findings could potentially also be explained by ease of reach. As such, the potential for increased visual attention to lead to changes in behaviour in real-world contexts is yet to be demonstrated.