The following figure integrates the fairness framework presented by Busch and Spiller (2016) with the concept of Environmental fairness, included by the Food Ethics Council (2020) in the Food Justice report (Figure 1). It provides an overview of the concept of fairness and its dimensions from a general perspective. It incorporates the concepts of distributive fairness, interpreted as the fairness of price received (Adams, 1965; Bolton et al., 2003; Haitao Cui et al., 2007; Gielissen and Graafland, 2009; Zitzmann and Dobhan, 2010; Briggeman and Lusk, 2011; Lu et al., 2021), procedural fairness, the perceived fairness of the procedures used to determine price distributions (Thibaut and Walker, 1978), interactional fairness, the quality of employees' interpersonal treatment (Bies and Moag, 1986; Colquitt et al., 2001), and environmental fairness, respect for the environment (Food Ethics Council, 2020). Past research conceptualised interactional fairness as the third dimension of fairness (Colquitt et al., 2001) or as the social part of procedural fairness (Folger and Konovsky, 1989).

The concept of distributive fairness was introduced by Adams in 1965 who stated that if the relationship between the single actor's inputs and outputs is balanced then the outcome is perceived as fair (Konovsky, 2000). The concept of outcome is often defined as the “price” that each actor in the chain receives for their products. In general, a price is fair when all parties are satisfied. This is why the concept of distributive fairness is often associated with the concept of fair price distribution (Haitao Cui et al., 2007; Lu et al., 2021).

Even in the case of distributive justice, there are different perspectives. From a producer's perspective, fairness concerns the price they get for their products (Hellberg-Bahr et al., 2012). From a consumer perspective, the distribution is fair when producers get the highest share and the rest is distributed equally to other stakeholders (Busch and Spiller, 2016). Consumers believe that they are treated fairly when the product they buy gives them good value for money (Nguyen and Klaus, 2013). Price increases are also seen as fair if small or poor stakeholders get a benefit rather than large and powerful ones (Gielissen and Graafland, 2009).

The concept of procedural fairness is commonly linked to agreements, negotiation processes and bargaining power. It was introduced by Thibaut and Walker (1978) and it deals with the procedures used to achieve outcomes rather than the actual outcome achieved. In fact, procedural fairness can be defined as equity related to the procedures used to achieve outcomes (Konovsky, 2000; Korsgaard, 2002). Outcomes will be perceived as fairer if the process that generated them is considered fair by those who participated in the decision making-process (Folger, 1977). A fair procedure must be consistent, impartial, open to all, transparent and credible (Lewicki and Bunker, 1995; Bolton et al., 2005). A procedure is considered unfair when the bargaining process does not take place or when it is perceived as unfair (Thal, 1988; Druckman and Wagner, 2017).

The literature does not clarify who is responsible for ensuring certain standards at economic and social level. It is not clear whether it should be the retailers or processors who guarantee a fair price to farmers for their products (Gielissen and Graafland, 2009; Busch and Spiller, 2016) or the consumers themselves by paying a higher price (Gielissen and Graafland, 2009). Policy makers also have a great responsibility in ensuring a fair food chain by creating policies that ensure farmers a decent livelihood and promoting information campaigns for more careful and conscious choices (Busch and Spiller, 2016).

Whether it is the third dimension of fairness or an aspect of procedural fainess, interactional fairness is about the intention behind every action (Rabin, 1993). Introduced by Bies and Moag (1986), the concept of interactional fairness was subsequently subdivided into interpersonal fairness and informational fairness (Greenberg, 1990). The former refers to the honest and respectful behaviour of chain trading partners. The latter refers to the quantity and quality of information shared (Busch and Spiller, 2016).

In the agro-food sector, integrity is mainly about how producers are treated. Everyone must be guaranteed a job that allows to have a good standard of living, both economically and in terms of safety, and that guarantees equal opportunities such as adequate education. No exploitation, intimidation or abuse should be accepted (Fairtrade, (n.d.); Food Ethics Council, 2020). Gender policies should be developed and support programmes for disadvantaged people should be guaranteed (Fairtrade, (n.d.); FAO, 2021). Any kind of discrimination, be it gender, marital status or ethnicity, should be avoided (Fairtrade, (n.d.); Food Ethics Council, 2020). Furthermore, an ethical agro-food chain should prohibit forced and child labour (Fairtrade, (n.d.); FAO, 2021), support the community by encouraging the work of small-scale producers (Chang and Lusk, 2009), guarantee producers long-term “contract” [Fairtrade, (n.d.)], and finally, ensure facilities to allow producers to manage the Premium price (FAO, 2021).

Often included within the concept of sustainability, environmental protection is included as a fundamental part to describe fairness in food systems (Food Ethics Council, 2020; Zamzow and Basso, 2022).

Past literature does not present a uniform picture in describing consumer interest in environmentally sustainable products. According to Kit et al. (2018), the environmental-conscious food market expansion is due to consumers' growing interest in the environment. However, although the topic is much debated today, the demand for eco-friendly products is lower than would be expected (Kamalanon et al., 2022). A study by Moslehpour et al. (2021), shows that it takes a tangible element like sustainable packaging to get positive attitude towards environmentally sustainable products. Attitudes, environmental concerns, environmental knowledge, and subjective norms, are among the major positive drivers of green purchase behaviour (Young et al., 2020; Wijekoon and Sabri, 2021). Some organisations, such as FAO or Fairtrade, deeply addressed the environmental topics setting various key issues for agro-food system regarding the environment. A much-debated topic when it comes to environmental ethics is the use of genetically modified organism (GMO). Opinion on genetically GMOs is still controversial (Wilson, 2021). The greatest concerns relate to the potential danger to human health or the environment (Fairtrade, (n.d.); de Olde and Valentinov, 2019; Peano et al., 2019; Zimmerer et al., 2019; FAO, 2022), but also to how right it is to 'unnaturally' alter nature (Weale, 2010). The naturalness of food is in fact perceived as a positive aspect (Rozin et al., 2004; Korzen et al., 2011; Román et al., 2017). Carbon footprints, e.g., carbon emissions from energy used in the manufacture of fertiliser and for transport, should be reduced. Waste, whether of food, water or materials, should be minimised (Fairtrade, (n.d.); Bagherzadeh et al., 2014; de Olde and Valentinov, 2019; FAO, 2021). Responsible use of resources should be ensured, especially in reducing the water footprint, i.e., how much water is used in food production and processing (Fairtrade, (n.d.); Peano et al., 2019; FAO, 2021).

The methodological framework of the research included four different steps (Figure 2): Step (1) literature review, Step (2) semi-structured online interviews with consumers and validation of the items list, Step (3) survey finalisation and submission, and Step (4) data elaboration. The first two steps allowed for the creation of the final questionnaire to be submitted to consumers (third step), while the fourth step lead to consumers' fairness perception and relevance in food purchasing.

In particular, the first step of the present study aimed at creating an initial group of phrases presented as items to capture the potential different aspects of fairness in the agro-food chain retrieved from several sources. The websites of Fairtrade International, Food Ethics Council, and Food and Agriculture Organisation (FAO) were consulted in addition to the most relevant academic studies on the research topic retrieved from Scopus database. Since the concept of fairness does not have a clear and unique definition, the purpose of step 1 was to collect aspects potentially related to a broad idea of fairness, focusing on the consumer's point of view. These aspects were then organised in an Excel file and divided into macro-dimensions, dimensions and sub-dimensions in which fairness can be classified. This step allowed the creation of an initial pool of 90 items. After merging and removing duplicate, redundant or non-applicable items, the authors' selection process brought to a set of 39 items.

The second step consisted in organising a series of online interviews with consumers. The aim was to select the fairness items and to add those not identified by the literature. Due to the COVID-19 global pandemic, the foreseen interviews were held online and involved a limited number of consumers (Burton and Bruening, 2003; Stewart and Williams, 2005; Tuttas, 2015; Kite and Phongsavan, 2017). Semi-structured interviews with 11 consumers were organised at the end of February 2021 and included open-ended questions asking about the perception of the concept of fairness in food chains. Consumers were contacted through the mailing list of an agro-food company selling and distributing local food. In order to invite consumers to participate, two reminder e-mails were sent at intervals of 1 week. Participants received a 15% discount on a minimum purchase of €30 as a reward for participating.

The purpose of the third step was to test and finalise the survey structure and submit it to consumers. Three items were added based on interviewees feedback reaching a final pool of 42 items (Table 1). The full set of items was reorganised into dimensions for easier understanding by consumers. The list of items was then tested with 4 experts in consumer food behaviour to refine unclear questions and develop a robust data collection instrument. After suggested fine-tuning the survey was submitted to consumers. The online survey has been administered with the support of Qualtrics, (Version number: March 2021; Qualtrics, Provo, UT, USA), an online data collection software.

The fourth step consisted in data analysis. First, an Exploratory Factor Analysis (EFA) was conducted to identify the various dimensions of fairness. Then, an ANOVA was conducted to analyse the relationship between sustainable consumption and socio-demographic characteristics.

The survey includes three main sections. The first section includes a list of 42 items to explore consumers' perceptions of fairness in the food sector. The items that share the same theme were grouped together into factors. Consumers were asked to provide their rating for each item with a 7-point Likert scale (from 1 = strongly disagree, to 7 = strongly agree). In the second section, consumers were presented with 13 factors potentially influencing their purchasing behaviour like attributes related to the environment, economic or product characteristics (e.g., taste, brand, etc) (Table 2). They were asked to rate (from 1 = not at all important, to 7 = extremely important) according to what extent they take those characteristics into account when buying food. The objective is to understand to what extent ethical aspects are among the purchasing drivers. The third section is about consumers' socio-demographic information (age, gender, nationality and members of family working in the food system).

The survey was distributed between March and July 2021. To increase the number of responses, the questionnaire has been distributed through various channels, such as a local agro-food companies, agro-food networking association and different online platforms such as Instagram, Facebook, and WhatsApp. Research team sent reminders during the following 2 weeks to maximise survey responses. The data collection ended when researchers observed that the survey promotion campaign was progressively yielding a lower number of responses. All of the respondents were provided with a participant information sheet signed an online consent form detailing their rights. By denying the consent forms, the questionnaire ended automatically. Moreover, participants' anonymity was guaranteed as no personal data were required.

The survey was filled in by 626 consumers. Data cleaning, which included rejecting questionnaires that were < 80% completed, yielded a final convenience sample of 529 questionnaires used for data analysis. The sample included mostly women (73%), and respondents had an average age of 34 years old, were Italian (98%), had three family members on average. More than 80% of them have no members working in the agri-food sector (Table 3).

Data elaborations were performed with the support of the software SPSS (IBM, version 27, Armonk, NY, USA). Data analysis followed three steps. The first step aims at understanding the perception of fairness in the agro-food chain. The first set of forty-two items of the questionnaire on the consumers' perception of fairness in food chain was processed using the EFA (Table 4). Principal Components Analysis (PCA) was used as an extraction method. The Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO) value was 0.880 therefore, above the required level of 0.6 (Tabachnick and Fidell, 2012); and the Bartlett's Test of Sphericity value was significant, p < 0.001. A Varimax rotation was performed to clarify and simplify the results of factor analysis. Items with factor loadings below 0.5 were excluded. The EFA grouped twenty-two items into five multi-items components (identified cumulated variance ~57%). Considering the number of missing values in the variables included in the factor analysis, the pairwise method was adopted. The pairwise method occurs when the statistical procedure uses cases that contain some missing data. This procedure does not exclude responses with missing data. The choice of factors was made on the basis of the Eigenvalue criterion being higher than 1. The reliability of each factor was checked with Cronbach's Alpha (CA) coefficients and considered acceptable (Ponterotto and Ruckdeschel, 2007; Bassioni et al., 2008; Wachter et al., 2012). The mean value was then calculated for each factor.

In the last phase, data analysis aimed at understanding if consumers' socio-demographic characteristics and fairness perception (independent variables) influence consumers' perceived importance of food with fairness-related characteristics (dependent variables). The “fair price for farmers” attribute was cross-analysed using ANOVA test with socio-economic characteristics and fairness factors previously identified by the EFA. Those variables that revealed statistical significance after ANOVA were further analysed to see if consumers' socio-economic characteristics and perception of fairness played a role in influencing consumers' perceived importance of fairness attributes. Factor mean values were dichotomized as above vs. below 4 within the 7-point Likert scale to assess how the perceived importance of fairness attributes of products changes among those with a high and low perception of the concept of fairness as “short supply chain” and “fair price”.

The research identified the fairness-related aspects and then sought to understand which aspects influence the perceived importance of fair attributes of food products. Results identified five factors defining consumers' perception of fairness (Table 4):

The factors' mean values provide insights on the consumers' fairness perception along the agro-food chain. The most relevant factors are “Fair price” (mean 6.4; SD 1.0) and “Environment” (mean 6.2; SD 0.8). Ensuring producers receive a fair price for their products is the most important aspect for consumers when it comes to ethics in agro-food chains. Soil protection is of prime importance to consumers who prefer a natural product with no modification of colour, shape or appearance for commercial purposes. Then, with decreasing importance, “Working condition” (mean 5.9; SD 1.0), “Networking” (mean 5.5; SD 1.0), and “Short chain” (mean 5.2; SD 1.2) contribute to the conceptualisation of a fair agro-food chain.

This section provides the results on the importance of product attributes when consumers are effectively purchasing food. The most important attribute is the taste (mean 6.0; SD 1.1), followed by the seasonality (mean 5.8; SD 1.3), the origin of the product (mean 5.6; SD 1.4), the sustainability in terms of environment (mean 5.4; SD 1.6), the value for money (mean 5.2; SD 1.5), fair price for farmers (mean 4.9; SD 1.8), habits (mean 4.7; SD 1.5), promotion or offers (mean 4.4; SD 1.6), packaging (mean 4.4; SD 1.8), and vegan or vegetarian (mean 4.0; SD 2.1). Product nutritional or health label, the lack of time to groceries, and brand are the least important in the choice of food products (mean 3.8, 3.4, and 3.3 respectively; SD 1.9, 1.8, 1.7 respectively). Standard deviation values support there is consistent consumers' perception of the relevance of taste and seasonality, but less for vegetarian or vegan attributes. Results support that environmental sustainability is more important to consumers than farmers' receiving a fair price for their products.

Results indicate that consumers' socio-demographic characteristics and fairness perception moderately impact on the consumers perceived importance of fair attributes. “Short Chain” and “Fair price” are the dimensions of fairness most affecting perception of importance of fair characteristics. The key socio-demographic factor influencing perceived importance of fair attributes is age. Results support that older consumers tend to value fairness more than younger consumers. Moreover, consumers with higher scores for the “short chain” and “fair price” factors consider a fair price for farmers to be more important when buying products (Table 5).

Table 5 Influence of socio-demographic characteristics and perceptions of fairness on perceived importance of 'fair price for farmers' attribute.