Currently, the estimation of food consumption, nutrient intake, and food safety are growing in interest worldwide. So, in its latest published study, the Food and Agriculture Organization (FAO) stresses the urgency for the needed broader food systems transformation, which is currently the focus of global attention (1). Therefore, if the focus is on consumption, the need to raise awareness and promote a sustainable diet has been gaining ground. Diets are drivers of sustainable food systems (2) and inextricably link human health and environmental sustainability (3). So, a sustainable diet comprises different dimensions, such as health, ecological and social, all within the framework of political, economic, and environmental contexts that limit or facilitate changes in diets and food systems (4).

In this study, focused on the ecological and health dimensions of a sustainable diet, we will analyze the consumption of organic products, consider the degree of food processing, and identify consumer profiles for each case. All this is in a particular context, such as that of Spain. Organic production refers to the agricultural and animal husbandry management procedure of food production whose objectives ensure a viable and sustainable agricultural management system that combines production techniques aimed at the optimal and sustainable use of natural resources, reducing the environmental impact (5). In a more multifaceted conception, organic production combines tradition, innovation, and science to benefit the shared environment and promote fair relationships and good quality of life for all involved (6).

At the world level, organic production and consumption of organic food follow a growing trend (7, 8). A quick review of the evolution of the latest published data is shown in Table 1.

An increasing number of consumers generally show positive perceptions of organic food products, understanding that they are healthier than their conventional counterparts and better for the environment (9, 10). There is a broad consensus that organic food production and consumption offer health and sustainability upsides (11–13).

In this context, sustainable diets protect and respect biodiversity and ecosystems, are culturally acceptable, accessible, economically fair and affordable, and nutritionally adequate, safe and healthy while improving natural and human resources (14). The issue has received considerable attention in Europe. From 2021, The European Council is pursuing the goal of transforming the way food is produced and consumed in Europe to reduce the environmental footprint of food systems and strengthen their resilience against crises.

Organic products are regulated by a series of regulations and standards by countries or areas. In this research, the definition of an organic product is in line with the description of European legislation. In 2007, the Council of the European Union (EU) agreed on Council Regulation 834/2007 (15), which establishes the principles, objectives, and general rules of organic production and defines how organic products should be labeled. Today, the Regulation in force is (EU) 2018/848 (16) on organic production and labeling of organic products. Specifically, an organic product is defined as a product resulting from organic production; organic production uses methods that comply with the Regulation at all stages of production, preparation, and distribution. The products of hunting or fishing wild animals are not considered organic products.

Specifically, the organic labels certify that a product complies with organic quality standards. So, consumers can be guided by this label when purchasing an organic product and ensure that the components comply with the organic standards. The EU organic logo can only be used on products containing at least 95% organic ingredients and respects other strict conditions for the remaining 5%. The same component cannot be present in organic and non-organic forms (17).

On the other hand, this research considers the classification of foods according to their degree of processing since many studies have evaluated their possible effect on consumer health. The bibliography shows that there are currently seven possible food classification systems based on the degree of processing. NOVA and SIGA are the two most globally applicable classification systems, i.e., they are not country or region-specific (18).

NOVA system has been used in most studies to analyze and document the effect of ultra-processed food consumption on various diseases or markers of disease, health, or mortality. Several cross-sectional and longitudinal studies have been conducted, with many indicating a direct relationship between increased consumption of ultra-processed foods and cardiovascular disease, obesity, type 2 diabetes, cancer, and generally a higher mortality risk (19–22).

Therefore, it is worth noting that NOVA has been used in nutrition research to assess the healthiness of diets and food environments where food composition data has not been readily available (23). It is the case we are dealing with in this paper. NOVA classifies all foods and food products into four groups according to their nature, the scope, and the purpose of the industrial processing to which they are subjected (24, 25). Currently, four groups are described (see Table 2), considering the physical, biological, and chemical processes to which the food is subjected before being consumed (26).

It should be noted that there are foods with organic certification in any of the four NOVA categories, and not only in NOVA 1 of unprocessed or minimally processed foods. For example, organic oils (NOVA 2), organic wines (NOVA 3), organic frozen pizzas (NOVA 4), and other possible processed or ultra-processed foods are marketed in the Spanish market.

So, not all products labeled as organics are healthy, as may be the case with some ultra-processed foods. For example, are ultra-processed products-that tortillas chip or candy- that we can buy with European organic labeling if 95% of their agricultural ingredients are of organic origin.

As regards the country studied, it should be pointed out that Spain is a Mediterranean country. The Mediterranean diet is characterized by high consumption of olive oil, vegetable products, fish and seafood; low consumption of dairy, meat and meat products; and a moderate ethanol intake (28).

As organic food grows in popularity, there is a growing need for academic research on consumer profiles, habits, and attitudes toward organic food (29). Previous research has addressed the relationship between organic food and sustainable consumption (30). That is, to understand organic consumption as a desire to be more sustainable. People are attracted to organic food because it is considered safe, healthy, and environmentally friendly (31), which impacts their daily-consumer preferences (32–34). However, although many studies show that organic food consumption is growing worldwide, there are still inconsistencies arising from differences in how people perceive organic food, their motivations, attitudes, or willingness to purchase (30, 35). These inconsistencies reveal a gap between intentions and attitudes at the center of behavioral research in this field (9, 36–38). Aspects such as price (39–44) and, to a lesser extent, availability and lack of knowledge and information (45–49) can be a barrier to buying organic products. Nevertheless, understanding that gap is made difficult by variations in consumer profiles. In previous research papers, the socio-demographic variables studied to profile consumers of organic products are gender, age, income, marital status, the existence of children in the household, education, and place of residence. The results are not conclusive. So, discussion of the pertinence of these variables for distinguishing between consumers of organic food and other consumers has not always enjoyed general agreement. There are studies suggesting no significant correlation between socio-demographic variables and purchase. Others offer a significant and negative correlation, while others have found a significant and positive correlation. A review of the literature on the different socio-demographic factors that impact the purchase of organic food is summarized in Table 3.

These studies indicate a relationship between socio-demographic variables and the consumption of organic products. However, the findings reveal that there is no identified prototype profile of buyers of organic products.

To promote a sustainable diet, it is essential to know who is buying organic products and what kind of organic products they are buying, given that the previous studies reviewed do not yield conclusive results on the characteristics of organic food consumers. A set of sequential research questions has been designed, not assuming a priori that there are differences in who buys and what is eaten. So, are three questions: (1) to explore whether there are statistically significant differences in the socio-demographic profile of Spanish households that buy organic products and those that do not; (2) to investigate whether there are statistically significant differences in the degree of processing of organic and non-organic foods purchased by the households; (3) to discover the association between socio-demographic characteristics and the purchase of organic products in the four categories NOVA of processed and unprocessed foods. The weight justifies the choice of Spain as the object of research that its organic production represents worldwide: the third place on the world podium of the largest organic agriculture producers in 2019, only surpassed by Austria and Argentina in the ranking. In terms of consumption, it is in tenth place, with the USA, Germany and France being the most prominent organic food consumers (8).

In the subject at hand, food purchases can be obtained mainly through different procedures, depending on the consumption unit studied in the data collection -household or individual-, the methodology used in the collection of information -stable sample or not over time, automatic or manual data collection-, the desired level of statistical significance -random or convenience sampling- and the nature of the entity that performs them -public agencies or private companies-. Fundamentally, the consumption data of a population can be obtained through panels or surveys carried out by public or private organizations, on households or individuals, chosen randomly or not, who undertake to collect all purchases made by manually recording them in purchase diaries or capturing the data using automated procedures.

In the case of Spain, the two primary official sources that collect food purchases were reviewed: the National Household Budget Survey and the National Household Food Consumption Panel. Although the first of these sources is available at the micro-data level, it could not be used because it does not measure organic food consumption. In contrast, the Household Consumption Panel of the Ministry of Consumption, Fisheries and Food has started to count the purchase of some categories of organic food in 2021 (72).

Another option, such as purchasing data from private organizations, has been discarded due to a lack of budget. The use of data from different sources has not been contemplated since some recent studies have detected that using different methodologies in data analysis may complicate or impair the combination of data from various sources (73).

The Household Food Consumption Panel (purchasing panel) is included in the National Statistical Plan. It represents all households in Spain and is statistically significant at the national level. The information is collected through a sample of 12,500 households, which are asked to record all their daily food purchases. Purchases are recorded daily using an optical barcode reader, which minimizes reporting errors. Food and beverages purchased to supply the household are collected every day of the year, at all times of purchase and by any household member. Products that do not have a barcode are collected by employing a code book that allows all types of products to be declared automatically. Self-consumption is not included (72). The composition of the average household basket in Spain, aggregating the data from the purchasing panel, is shown in Table 4.

The official statistics present the information disaggregated according to the following socio-demographic criteria: the age of the person responsible for the expenditure, size of the municipality where the household is located; social class by income; the presence of children in the home by age; employment situation of the person responsible for purchasing; the number of persons in the household; household life cycle and autonomous territories. In this article, we have not studied information about the autonomous region of ascription (we were only interested in Spain as a whole). Neither has it been used family life cycle, as it is a variable that combines the age of the person responsible for purchasing, the number of family members, and the presence of children in the household. Unfortunately, it should be noted that the official Spanish statistics do not provide information on gender and education level.

Regarding the environment, the data on household consumption of organic products was incorporated for the first time in January 2020, so only one year’s information is available. On the other hand, data are only available for 19 organic food categories without disaggregating into subcategories. In this study, we have started from the information published and revised for 2020, choosing data measuring quantities purchased–kilograms. The numerical data available from this official statistic are totals−they can only be obtained in aggregate form in Excel files−it is not possible to get individualized micro-data by household.

Thus, the information necessary for this work has been extracted in the following way:

Household consumptions from official data are total, not per capita. Since this paper aims to analyze the differences between organic and non-organic consumption, the data from national statistics have been kept, not transformed into an average per person. However, the number of household members is one of the characteristics analyzed.

The 19 food categories, provided by official statistics, were assigned to the appropriate groups of NOVA. This assignment was done independently by two authors (BG and RM). Discrepancies between the two were recorded, and a final selection was agreed upon through consultation.

Hence, for the subsequent analyses, the kilograms of food purchased per household in Nova families have been added, which has shortened an initial list of 19 to only four groups. The quantities have been reduced to hundreds of thousands of kilos to facilitate the readability of the results (Table 5).

The use of contingency tables is the most common technique to study the relationships between variables taken in pairs. In this paper, we have chosen to analyze the aggregate data available according to the mosaic plots instead of presenting only contingency tables.

Mosaic plots, proposed by Hartigan and Kleiner (73, 74) and further developed (75), can help visualize contingency tables, even those complex ones consisting of many categorical variables. This kind of display can be beneficial in understanding simple and complex associations among categorical variables. So, the contingency tables can (but need not) be easy to understand simply by scanning the numbers (76), but this need not be the case with more complex tables, especially when the variables are associated, as is our case. In such cases, mosaic plots (or a fourfold visualization for 2 × 2 tables with an additional stratum variable) can help to capture associations between categorical variables.

To better understand the results presented in the following section, it is helpful to note that a mosaic plot is a graphical display of the cell frequencies of a contingency table in which the area of boxes of the plot is proportional to the cell frequencies of the contingency table. Both the size and position of the rectangles are relevant to the interpretation of the mosaic chart, making them one of the most advanced charts (77). For investigation, the mosaic function results in a mosaic plot in which the shading of the cells represents the Pearson residual for independence: red for negative values, blue for positive values, and more substantial shading for higher absolute residual values.

Analyses have been combined and performed sequentially: first, we checked, using contingency tables and the associated chi-square statistic, the existence of a one-to-one relationship between each socio-economic variable of the households and the purchase or not of organic products; subsequently, the distribution of food consumption according to NOVA was explored using mosaic plots and Pearson’s p-values of the residuals. Finally, cross-checking of socio-demographic categories with NOVA categories was conducted to detect possible differences in the case of organic NOVA products.

All analyses were carried out using R software.

It has not been possible to investigate multivariate relationships. Although we contacted the public body that publishes this information, obtaining the micro-data from the consumer panel was impossible. Only aggregated data were available, so applying multivariate analysis methods was impossible.

The analysis of the differences in the consumption of organic and non-organic foods was always significant; statistically significant differences were found for the six socio-demographic characteristics studied in Spanish households (p = 0,000 in each case). A closer inspection of the results, variable to variable, shows:

Conclusively, the findings are statistically significant and show differences in the socio-demographic characteristics of the households buying organic food, at least in some aspects of the six variables studied.

As a next step, it is helpful to examine whether there is a relationship between organic food consumption and the group consumed according to the Nova classification. Figure 1 contains the distribution of consumption according to both criteria, in absolute figures and percentages, taking in the second case as a reference the type of consumption (organic or not) and the mosaic plot with residual-based shading (organic × NOVA).

Calculating Pearson’s statistic, its p-value of 0.000 indicates that these characteristics are related so that the distributions of consumption by the NOVA group are not similar in the case of organic and non-organic products.

The differences found are as follows:

The results will be extracted by comparing two mosaic plots associated with each of the contingency tables of the combination “socio-demographic characteristic × organic or non-organic × NOVA.” The six Mosaic plots for NOVA and socio-demographic characteristics are presented in Figures 2–7.

There are a few differences in the NOVA classification for the spectrum of organic foods, segmented according to each of the six socio-demographic variables available. However, quite a few differences are observed when the analysis is compared with Spanish households’ consumption of non-organic food. Therefore, the following paragraphs show the results for each organic category.