A problem unstuck? Evaluating the effectiveness of sticker prompts for encouraging household food waste recycling behaviour.
Linzi Shearer ab *, Birgitta Gatersleben c, Stephen Morse a, Matthew Smythb, Sally Huntb
aCentre for Environmental Strategy, University of Surrey, Guildford, Surrey, GU2 7XH, United Kingdom
bWaste Development, Surrey County Council, Kingston-Upon-Thames, KT1 2DW, United Kingdom
cDepartment of Psychology, University of Surrey, Guildford, Surrey, ,GU2 7XH, United Kingdom
Abstract
This Randomised Control Trial (RCT) investigated the effectivenessofusing stickers as a visual prompt to encourage the separate collection of household food waste for recycling in two local authorities in South East England. During a baseline period of up to 15 weeks,separately collected food waste was weighed (in tonnes) and averaged across households in both treatment (N = 33,716 households within 29 defined areas) and control groups (N = 30,568 households within 26 areas). A sticker prompt was then affixed to the lids of refuse bins in the treatment group area only. Weightsfor both groups were subsequently measured acrossa 16-week experimental period.Results showed that, in the control group, there was no change in the average weight of food waste captured for recycling between the baseline and experimental period. However, there was a significant increase (20.74%) in the treatment group, and this change in behaviour persisted in the longer term.Sticker prompts therefore appear to have a significant and sustained impact on food waste recycling rates, while being simple, practically feasible and inexpensive (£0.35 per household) for local authorities to implement at scale.
Key words:
Visual prompts, household food waste, recycling behaviour, nudge, behaviour change.
* Corresponding author
Email addresses: , , , ,
1.Introduction
Changing patterns of human production and consumption in industrialised nations have resulted in increased levels of household food waste (Parfitt et al., 2010). In the United Kingdom (UK), households are responsible for generating around half of the15 million tonnes of food and drink waste that is produced each year (House of Commons, 2015; WRAP, 2011).The EU Landfill Directive (1999/31/EC) specifies that Member States must reduce levels of biodegradable waste sent to landfill to 35% of 1995 levels by 2020, but does not prescribe the treatment options, collection systems, or other policies that should be introduced to meet these targets (Defra, 2011).
In the UK, local authorities are responsible for the collection and disposal of biodegradable waste,whichhas traditionally been sent for disposal to landfill sites (House of Commons, 2015). Methane, a greenhouse gas far more potent than carbon dioxide (CO2), is released when biodegradable waste (which includes household food waste) decomposes anaerobically at landfill sites (Graham-Rowe et al., 2015).The UK Waste and Resources Action Programme (WRAP) (WRAP, 2011) estimate the annual environmental impact of manufacturing, distributing, storing, using and disposing of edible food and drink in the UK to be around 17 million tonnes of CO2 equivalent.
If collected separately from residual waste (refuse), food waste can be used as a feedstock for anaerobic digestion (AD), a ‘recycling’ process that produces methane-rich ‘biogas’ that can be used to generate renewable energy, and ‘digestate’ which can be used to produce agricultural bio-fertiliser (Zhang et al., 2007) and has a lower disposal cost (approximately 50%) than landfill (Nomura et al., 2011). Households are issued with a small food waste bin (known as a ‘caddy’) for use inside the house and a larger caddy that is stored outside. The purpose of introducing the service is to encourage households to separate their food waste from their refuse andstore it in their indoor caddy before transferring it to the outdoor caddy in advance of theirweekly collection day. By diverting food waste in this way, local authorities can increase their overall recycling rate, while saving money and improving their environmental performance.
When the first local authorities in the UK introduced separate food waste collection servicesin 2006, just 1% of household food waste wasbeing collected separately (Defra, 2015). By 2013/14, more than half of the local authorities responsible for waste collection in England had introduced some form of food waste collection scheme (House of Commons, 2015). As a result, the total amount of separately collectedfood waste in Englandincreasedfrom 118,000 to 290,000 tonnes between 2010 and2014 (Waste Data Flow, 2016).Despite this dramatic improvement,almosta third of the refuse waste stream in England is still composed of food waste (WRAP, 2011). This suggests that manyhouseholdsare still notparticipating in the scheme, or those that are taking part are not using their caddies as effectively as they potentially could be. Since the successful management of any household recycling scheme is dependent on the effective participation of a sufficient number of households, local authorities must introduce policy interventions designed to encourage public participation (Karim Ghaniet al., 2013).
To encourage food waste separation behaviour,local authorities have a number of policy interventions at their disposal (Dahlén & Lagerkvist, 2010; Steg & Vlek, 2009). The decision about which policy interventions should be introduced is not a simple one for decision-makers as they must be clearly effective, in the sense of producing changes in behaviour that are sustained in the longer term (Steg & Vlek, 2009). An intervention must also be practically feasible to introduce and there must be a sufficient means of accurately monitoring the impact of the intervention to understand whether it was successful or not (Steg & Vlek, 2009). They must also be cost efficient to implement as local authorities in England are currently under increasing pressure to deliver ‘more with less’, following a 40% reduction in funding from central government (LGA, 2014). Hence there is an obvious attraction for the use of relatively simple and cost-effective approaches.
An emerging body of literature has advocated the use of ‘nudge interventions’ as alternatives to more traditional forms of behavioural intervention (Dolan et al., 2010; Thaler & Sunstein, 2008). Nudge approaches have been hailed to be a potentially powerful, low-cost set of tools for policy makers who are faced with addressing the challenge of weighing environmental sustainability against fiscal responsibility when making decisions, particularly during times of austerity (Dolan et al., 2010; John, 2013;Kallbekken & Sælen, 2013).
Nudge is a valuable theoretical framework that summarises ideas drawn from the field of behavioural economics, which itself draws insights from the field of psychology (Kirakozian, 2016). Proponents argue that traditional policy tools have ignored the fact that much human behaviour is automatic in nature (Michie, 2015), recognising that behaviours most often occur as a consequence of both automatic and reflective processes running in parallel. Nudge interventions, therefore, mainly target the automatic system and seek to change the ‘choice architecture’ of individuals to encourage changes to attitudes and behaviour (Sunstein, 2015). The approach assumes that people will rely on past ways of thinking and acting unless they are encouraged to act or think differently. The options for changing behaviour centre on providing reminders and cues that both recognise where the individual currently is while also placing them in a choice environment.
Good designers of nudge policy interventions can steer individuals down new decision pathways without them necessarily noticing that it is happening. Behavioural change is achieved by altering how individuals view the attractiveness of an alternative course of action by improving the messages they receive or the opportunities they have. While no “precise, operational definition of nudging” (Martaeu et al., 2011: 263) currently exists, ataxonomy of interventions published in a recent House of Lords report (House of Lords, 2015) described nudges as being any one of the following: changes to the physical environment; information provision; changes to the default policy and the use of social norms and salience.
One form of nudge intervention, ‘visual prompts’ has had a wide applicability within a variety of behavioural fields. Visual prompts are a form of informational intervention designed to stimulate action or serve as a reminder to engage in a behaviour that might otherwise beforgotten (Chui, 2015). Visual prompts usually take the form of posters, signs, stickers or flyers (Bartram, 2009), and display factual or persuasive information, or provide cues to aid behavioural decision-making (Geller, 1982; Sussman & Gifford, 2012). Their intensity can vary from simple notices that raise awareness or provide procedural information, to more comprehensivestatements that provide context and rationale (Tucker, 2001).
Several studies have demonstrated the effectiveness of visual prompts as a means for encouraging transportation (e.g., Cope et al., 1991; Huyberset al., 2004) and health behaviours (e.g., Amass et al., 1993; Andersen et al., 2012). They have also been used toencouragecertain pro-environmental behaviours, for example: litter reduction in public places (Baltes & Hayward, 1976; Geller et al., 1976); increasing workplace recycling (Austin et al., 1993); and reducinghousehold energyconsumption (Sussman& Gifford, 2012; Winett, 1978).
The effects of prompts on general household recycling behaviour specifically have also been widely documented, but results are not consistent. Some research has shown that single prompts in isolation can be an effective way of increasing recycling (e.g., Arbuthnot et al., 1976; Burn, 1991; Jacobs & Bailey, 1982; Oskamp, 1995; Spaccarelli et al.,1990), while other studies suggest that prompts are less effective than other types of intervention (Goldenhar & Connell, 1992; Schultz, 1999; Werner et al.,1998; Witmer & Geller, 1976). A recent meta-analysis found that prompts were one of the most effective intervention types for encouraging pro-environmental behaviour (Osbaldiston & Schott, 2011). However, as 78% of the studies included in the analysis tested interventions in combination, it was not possible for the authors to make definitive conclusions about which interventions were most effective in isolation. It is therefore possible that prompts are effective but only when delivered in combination with other types of intervention.
Other research has explored the elements of design that can improve the effectiveness of visual prompts. Several authors have indicated that ensuring the final product isnoticeable, simple and clear is important (Sussman et al., 2012). Adding pictures to written information may also improve effectiveness(Roberts et al., 2009), provided the images used are congruent with the text (Jae et al., 2008) and they do not ‘cloud’ the message (van Meurs & Aristoff, 2009). Some studies have shown that certain attempts to persuade using visual prompts can cause individuals to protest and engage in undesirable behaviours (Sussman & Gifford, 2012). This phenomenon, which threatens the perceived freedom of individuals, is known as psychological reactance (Brehm, 1966; Dillard & Shen, 2005) and can be reduced by constructing messages using positive and polite language (Aronson & O’Leary, 1983; Reiter & Samuel, 1980). Finally, prompts work most effectively for those behaviours that are ‘simple, easy, effortless and repetitive’ to perform (Frederiks et al., 2015: 1391), and on those individuals who already feel motivated to engage in the target behaviour (Schultz, 2013).
The persuasive impact of a visual prompt will not only depend on the message and its design, but also on the recipient’s capacity to attend to and cognitively process the information (Borgstede & Andersson, 2010). The same authors also suggest that the most important factor for any behavioural intervention is attracting the attention of the target audience. Most previous studies on prompting used leaflets or posters as the medium of delivery, yet their effectiveness has been called into question (cf. Read, 1999). The more permanent nature of a bin sticker may allow for repeated exposure to the prompt message, thus providing more opportunity than a leaflet for individuals to notice and cognitively process the message.
Further, it is noteworthy that most experimental research on visual prompts and recycling behaviour was conducted between 1980 and 2000, at a time when recycling had not yet been established as a societal norm (Bedford et al., 2010; Thomas & Sharp, 2013). Since this period, there has been a substantial societal shift in attitudes towards environmental issues in general (Thomas & Sharp, 2013) and hence further research into the effects of prompting may be warranted. Little is also known about the longer-term impact and cost effectiveness of behavioural interventions (Abrahamse & Steg, 2013). With some exceptions (cf. Burn, 1991; Spaccarelli et al., 1990), most existing research on prompting and recycling behaviour had short-term monitoring periods or did not include longer-term follow-ups (Schultzet al., 1995). Finally, since most research on behaviour change interventions is academic in nature, little is known about the actual costs of implementing such strategies in the real world (Schultz, 2013). Academic studies tend to be small in scale and staff-intensive, often using research assistants to carefully control implementation. However, when these techniques are implemented at scale, costs usually increase. It is therefore important for researchers to provide detailed cost information in addition to reporting results.
Despite the mixed findings noted above, it seems logical to assume that visual prompts, if properly designed and implemented, could provide local authorities with one practically feasible and inexpensive strategy for targeting a large number of households. However, the evidence about their effectiveness on general recycling behaviour is mixed and, to our knowledge, no academic studies have investigated their effectiveness for promoting food waste recycling behaviour.
A recent non peer-reviewed research study (WRAP, 2016) sought to assess the effectiveness of various combinations of the following interventions: a year’s free supply of caddy liners, a leaflet providing information about how to recycle food wasteand a ‘visual prompt’ in the form of a bin sticker designed to be stuck on the lid of refuse wheelie bins (WRAP, 2016). Working in partnership with 11 local authorities, a total of 19 pilot studies were carried out between 2013 and 2015, 11 of which specifically tested the interventions outlined above. The weights of food waste (in tonnes, t) collected from pre-defined groups of households were monitored before and after the delivery of each intervention combination. Results were quite variable. The one study that tested the sticker and liner combination achieved a 13% increase in weight of food waste collected. The mean increase in weight achieved by the sticker and leaflet combination was slightly higher at 14% (two studies, range = 4% to 24%). The full ‘package’ of interventions achieved a mean increase of 32% (six studies, range = 18% to 62%). WRAP concluded that the effect on food waste collection rates was amplified when refuse bin stickers were included. They recommended that local authorities should introduce the full ‘package’ of interventions to have the greatest impact on behaviour, at a cost of £1.12 to £2.10 per household (WRAP, 2016).
While the addition of the sticker prompt to the other interventions appeared to have an amplifying effect, it was never tested in isolation. WRAP did not incorporate basic experimental design principles (e.g., randomisation, replication and control groups) into their study and methodologies for analysing results also varied between authorities. While sample sizes were often large (up to 15,000 households in some study areas), the outcome measurement was the weight of food waste collected from defined groups of households known as ‘collection rounds’. For each pilot study there was an insufficient number of collection round weights to statistically analyse results. The collection round areas chosen to receive the interventions were also not selected at random and no control groups were included for comparison against a ‘do nothing’ scenario on the part of the local authority. Further, results from one authority could not always be directly compared with those from another authority, as there were subtle, yet discernable, differences in the way in which the tonnage data were recorded and analysed. As a result, WRAP’s findings cannot be generalised to other populations, as these issues did not allow for a statistically robust set of conclusions to be produced about the relative effectiveness of each intervention combination.
Separate food waste collection services were first trialled in England between 2006 and 2008 (Bridgwater & Parfitt, 2009). It is therefore surprising that so few academic studies have examined the effectiveness of policy interventions for promoting this behaviour. A review of the literature found just three studies that have examined the impact of informationalstrategies on food waste separation behaviour. Bernstad et al. (2013) examined the effect of delivering oral information via a door-stepping campaign. By weighing the separately collected food waste pre and post intervention, they found that the difference in average weights collected by Swedish households in the control and treatment groups was not statistically significant. Using the same outcome measures, Bernstad (2014) assessed the effectiveness of written information (a leaflet about how to recycle and why it is important) on a different group of Swedish households and found that this type of written information also did not significantly increase the weight of food waste collected. Thethird study, a Randomised Control Trial (RCT) conducted in Manchester, England (Nomura et al., 2011), showed that feedback comparing a household’s food waste recycling behaviour with other households in the same street resulted in a statistically significant increase in participation of 2.8% compared with a control group.
In light of the above, the aim of the present study was to investigate the real world effectiveness, practical feasibility and cost-efficiency of using sticker prompts to encourage households to recycle their food waste.Since the true test of any behaviour change intervention lies in its ability to elicit a change in behaviour, the first objective of the field experiment was to test the effectiveness of the visual prompt (a bin sticker) amongst households in two local authorities in Surrey, England. It was hypothesised that, compared with the baseline period, significantly higher weights of food waste would be collected in the treatment group during the experimental period than would be collected in the control group (H1). A secondary objective was to assess the extent to which the impact of the intervention persisted in the longer term. It was hypothesised that significantly higher weights of food waste would be collected in the treatment group during each post-intervention period (short, medium and long term) than in the control group (H2). The final objective of the study was to conduct a cost-benefit analysis to estimate the payback period for this intervention.
2.Materials and methods
2.1Participants & setting
Surrey is acounty located in the South East of England that operates across 642 square miles and has a population of over 1.1 million people (ONS, 2011). In Surrey there are a total of 11 district and borough councils, each acting as Waste Collection Authorities (WCAs), responsible for the collection of refuse and recycling.Two of these WCAs participated in this study. WCA1 and WCA2 are responsible for the collection of waste and recycling from 33,538 households and 30,746 households respectively (Ntotal = 64,284 households).