Identification of different user types for designing household food waste interventions

ABSTRACT

To date, designers have developed various interventions to support users in reducing their household food waste. When designing food waste interventions, it is essential to have a clear understanding of target users’ characteristics and consider the diversity in these characteristics. Although there are studies exploring user diversity in the context of sustainable behaviours, the literature lacks studies investigating food waste as a case. This paper explores user diversity in household food waste behaviour and identifies four user types that need to be considered when designing food waste interventions: Conservers, Considerates, Reluctants, and Prodigals. It further gives suggestions on how designers can select target user types and interventions according to their characteristics. It concludes that a design team should target all of the user types in a target population as much as possible and use a combination of different interventions types to create an impactful solution.

KEYWORDS:

• Household food waste
• design interventions
• design for sustainable behaviour
• user diversity

1. Introduction

Food waste and loss is an a global issue. One-third of the food globally produced for human consumption is lost or wasted (Gustavsson et al. 2011; FAO 2019). The United Nations has identified the prevention of food waste as an important sustainable development goal (SDG12.3; UN 2015) because it has adverse effects on climate, water, land, and biodiversity (FAO 2013). Household food waste is a significant contributor to global food waste. It constitutes 53% of total food waste produced in European Countries (Stenmarck et al. 2016) and 60% of the food waste produced in the United States (Griffin, Sobal, and Lyson 2009). Since a diverse array of behavioural and attitudinal factors causes consumers to waste food (Quested et al. 2013; Evans 2012; Aschemann-Witzel et al. 2015; Stefan et al. 2013), consumer/user behaviour is instrumental in the amount of waste created at the household level. Thus, influencing user behaviour seems to be a promising way to deal with household food waste.

Within the last decade, designers have been increasingly interested in developing novel solutions to environmental problems by influencing user behaviour (Cash, Hartlev, and Durazo 2017; De Medeiros, Da Rocha, and Ribeiro 2018; Bhamra, Lilley, and Tang 2011). To date, various technological interventions have been designed to support the performance of food waste prevention behaviours; to name a few, smart waste bins (Thieme et al. 2012; Comber and Thieme 2013), augmented fridges (Farr-Wharton, Choi, and Foth 2014), and mobile apps for food waste management (Ganglbauer, Fitzpatrick, and Molzer 2012; Lim et al. 2017). These interventions showed that design could have an active role in reducing household food waste by increasing users’ awareness and triggering self-reflection.

Behavioural, psychographic, and sociodemographic factors influence household food waste. Since these factors vary according to individual characteristics, interventions adapted to different needs of target users tend to be more successful (Broms, Bang, and Hjelm 2009; He, Greenberg, and Huang 2010). For example, an intervention designed for university students who live in dormitories with their friends (e.g., showing the amount of food waste in social media apps) would be different from the one designed for senior citizens who live alone (e.g., smart containers to preserve the quality of food for a long time). Thus, designers should consider the diversity in these factors when designing food waste interventions (Coskun and Erbug 2014).

There have been some efforts, outside the design field, to categorise consumers according to their attitudes, beliefs and behaviours pertaining to sustainable behaviours, for instance, Ottman’s (2011) ‘green consumers’ and Barr and Gilg (2006) lifestyle groups for environmental action. This work can provide initial guidance for designers who would like to consider user diversity when designing interventions. In the field of design, some design researchers proposed generic user types to promote sustainable behaviours for different users; for example, hypothetical personas for sustainable behaviours (Coskun and Erbug 2014), and designers’ models of the users when designing for sustainable behaviours (Lockton, Harrison, and Stanton 2012). As these user types are generic, they can be used across various behaviours like electricity and water conservation, choosing public transportation, food waste prevention and etc. Other design researchers identified user types for specific sustainable behaviours by analysing empirical data about users, e.g., mending typologies for product repair (Lilley, Bailey, and Charnley 2013), user types for using electrical appliances (Cor and Zwolinski 2014), and user types for eco-friendly driving (Coskun and Erbug 2017). However, none of these studies has examined user diversity in the context of reducing household food waste. The only exceptions are two studies outside the design field, which identified user types based on German consumers’ knowledge and perception of food waste (Richter 2017) and UK consumers’ food waste behaviour (Metcalfe et al. 2012). Furthermore, even though all of these studies can help designers understand user diversity for sustainable behaviours, none explores which user type would prefer which intervention type, essential knowledge to tailor a design intervention to a specific target user group’s characteristic.

This study addresses these problems by investigating household food waste behaviour of 150 Turkish individuals and answering the following questions: 1) what are the different user types for designing technological food waste interventions and 2) which intervention types are suitable for each user type. Turkey is selected as a case due to two reasons. First, despite the recent campaigns initiated by the Turkish Ministry of Agriculture and Forestry, household food waste is a growing problem for Turkey (Ministry of Customs and Trade 2018). Total edible food loss and waste in Turkey is 26,04 million tons per year, and food loss and waste in the consumption step account for 8% of this value (Salihoglu et al. 2018). Household food waste in Turkey (8.61 million tons/year) is approximately 28% of the total municipal solid waste (Pekcan et al. 2006). Second, most household food waste studies have investigated developed countries (e.g., the United States and European countries). A study exploring Turkish consumers’ household food waste behaviours would be a good addition to the existing knowledge.

The study makes two contributions to design for sustainable behaviour literature. First, it identifies and describes four user types for designing food waste interventions (Considerates, Conservers, Reluctants, and Prodigals). Second, it makes recommendations on selecting the right target user types and deciding on the right interventions. Designers and researchers working on generating design interventions to reduce household food waste can utilise these user types in their future endeavours by following these recommendations.

2. Materials and methods

2.1. Sample and the data collection

A questionnaire was used for data collection. Non-probabilistic quota sampling was used to recruit participants. Gender and Socio-economic Status (SES) were used as sampling variables. Female participants are the majority because women usually have greater responsibility in household food provision in Turkey (Karaca and Altun 2017). The participants are from low-middle (SES-C1), middle (SES-B), and upper-middle (SES-A) socioeconomic status. This was because these groups represent the top three high-income groups¹ in Turkey (TUAD 2012), and previous studies reported a positive correlation between income level and food waste (Stancu, Haugaard, and Lähteenmäki 2016).

The participants were recruited through field visits to public places in different neighbourhoods of Istanbul (e.g., squares, parks, cafes, and shopping malls). Istanbul was selected since it is the biggest city in Turkey, with a population of over 16.000.000. First, the researcher asked for potential participants’ consent. After they agreed to participate and signed the form, the researcher asked a series of questions for screening participants who meet the inclusion criteria, i.e., sex, household income, education level, and occupation. If a potential participant had the targeted sample characteristics, s/he was asked to fill the questionnaire. Completing the questionnaire approximately took 25 minutes. In total, 150 participants filled the questionnaire between November 2019 and December 2019. Table 1 summarises the sample characteristics.

Table 1. Participant characteristics

Variable	Characteristic	Percentage
Variable	Characteristic	Percentage
Sex	Female	61 %
	Male	39 %
Age	18-25	21 %
	26-35	27 %
	36-45	17 %
	46-55	25 %
	> 55	10 %
Socio-economic status	Upper (SES-A)	20 %
	Middle (SES-B)	45 %
	Low-middle (SES-C1)	35 %
Household type	Lives with family	78 %
	Lives with friends	11 %
	Lives alone	11 %
Having children	yes	40%
	no	60%

2.2. Measures

The questionnaire has five parts, each aiming to measure different variables.

2.2.1. Sociodemographic factors

Although sociodemographic factors do not play a significant role in explaining food waste behaviour, some of these factors, including age, number of individuals in a household, and socioeconomic status, have a relationship with household food waste. For example, older people tend to produce less waste than younger individuals (Stancu, Haugaard, and Lähteenmäki 2016; Secondi, Principato, and Laureti 2015; Schanes, Dobernig, and Gözet 2018). Bigger households tend to produce less food waste than smaller households (Quested et al. 2013). Households with children tend to produce more food waste than singles (Visschers et al., 2016; Schanes, Dobernig, and Gözet 2018; Parizeau, Von Massow, and Martin 2015). High-income households produce more waste than low-income households (Parfitt, Barthel, and Macnaughton 2010; Thi, Kumar, and Lin 2015). Thus, the first part of the questionnaire included questions measuring age, sex, income, education level, and household type.

2.2.2. Personality traits

Individual differences and personality traits also influence food waste prevention behaviours (e.g., Swami et al. 2011). Thus, the second part included Ten-items Big Five Personality Inventory to measure participants’ personality traits (Rammstedt and John 2007).

2.2.3. Self-reported food waste and related household practices

Food turns into waste due to various behaviours related to its journey through home. These behaviours include planning, shopping, storage, preparation, and consumption (Evans 2012; Ganglbauer, Fitzpatrick, and Comber 2013). Poor decision making and lack of knowledge in preventing food waste when performing these behaviours (e.g., excess buying or being unaware of how to store a particular type of food) increase household food waste (Farr-Wharton, Choi, and Foth 2014). Therefore, the third part included questions asking participants’ agreement levels to various food-related practices for planning, shopping, cooking, and storing. This part also included questions about participants’ self-reported food waste measured by the items adapted from Visschers et al. (2016). These items ask the frequency and the amount of wasted food in four different food categories: meat, dairy, bakery, and fruits and vegetables.²

2.2.4. Attitudinal factors influencing household food waste

Individuals’ intention to avoid food waste is an influential factor in their food waste behaviour (Aschemann-Witzel et al. 2015). Among the factors that moderate intention to reduce one’s food waste, while attitudes and personal norms have a direct influence (Visschers et al., 2016; Stancu, Haugaard, and Lähteenmäki 2016), subjective norms, and perceived behavioural control have an indirect influence (Graham-Rowe, Jessop, and Sparks 2014; Stancu, Haugaard, and Lähteenmäki 2016). Furthermore, studies indicate that global environmental concern (GEC) influences individuals’ food waste behaviour (e.g., Watson and Meah 2012). Thus, all of these variables were included in the questionnaire. Items measuring participants’ intention, perceived behavioural control, attitudes, subjective norms, and personal norms were adapted from Visschers et al. (2016). New Environmental Paradigm (NEP) scale (Dunlap et al. 2000) was used to measure GEC.

2.2.5. Preferences for food waste interventions

The last section included questions measuring participants’ preference for food waste interventions. Hebrok and Boks (2017) categorised various design interventions according to seven major food waste drivers via a systematic literature review. These categories are information and awareness, technology and planning, leftovers and portioning, storage, packaging, food risk, and policy and regulation. One example intervention for each of these categories was prepared to communicate them to the participants. These are Waste diary; A mobile application and a smart bin that monitors household food waste, Smart stock management; A smart fridge that help creating shopping lists and inventory control, Food share; A social media platform for sharing (left-over) food with others, Freshness booster; A smart container that is specialised for storing different food types longer, Portion packaging; A packaging solution that supports consuming food in small portions, Expiry date reminder; A smart packaging that gives expiry date warnings via colour change, Waste tax; A system that regulates trash tax based on the amount of food waste (Figure 1). During the data collection, the participants were asked to select one intervention that they thought would help them reduce their food waste and state the reasons for their preference.

short-legendFigure 1.

2.3. Data analysis

Agglomerative hierarchical clustering analysis was used to identify user types. This analysis is preferred as it allows the identification of groups with a bottom-up approach (Kaufman and Rousseeuw 2005); that is, researchers do not need to determine the number of clusters before the analysis. The variables included in the cluster analysis were self-reported food waste, intention, attitude, behavioural control, subjective norms, personal norms due to their direct influence on household food waste behaviour (Ajzen 1991; Visschers et al., 2016; Aschemann-Witzel et al. 2015; Stancu, Haugaard, and Lähteenmäki 2016).

To prepare the data for cluster analysis, first, self-reported food waste was calculated, using the procedure applied in Visschers et al. (2016). The frequency of food waste was multiplied with the amount of food waste. For example, if a participant indicates that s/he wastes 2–3 portions of meat 3–5 times per week, his or her meat waste is calculated as 2.5 × 4 = 10 portions a week.³ A participant’s total food waste was determined by summing up all the waste for each food category. Second, the items measuring intention, attitude, perceived behavioural control, subjective norms, personal norms, personality traits, and GEC were converted into numbers: (1: strongly disagree, 2: disagree, 3: undecided, 4: agree, strongly agree).⁴ Then, the averages of corresponding items were calculated.

After clustering the data by using different linkage methods (single, complete, and average linkage, and Ward’s method), the final analysis was performed by using the Ward’s method. This analysis led to the identification of four clusters. Non-parametric statistical tests were used (i.e., Kruskal-Wallis Test and Mann-Whitney test) to identify significant differences between these clusters because the data was not evenly distributed between groups. As a final step, the frequencies of the remaining variables (i.e., age, socioeconomic status, sex, household type, food-related household practices) were used to further describe each cluster and differentiate them from each other.

3. Results

Cluster analysis led to four user types: Conservers, Considerates, Reluctants, and Prodigals. Table 2 presents the descriptive statistics for these user types. Table 3 presents their preferences for different interventions.

Table 2. Descriptive statistics for the user types

	Conservers	Considerates	Reluctants	Prodigals
Cluster size	75 (50%)	42 (28%)	24 (16%)	9 (6%)
Socio-demographics
Age (average)	41,4	38,4	32,3	29
Sex
Female	38 (51%)	31 (74%)	16 (67%)	6 (67%)
Male	37 (49%)	11 (26%)	8 (33%)	3 (33%)
SES
SES-A	13 (17%)	19 (45%)	4 (17%)	3 (33%)
SES-B	32 (43%)	10 (24%)	14 (58%)	2 (22%)
SES-C1	30 (40%)	13 (31%)	6 (25%)	4 (45%)
Household type
Lives with family	62 (83%)	32 (76%)	20 (83%)	4 (45%)
Lives with friends	6 (8%)	6 (14%)	1 (4%)	3 (33&)
Lives alone	7 (9%)	4 (10%)	3 (13%)	2 (22%)
Having children
Yes	39 (52%)	17 (40%)	5 (21%)	0
No	36 (48%)	25 (60%)	19 (79%)	9 (100%)
Personality traits
Extraversion	3,56	3,77	3,66	3,77
Conscientiousness*	3,94	3,86	3,47	3,50
Agreeableness	4,02	4,00	3,93	3,88
Openness	4,03	4,11	4,02	4,00
Emotional stability	2,97	2,76	3,02	3,00
Food related household practices
I am usually responsible for ...
planning for shopping*	3,51	3,98	2,92	2,89
shopping	3,95	3,86	3,50	3,56
cooking	3,48	3,67	3,54	3,11
storing	3,01	3,12	2,79	2,67
I do not buy food nearing its expiry date *	4,37	4,50	4,04	4,67
I give left-over food to animals.	4,27	4,21	4,04	3,78
I usually cook no more than I can consume. *	4,37	4,45	3,92	4,00
I store the left-over food in the freezer.	4,08	4,19	3,75	3,89
I reuse left-over food to prepare new meals. *	4,01	3,76	3,42	3,44
I usually buy seasonal fruits and vegetables. *	4,53	4,57	3,92	4,44
I buy food no more than I need.	4,35	4,29	3,96	3,78
I prepare a shopping list. *	3,59	3,88	3,17	2,89
I avoid buying food outside my shopping list. *	3,32	2,69	2,63	2,78
I know the exact amount of food I need to buy as I plan one week ahead what to cook. *	3,33	2,88	2,75	2,33
Self-reported food waste * (n portions a week)	2	9	19	38
Attitudinal factors
Intention*	4,65	4,56	4,13	3,91
Attitude*	4,16	4,29	3,70	3,84
Subjective norm*	4,09	4,17	4,12	4,08
Personal norm*	4,46	4,34	3,79	3,77
Perceived behavioral control*	4,00	4,03	3,55	2,95
Global environmental concern	3,50	3,56	3,58	3,27

*Variables where there is a significant difference between clusters, for the significance level of 0,05

Table 3. Intervention preferences of the user types

	Conservers	Considerates	Reluctants	Prodigals
Cluster size	75 (50%)	42 (28%)	24 (16%)	9 (6%)
Intervention preferences
Food share	16 (21%)	9 (21%)	3 (13%)	4 (45%)
Waste diary	14 (19%)	8 (19%)	6 (25%)	1 (11%)
Smart stock management	11 (15%)	9 (21%)	7 (29%)	1 (11%)
Freshness booster	11 (15%)	7 (17%)	2 (8%)	1 (11%)
Food tax	8 (11%)	4 (10%)	3 (13%)	1 (11%)
Expiry data reminder	10 (14%)	2 (5%)	1 (4%)	1 (11%)
Portion packaging	4 (5%)	3 (7%)	2 (8%)	0

3.1. Conservers

Conservers is the biggest user type, representing half of the participants (n:75). They produce the least amount of food waste (M: 2 portions a week). The majority of them indicated that they almost produce no waste (n:45). Conservers produce significantly less food waste than all other user types: Considerates, Reluctants and Prodigals. They have the highest level of intention to prevent household food waste (M: 4,65), which is significantly higher than Prodigals, and Reluctants. They feel the highest moral obligation to prevent food waste (M: 4,46), which is significantly higher than Reluctants and Prodigals. Furthermore, Conservers have significantly more perceived control over their food waste behaviour than Reluctants and prodigals.

Conservers include individuals who are usually in charge of shopping and planning routines. Since they usually plan what to cook one week before, they are more likely to avoid making unexpected purchases than Reluctants and Prodigals. In parallel, they are less likely to buy items that are not listed in their shopping list than Considerates and Reluctants. They tend to use leftovers in preparation of new meals. This tendency is significantly bigger than Reluctants.

Conservers tend to be more conscientious. On average, they show the traits of conscientiousness significantly more than Prodigals and Reluctants. A high majority of Conservers consists of families (83%) who either belong to the middle (SES-B) and low-middle socioeconomic status (SES-C1). Conservers represent a balanced sample in terms of sex and having children. Their intervention preferences are as follows: food share (n:17), waste diary (n:14), smart stock management (n:11), freshness booster (n:11), expiry date reminder (n:10), and portion packaging (n:4).

3.2. Considerates

Considerates is the second biggest user type, representing 28 % of the participants (n:42). With a weekly average of 9 portions of food waste, Considerates produce significantly more waste than conservers, and significantly less food waste than Reluctants and Prodigals. They have the highest positive attitude towards reducing their household food waste (M:4,29), which is significantly higher than Reluctants and Prodigals. They have the highest perceived control over their food waste behaviour (M:4,03), which is significantly higher than Reluctants and Prodigals. They have the highest social support to reduce their food waste (M: 4,17), which is significantly higher than Prodigals. Furthermore, they feel significantly more moral obligation to reduce their food waste than Reluctants and Prodigals.

Considerates are more likely to be responsible for planning practices compared to Reluctants and Prodigals. In parallel, they are more likely to prepare a shopping list compared to Reluctants and Prodigals. The majority of the Considerates is women (74%) and live with their families (76%). They represent a balanced sample in terms of having children and socio-economic status. Their intervention preferences are as follows: food share (n:9), smart stock management (n:9), waste diary (n:8), freshness booster (n:7), food tax (n:4), portion packaging (n:3), and expiry date reminder (n:2).

3.3. Reluctants

Reluctants is the third biggest user type, representing 16 % of the participants (n:24). Their self-reported food waste (M:19 portions a week) is significantly less than Prodigals, but significantly more than Conservers and Considerates. Reluctants have the least positive attitude towards reducing their household food waste (M:3.70), significantly less than Considerates and Conservers. Furthermore, they have significantly lower intention to reduce household food waste than Conservers and Considerates.

Reluctants are more likely to purchase food nearing its expiry date than all other user types; Conservers, Considerates, and Prodigals. In parallel, they are less likely to consume fresh fruits and vegetables compared to Conservers and Considerates. Furthermore, they are more likely to cook more than they can consume in a single meal compared to Conservers and Considerates. The majority of Reluctants is women (67%), and singles or families who do not have children (79%), and who represent middle socio-economic status (SES-B). Their intervention preferences are as follows: smart stock management (n:7), waste diary (n:6), food share (n:3), food tax (n:3), portion packaging (n:2), freshness booster (n:2), and expiry date reminder (n:1).

3.4. Prodigals

Prodigals is the smallest user type, representing 6% of the sample (n:9). They produce the highest amount of self-reported food waste (M:38 portions a week), which is significantly more than all other user types; Conservers, Considerates and Reluctants. They have the lowest intention to reduce their household food waste (M: 3,91), significantly lower than Conservers and Considerates. They have the lowest social support to reduce their household food waste (M: 3,38), significantly lower than Conservers and Considerates. They have the lowest perceived control over their food waste behaviour (M:2,95), which is significantly lower than Conservers and Considerates.

Prodigals are less likely to plan for shopping than Considereates, and are more likely to make unexpected purchases than Conservers. They are more likely to purchase seasonal food and avoid purchasing food nearing its expiration date compared to Reluctants. Prodigals represents a balanced sample in terms of household type, socio-economic status, and sex. However, the distinguishing characteristic of this user type is that none has children. Their intervention preferences are as follows: food share (n:4), smart stock management (n:1), waste diary (n:1), food tax (n:1), portion packaging (n:1), freshness booster (n:1), and expiry date reminder (n:1).

4. Discussion

Behaviour change is a complex and challenging that requires the consideration of many interrelated factors. Designing successful interventions that encourage individuals to reduce household food waste depends on a holistic and collaborative approach where various stakeholders like governments, industry and households work together at various scales individual, community and national (Tudor et al. 2011). Designers have a central role in this collaborative process as they can turn ideas and insights into materialised solutions, e.g., designing smart waste bins that track household food waste in a neighbourhood and provide information for the local government and policymakers to shape their waste management strategies. However, as discussed previously, selecting the right target user group and deciding on the right intervention type is also critical for motivating households to comply with these strategies (Coskun and Erbug 2014) The remainder of this section elaborates on how this study support designers in these tasks, i.e., how designers can utilise the user types when designing food waste interventions.

4.1. Selecting the target group based on user types

When designing food waste interventions, designers can select the target groups by looking at previous work, exploring user characteristics that influence household food waste behaviour. For example, studies revealed that food waste is negatively correlated with intention, attitude, perceived behavioural control, subjective norms and personal norms (Aschemann-Witzel et al. 2015; Visschers et al., 2016), age (e.g., Stancu, Haugaard, and Lähteenmäki 2016; Secondi, Principato, and Laureti 2015), having children (e.g., Visschers et al., 2016; Schanes, Dobernig, and Gözet 2018; Parizeau, Von Massow, and Martin 2015) and unconscientiousness (e.g., Swami et al. 2011). Thus, designers can select individuals with high intention to, attitude towards, control over, social support, and moral obligation to reduce their food waste. They can also target older and conscientious people who do not have children.

However, selecting a target group by consulting with previous studies might be problematic because sometimes these studies reveal conflicting results. For example, while some studies indicate that households with children tend to produce more food waste (e.g., Schanes, Dobernig, and Gözet 2018), others found that single households waste the most (Jörissen, Priefer, and Bräutigam 2015). In another example, although some studies indicate a positive correlation between income level and food waste (Parfitt, Barthel, and Macnaughton 2010; Thi, Kumar, and Lin 2015), there are a few studies that indicate that low-income households can also produce food waste as much as high-income households (Porpino, Parente, and Wansink 2015). Thus, it is critical to identify different user types and their characteristics before designing an intervention.

Designers can identify user types in a target population by following the same methodology described in this study, i.e., collecting data about users’ food provision and waste practices, and socio-demographic, behavioural, and psychographic characteristics that influence these practices and clustering the users based on this data. However, this may not be feasible due to the tight deadlines of many design projects. When designers have limited time and resources to do extensive user research and analysis, they can use previously identified user types, as in this study. In such a case, they can use the size and average food waste as a reference to select the right user type. For example, they can start selecting the largest user type in a population (e.g., Conservers) and continue comparing this user types food waste with the population’s average food waste. When the biggest user type involves individuals whose food waste is high, selecting them would be a good choice. However, when this user type consists of individuals who produce less food waste and have already adopted many food waste prevention practices, as in the case of Conservers, selecting them may not be the best solution. Hence, the alternative could be selecting the user types that involve fewer people but produce more substantial food waste (e.g., Reluctants).

4.2. Selecting the design intervention based on user types and their intervention preferences

Selecting the right user type is not always sufficient to create interventions that encourage behaviour change because users’ characteristics influence their compliance with an intervention. For example, being able to track the amount of food waste (e.g., waste diary) may encourage behaviour change for a user who has a high intention to reduce his or her food waste (e.g., Conservers) but may not encourage behaviour change for a user without such an intention (e.g., Prodigals). Thus, as a next step, designers should consider the intervention preferences of different user types. For instance, smart stock management can be selected for Reluctants, and food share can be selected for Conservers.

Selecting one specific user type and a suitable intervention may not be possible when designers work with bigger samples that include diverse user types. For example, a target sample consisting of university students can include both Reluctants and Considerates, or a big neighbourhood can include all the user types from Conservers to Prodigals. Furthermore, selecting one specific user type is not always desirable. On the one hand, motivating Prodigals, who produce large amounts of food waste but do not intend to reduce it, would be very hard. On the other hand, motivating Conservers, who produce less waste and have a high intention to reduce it would be easier; however, this would not create a significant change in total household food waste since their average food waste is already lower than the others. Thus, a combination of two or more user types would increase an intervention’s impact on reducing household food waste. For example, the study results showed that food share is the first popular intervention type for Conservers, Considerates, and Prodigals and that Waste diary is the second most popular intervention for Conservers, Considerates, and Reluctants (Table 3). It seems that the combination of these interventions (Waste diary and Food share) would be more likely to influence all the user types in this study’s sample.

5. Limitations

Although the study sample includes participants with diverse characteristics (age, gender, socio-economic status, household type), it does not fully represent the Turkish population. For example, it does not include low-income households even though they constitute almost 25% of the population (TUAD 2012). The rationale behind selecting people from low-middle, middle, and upper-middle socio-economic status is the literature indicating that low-income households tend to produce less waste. Nevertheless, the limited extent of the generalisability of the findings to the Turkish population should be considered.

6. Conclusion

Minimising household food waste requires integrating various perspectives into intervention design, not only from the design but also from other fields like sociology, psychology, and policymaking. Besides collaborating with experts from these fields, during the intervention design process, designers should consider many psychological, behavioural and socio-demographic factors that influence individuals’ food waste behaviour as well as the variety in these factors to better tailor an intervention to the needs of different user groups. This brings a great challenge for designers who usually imagine a single user type (a persona) when designing a product or a service. This paper’s main contribution is to help overcome this challenge in reducing household food waste, by summarising the factors influential in household food waste behaviour, explaining how they can be used to identify different user groups and discussing how designers can select the right target user groups and the appropriate intervention types.

A design team can utilise the user types from the current study when there are time and resource constraints, similar to the generic user types identified in previous work. Ideally, they should start with identifying relevant user groups by following the methodology proposed here since a user group identified in a setting may not be directly applicable to another. Furthermore, the team should target all of the user types in a target population as much as possible and use a combination of different interventions types to create an impactful solution. However, identifying the user groups alone does not guarantee an effective behaviour change. To really understand what works and what does not work for particular user type(s), one needs to implement a design intervention and measure its effectiveness. As this paper has already illustrated the process of user type identification, presented four user types for household food waste based on data from Turkish context, and suggested intervention types for these user types, future studies should advance on this work by identifying user types in other contexts, designing new interventions, and evaluating their impact on food waste prevention.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Notes on contributors

Aykut Coskun

Aykut Coskun is an Assistant Professor of Media and Visual Arts at Koç University. He is also a design researcher at Koç University Arcelik Research Center for Creative Industries. He received his B.Sc. M.Sc. and Ph.D. from Middle East Technical University Department of Industrial Design. He attended Carnegie Mellon University Human-Computer Interaction Institute in 2015 as a Fulbright Visiting Student Researcher. His current research focuses on design for behavioral change, design for sustainability, and design for wellbeing.

Notes

1. There are six socio-economic groups in Turkey. These are SES-A (4%), SES-B (9%), SES-C1 (22%), SES-C2 (29%), SES-D (28%) and SES-E (9%).

2. Visschers et al. (2016) measured self-reported food waste for meat, dairy and bakery products arguing that they are the most wasted food types. In his study, fruits and vegetables were added to this list.

3. One portion is considered as a handful of food.

4. Despite the criticism towards treating Likert scale as interval scale, Norman illustrates that parametric and non-parametric statistics can be used with Likert data without reaching the wrong conclusion (2010).