Complexity of nutritional behavior: Capturing and depicting its interrelated factors in a cause-effect model

ABSTRACT

The aim of this article is to demonstrate the complexity of nutritional behavior and to increase understanding of this complex phenomenon. We developed a cause-effect model based on current literature, expert consultation, and instruments dealing with complexity. It presents factors from all dimensions of nutrition and their direct causal relationships with specification of direction, strength, and type. Including the interplay of all relationships, the model reveals cause-effect chains, feedback loops, multicausalities, and side effects. Analyses based on the model can further enhance understanding of nutritional behavior and help identify starting points for measures to modify food consumption.

KEYWORDS:

• Cause-effect model
• complexity
• nutritional behavior
• strength of causal relationships

Introduction

Nutritional behavior is “the sum of all planned, spontaneous, or habitual actions of individuals or social groups to procure, prepare, and consume food as well as those actions related to storage and clearance. In this context, the term ‘nutritional behavior’ not only refers to influencing factors but also to health, environmental, social, and economic implications along the entire product chain from farmer to consumer” (Department of Nutritional Behavior 2010; based on definitions by Oltersdorf 1984, pg. 189; Leonhäuser et al. 2009, pg. 20). The definition illustrates the multidimensional character of nutritional behavior. It is associated with multiple factors from the four dimensions of nutrition: health, environment, economy, and society (Schneider and Hoffmann 2011a). These factors are highly interrelated. Their interrelatedness leads to cause-effect chains, feedback loops, multicausalities, and side effects. Thus, nutritional behavior is a complex phenomenon.

This complexity of nutritional behavior is often emphasized (e.g., Beaton and Bengoa 1976; Nestle et al. 1998; Hoffmann 2003; Brug, Oenema, and Ferreira 2005; Contento et al. 2006; Renner et al. 2012). Nevertheless, factors determining nutritional behavior are named and discussed in isolation in numerous publications (e.g. Bellisle 2005; Brug 2008; Köster 2009; Etiévant et al. 2010; Kearney 2010). Research on the interrelatedness of the factors is missing (Ball, Timperio, and Crawford 2006; Köster 2009).

To increase understanding of the complex phenomenon of nutritional behavior, the aim of the present article is to demonstrate the complexity of nutritional behavior by depicting its factors in their interrelatedness in a cause-effect model. The model includes factors identified by literature research and expert consultation on an aggregated level as well as causal relationships between all factors. These were assessed by experts, which allows specification of the causal relationships: their direction (factor A influences factor B and/or vice versa), strength (weak, medium, or strong), and type (promoting or restricting).

Methods

To capture and depict factors of nutritional behavior in their interrelatedness in a cause-effect model, elements of three instruments were combined (table 1): (1) nutrition-ecological modeling NutriMod (Schneider and Hoffmann 2011b; Schneider, Hummel, and Hoffmann 2011), which was further developed to NutriMod+ST by depicting strength and type of relationships; (2) sensitivity model (SeMo) version 8.6 (Vester 2007; Malik Sensitivity Model® Prof. Vester 2013); and (3) cross-impact balance analysis (CIB) with ScenarioWizard 4.11 (Weimer-Jehle 2006, 2014).

Table 1. Instruments Applied to Capture and Depict Factors of Nutritional Behavior in Their Interrelatedness in a Cause-Effect Model.

	To capture factors of nutritional behavior	To capture direct causal relationships in a cross-impact matrix	To depict cause-effect model
Nutrition-ecological modeling + ST	X		X
Sensitivity model	X	X
Cross-impact balance analysis	X	X

The model consists of usual food consumption¹ as core factor and of factors from all dimensions of nutrition (health, environment, society, economy) that have a direct or indirect effect on food consumption. The core factor together with the other factors represents nutritional behavior. In addition, existing causal relationships between all factors were included in the model. For aspects that are not universally valid, the current situation in Germany has been determined as system boundary.

For the assessment of the causal relationships, extreme statuses of the factors were defined (e.g., good or poor state of health). If necessary, the focus of these statuses was on the dimension health even for factors from the other dimensions (e.g., the factor social identity belonging to the dimension society has the statuses favorable and unfavorable for one’s health). This means that the relationships in the model are supposed to be relevant for food consumption which is favorable for one’s health, independent of the factors’ relevance for the other dimensions.

Factors

The identification of factors involved in nutritional behavior started with a list of 41 factors mirroring current literature based on searching in topic-specific databases (e.g., Web of Science or PubMed) from October 2009 to February 2010 (Münkel, Metz, and Hummel 2010). In addition, an expert workshop with seven experts from different research areas (see table 2) was carried out in May 2011. Experts covered the different aspects of the complete scope of nutritional behavior and therefore the interdisciplinarity of the subject and came from the Max Rubner-Institut. The subject of the workshop was a brainstorming on factors of nutritional behavior, which was independent of the existing list of factors. The experts named 92 different aspects, which were used to identify, denominate, and describe aggregated factors.

Table 2. Experts Who Participated in the Workshop on Factors of Nutritional Behavior.

Dr. Erika Claupein, Department of Nutritional Behavior, Max Rubner-Institut, Federal Research Institute of Nutrition and Food, Karlsruhe, Germany

Dr. Anke Hanssen-Doose, Institute of Physical Education and Sports, University of Education Karlsruhe, Germany

Dr. Thorsten Heuer, Department of Nutritional Behavior, Max Rubner-Institut, Federal Research Institute of Nutrition and Food, Karlsruhe, Germany

Alexander Hose, Dr. von Hauner Children’s Hospital, Ludwig Maximilian University, Munich, Germany

Dr. Cornelie Pfau, Department of Nutritional Behavior, Max Rubner-Institut, Federal Research Institute of Nutrition and Food, Karlsruhe, Germany

Prof. Dr. Pirjo Schack, Faculty of Oecotrophologie (Home Economics and Nutrition Science), Münster University of Applied Sciences, Germany

Corinna Willhöft, Department of Nutritional Behavior, Max Rubner-Institut, Federal Research Institute of Nutrition and Food, Karlsruhe, Germany

For the final list of factors, aspects stated by the experts were combined with the literature-based list of factors by Münkel et al. (2010) and then aggregated. During this process, experts from the workshop were consulted individually to discuss details as needed. To complement the final list, additional literature searches were conducted. Databases of social sciences, natural sciences, and economics were used to address the multidimensionality of the issue. Factors mentioned in other published models as well as in publications on nutritional behavior were checked concerning additional aspects not yet considered in the list of factors or by the experts. The factors were specifically compiled with the aim to include all important aspects and to define them without overlapping. For unambiguousness, the final factors are described in detail (see table 4).

Table 3. Experts Who Were Interviewed Concerning Direct Causal Relationships between the Factors of the Complex Phenomenon of Nutritional Behavior.

Dr. Erika Claupein, Department of Nutritional Behavior, Max Rubner-Institut, Federal Research Institute of Nutrition and Food, Karlsruhe, Germany

Prof. Dr. Reinhard Fuchs, Institute of Sport and Sport Science, University of Freiburg, Germany

Prof. Dr. Angela Häußler, Department of Social Sciences, University of Education Heidelberg, Germany

Dr. Anke Hanssen-Doose, Institute of Physical Education and Sports, University of Education Karlsruhe, Germany

Dr. Sabrina Hense, School of Medicine and Health Sciences, Carl von Ossietzky University Oldenburg, Germany

Prof. Dr. Christoph Klotter, Department of Nutritional, Food and Consumer Sciences, Fulda University of Applied Sciences, Germany

Prof. Dr. Ingrid-Ute Leonhäuser, Institute of Nutritional Sciences, Justus-Liebig-University Giessen, Germany

Prof. Dr. Uta Meier-Gräwe, Department of Household Science, Justus-Liebig-University Giessen, Germany

Prof. Dr. Barbara Methfessel, Department of Social Sciences, University of Education Heidelberg, Germany

PD Dr. Anke Möser, ConsumerTrail, Gemünden, Germany

Dr. Gesa Schönberger, Dr. Rainer Wild-Stiftung, Foundation for Healthy Nutrition, Heidelberg, Germany

Table 4. Description and Statuses of the 19 Factors Identified for the Cause-Effect Model of Nutritional Behavior (in Alphabetical Order, Based on Expert Statements and the Cited References).

Factors	Statuses
The four main agents of socialization, following Hölscher (2008), are separate factors in the model: family, peers/peer groups, school/kindergarten, and media. The factors relate to the way agents of socialization influence peoples’ thinking and acting. As education is inducing learning with intent (Prahl and Setzwein 1999), this aspect is included in the factor. The influence of the agents of socialization leads to a process of socialization. According to Hölscher (2008), this is a lifelong, permanent, and fluent process.	Favorable for one’s health, unfavorable for one’s health
Food availability (private) refers to the availability of foods; for example, in private households (among others dependent on storage facilities and possible own gardening), at workplaces, in the bag away from home, and at kindergartens where children have free access to the present food. It is related to quantity and quality of the available food, meaning that diversity and amount are of relevance.	Good, poor
Food consumption refers to usual food consumption instead of consumption on a certain day. It is related to the kind and amount of foods being consumed and not related to where, when, and with whom food consumption takes place. The factor includes purchase decisions and therefore market demands.	Favorable for one’s health, unfavorable for one’s health
Food supply aggregates food offered in food stores and food offered for out-of-home consumption (e.g., restaurants, cafeterias). It is related to quantity and quality of the offered food, meaning that diversity and amount are of relevance. The factor also includes prices and access to food (i.e., infrastructure: choice of food stores and offers for out-of-home consumption as well as distances, accessibility (e.g., by car, bike, foot), and opening hours of food stores and places for out-of-home consumption).	Good, poor
Hunger/thirst/appetite aggregates hunger, thirst, and appetite. Instead of a current state, which changes frequently throughout the day, the factor refers to a more fundamental state.	A lot, little
Nutrition competences aggregates a person’s knowledge, experiences, as well as mental and physical ability and skills related to nutrition. Handling of knowledge and adequate implementation are also included in the factor (BMBF 1998; Eberle et al. 2005).	Existing, missing
The different phases of life are classified by chronological age as well as according to Schienkiewitz and Walter (2003) by biographical, biological, cultural, social, and subjective aspects.	Childhood, adolescence, adulthood, senior age
Physical activity refers to movements induced by skeletal muscles that increase energy expenditure (Mensink 2003) and not only to sports.	Sufficient, insufficient
Psychological resources aggregates self-efficacy and emotional control (handling of emotions, especially stress).	Existing, missing
Sex/gender aggregates the biological and the socially constructed sex/gender. Examples for aspects characterizing sex/gender of a person are stereotypes like norms of beauty and concerns about the body (Methfessel 2004), symbolism of food (e.g., meat as a male food) (Prahl and Setzwein 1999), and energy requirement.	Female, male
Sleep duration relates to the habitual time spent asleep (according to the definition by Grandner et al. 2010). Time being awake in bed is not included.	Habitually optimal, habitually deviating from optimal
Smoking relates to physiological effects of smoking.	Yes, no
Social identity refers to the current state of the process of socialization in relation to nutrition and health. Aspects characterizing the social identity of a person are values and norms (Hölscher 2008). These include restrictions of food consumption (Prahl and Setzwein 1999), ideals, rituals, symbols, traditions, habits, and customs. The factor includes attitudes towards food (e.g., interest in and appreciation of nutrition) as well as importance of eating, drinking, and food. The latter is also linked to the willingness of spending money for nutrition. Furthermore, awareness of sustainability and health is included.	Favorable for one’s health, unfavorable for one’s health
Socioeconomic status aggregates occupation (or professional reputation), education, and income of a person (Springer Gabler Verlag n.d.). The factor relates to the availability of money/resources and whether somebody is employed, unemployed, going to school, or undergoing an apprenticeship.	High, low
State of health is based on the definition of health by the WHO: “Health is a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity” (WHO 1946, pg. 2). However, this factor aggregates physical and mental well-being only.	Good, poor
Time spent for nutrition aggregates time spent for nutrition-related work of a private household and time spent for eating. Time spent for nutrition-related work includes aspects like cooking, shopping, stock-keeping, and other nutrition-related housework. Time spent by all household members (i.e., of the cook within a family and of the people cooked for) is accumulated.	Extensive and regularly, shortly and infrequently

Direct causal relationships

To compile and describe the existing direct causal relationships between the final factors, all theoretically possible direct relationships were systematically assessed by qualitative expert statements into a cross-impact matrix. In doing so, the core factor food consumption and its relationships were treated the same way as the other factors and relationships. For all direct causal relationships, the strength was assessed according to Weimer-Jehle (2006) on a scale reaching from minus three (strongly restricting) to plus three (strongly promoting) with zero for no causal relationship. For each factor, two extreme statuses were defined (exception: phases of life has four statuses). The relationships were assessed between the statuses, meaning that an influence of one factor on another was expressed by four relationships between the statuses. Relationships that are not causal influences but pure statistical associations/correlations were not included. The experts’ statements on the direct causal relationships were assessed by individual interviews with eleven experts (see table 3). For each factor, two experts were interviewed concerning the factor’s influencability (i.e., influences on the respective factor). These experts were selected because their main research areas and publications matched one or more factors of the model. The interviews were guideline based and all conducted by the same interviewer. As the factor phases of life is not modifiable, influences on this factor were not assessed. The two experts’ statements on every causal relationship were brought together with a decision scheme, taking into account the experts’ arguments. If necessary, targeted consultation was held with the experts concerning single relationships.

Cause-effect model

To depict the complex phenomenon of nutritional behavior, factors are visualized as boxes, and causal relationships between these factors and their direction are visualized as arrows. Strength (one to three) and type (promoting, restricting, or ambiguous) of the respective relationship are indicated by line style (width and dash type).

The cause-effect model does not depict the statuses of the factors. As each factor normally has two statuses, strength and type of four relationships between statuses of factors assessed by experts were combined to one overall relationship between two factors. In case of diverging assessments of strength for an overall relationship, the strongest relationship between statuses was used. If the type of relationship could not unambiguously be determined, it was marked as ambiguous in the model.

Results

The developed cause-effect model of nutritional behavior (figure 1) consists of 19 factors (for detailed descriptions, see table 4). Since the core factor is defined as usual food consumption, the model describes usual nutritional behavior. Fourteen of the factors are individual related (e.g., nutrition competences). Five factors (food supply, the four agents of socialization family, peers/peer groups, school/kindergarten, and media) represent the environment of the people whose nutritional behavior is modelled.

Figure 1. Cause-effect model of nutritional behavior. A (top): Strong and medium causal relationships. B (bottom): Weak causal relationships. For reasons of clarity, factors are sorted according to topics.

Of the 361 (19 x 19) theoretically possible direct relationships, 203 were assessed as existing. Eighteen of these are strong, and 79 are medium (figure 1, part A); 106 are weak (figure 1, part B). Concerning the types, 169 relationships were identified as promoting. Of these, 16 are strong, 66 medium, and 87 weak. Of the 15 restricting relationships, none is strong, 1 is medium, and 14 are weak. Nineteen relationships are indicated as ambiguous. Two of these are strong, 12 are medium, and 5 are weak.

Discussion

To demonstrate the complexity of nutritional behavior and to increase understanding of this complex phenomenon, knowledge on three levels is necessary: (1) knowledge on the factors decisive for nutritional behavior; (2) knowledge on the direct causal relationships between the factors, specified by their direction, strength, and type; and (3) knowledge on the interplay of these direct relationships, which leads to cause-effect chains, feedback loops, multicausalities, and side effects.

Extensive research conducted during the past decades (e.g., Popkin and Haines 1981; Köster 2009) resulted in differentiated knowledge on the factors decisive for nutritional behavior (1). For the presented model, relevant aspects were identified based on literature and expert consultations and aggregated into 19 factors. The factors reflect the multidimensionality of the issue, as factors from all dimensions of nutrition were included. The factors are described (table 4) to clearly differentiate the factors from each other and to name subfactors.

Knowledge on the direct relationships between the factors (2) is fragmentary (Ball, Timperio, and Crawford 2006; Köster 2009). Knowledge on strength and type of the relationships is especially rare but important for conducting analyses based on the model (see Outlook). Those published models of nutritional behavior that include strength and/or type of relationships (examples listed in table 5) do not focus on the overall phenomenon like the presented model. Besides these models, there is one database listing influences on nutritional behavior including the strength of relationships (table 5). As the required information on the causal relationships for the presented model could not be found in the literature, they were assessed by expert consultations. The model includes 203 direct causal relationships between the 19 factors, specified by their direction, strength, and type.

Table 5. Examples for Models (and One Database) of Overall Nutritional Behavior and of Specific Aspects of Nutritional Behavior That Include Strength and/or Type of the Relationships between Factors and/or the Interplay of the Direct Relationships (i.e., Cause-Effect Chains, Feedback Loops, Multicausalities, and Side Effects).

Source	Focus of the model	Strength of relation-ships	Type of relation-ships^1	Cause-effect chains	Feedback loops	Multicausalities	Side effects
Presented model	Overall nutritional behavior	X	X	X	X	X	X
Model from literature focusing on overall nutritional behavior
Bodenstedt et al. (1985)	Overall nutritional behavior			X	X	X	X
Database from literature focusing on overall nutritional behavior
Booth et al. (2001)	Overall nutritional behavior	X^2				X
Models from literature focusing on specific aspects of nutritional behavior
Baranowski et al. (2008)	Home fruit and vegetable availability and consumption			X	X^3	X	X
Drewnowski and Eichelsdoerfer (2009)	Consumer food choice comparing Western and Mediterranean diets	X^4		X		X
Fisher and Birch (2000)	Snack foods	X	X	X			X
Krebs-Smith and Kantor (2001)^5	Food choice			X	X^3	X	X
Rose et al. (2010)	consumer choice and food purchases			X		X
Sandvik et al. (2010)	Fruit intake	X	X	X		X	X
Schwarzer and Renner (2000)	Low-fat and high-fiber diet	X	X	X	X	X	X
Sharkey, Horel, and Dean (2010)	Food access			X	X^6	X	X
Shepherd (1985)	Food choice and intake			X		X	X
Sims, Paolucci, and Morris (1972)	Nutritional status			X		X
Wang, Worsley, and Cunningham (2008)	Consumption of healthy foods (e.g., vegetables, fruits) and unhealthy foods (e.g., sugar, fats)	X	X	X	X	X	X
Westenhöfer and Pudel (1989)	Bulimia nervosa		X^7	X	X	X	X

Note. ¹Promoting or restricting.

²Expert statements concerning the strength on a 3-point scale (low, medium, high).

³Model includes one feedback loop only.

⁴The importance of the factors is indicated by their size depicted in the model.

⁵Adapted from a model by Life Science Research Office-Federation of American Societies for Experimental Biology (1989).

⁶Model includes one feedback loop only. More feedback loops are implied as the shown feedback loop is a bidirectional relationship between two factors with many enlisted subfactors.

⁷In this model, types of relationships are roughly indicated by the specification “reinforcement” or “reduction”.

Apart from strength and type of the relationships, a particular feature of the presented model is the assessment of the interrelatedness between all factors, not only of direct influences on food consumption. This creates knowledge on the interplay of the direct relationships (3). The interplay leads to cause-effect chains, feedback loops, multicausalities, and side effects. In the presented model, cause-effect chains reveal indirect effects on food consumption. For example, media as agents of socialization have no direct influence on food consumption, but several indirect influences via cause-effect chains (e.g., media influence nutrition competences, which influence food consumption). Feedback loops, where a factor or a cause-effect chain affects its own cause in turn (Senge 2006), are important for a system’s behavior as they can cause changes in the system (Bossel 2007). Therefore, they are of relevance for the understanding of the complex phenomenon. With the presented model it is possible to identify them. For example, nutrition competences strongly influence food consumption, which has a strong effect on health status. Health status influences psychological resources with medium strength. These in turn influence nutrition competences with medium strength. An example for multicausalities is that food consumption as core factor is not only directly influenced by one factor, but by 16 factors, whereas indirect influences still have to be added. Side effects can be illustrated by the example of nutrition competences, which influence not only food consumption but also nine other factors.

Other models of nutritional behavior considering the interplay of the direct relationships are rare, and only a few include strength and type of the relationships (examples listed in table 5). Whereas cause-effect chains and multicausalities are depicted in most cause-effect models of nutritional behavior, feedback loops and side effects are rarely shown. One reason may be that many models only comprise selected relationships. As these are mostly the main effects, side effects are rarely included. In addition, these selected relationships are mostly those indirectly or directly affecting nutritional behavior, not the other way around. Excluding effects of nutritional behavior often means not depicting feedback loops.

In this context, two other models need to be mentioned as they were developed with the same methods (one with NutriMod and one with CIB) and have a comparable topic as the presented model. However, they focus on obesity with nutritional behavior as one aspect in this complex phenomenon. They consider cause-effect chains, feedback loops, multicausalities, and side effects (Schneider et al. 2009) as well as strength and type of the relationships (Weimer-Jehle, Deuschle, and Rehaag 2012).

The presented model is characterized by the broad scope of the theme: Because of the underlying definition of nutritional behavior, the model encompasses procurement, preparation, and consumption of food, actions related to storage and clearance, as well as the multidimensionality of the issue. To identify the factors in an interdisciplinary way, databases of social sciences, natural sciences, and economics were used, and experts brought in knowledge, backgrounds, and experiences from different disciplines according to their professional focus. Solely the identification of causal relationships between the factors was limited to the dimension health. Therefore, a limitation of the model is that direct and indirect influences on environmentally, socially, or economically friendly food consumption were not included.

Another characteristic of the presented model is the high level of aggregation on which the 19 factors were defined. Several subfactors were summarized in a single factor on a higher level of aggregation. These subfactors can be retrieved from the descriptions of the factors (table 4). For example, the factor food supply comprises subfactors such as offered food amount and diversity of foods, food prices, and access to food. The strength of the causal relationships was assessed for the high level of aggregation, even when the subfactors were also interrelated. Thus, statements based on the model allow increasing understanding of the overall complex phenomenon of nutritional behavior and can only be made on the chosen level of aggregation, which makes another limitation of this model. By contrast, another model on a more detailed level is required to answer more detailed questions; for example, concerning the interrelatedness of subfactors or concerning the relevance of subfactors for the entire system. To develop such a detailed model, steps to take would be similar to those undertaken for the presented model.

Many factors and relationships of the model are universally valid, which means that they are probably the same in most countries and periods (e.g., psychological resources or hunger/thirst/appetite and their relationships). Some factors however (e.g., available media or the importance of family in a society) differ between countries, resulting in specific causal relationships or specific strengths of the causal relationships. For those aspects, the presented model represents the current situation in Germany. Future research could focus on the development of models for other countries and study how they compare.

Outlook

Conducting several analyses based on the presented model of overall nutritional behavior and the underlying cross-impact matrix can further enhance this understanding and help to identify possible starting points for modifying food consumption. Instruments for such analyses can be SeMo and CIB, which were also used to develop the presented model. As examples, three potential analyses and their benefit are briefly described.

First, SeMo allows for systematic assessments regarding feedback loops. All feedback loops of a system are systematically identified and analyzed; for example, concerning their number and length. The knowledge about the type of relationships allows to differentiate between positive (reinforcing) and negative (stabilizing) feedback loops (Vester 2007). This analysis mainly enhances the understanding of the complex phenomenon of nutritional behavior and gives hints concerning possible starting points for a modification of food consumption.

Second, roles of factors in the system can be analyzed as the strength of relationships is known. The role of a factor results from the factor’s relationships with other factors. The four key roles are active, reactive, critical, and buffering. The role of each factor is somewhere between these key roles. It indicates, for example, whether a factor is suitable as a control lever for interventions in the system or as an indicator (Vester 2007). Therefore, this analysis may be conducted with the aim of identifying starting points from a systems perspective.

Third, because type and strength of relationships are known, effects of external impulses on the system can be analyzed with CIB based on the identification of consistent scenarios. A scenario is a configuration of factor statuses (e.g., good food supply, little hunger/thirst/appetite, existing nutrition competences, adulthood, insufficient physical activity, etc.). Consistent scenarios are configurations in which factor statuses interact without contradiction to the promoting and the restricting causal relationships assessed by the experts. External impulses on the system can be created by fixing some of the factor statuses. This simulates that the factor is stabilized by an external effect. The consistent scenarios with and without an impulse are compared (Weimer-Jehle 2013). This analysis also allows the identification of starting points for modifying food consumption from a systems perspective.

To our knowledge, the presented model is the first cause-effect model of overall nutritional behavior integrating knowledge on the factors decisive for nutritional behavior, on the direct causal relationships between the factors, specified by their direction, strength, and type, and on the interplay of these direct relationships. Especially inclusion of strength and type of causal relationships and consideration of the interrelatedness between all factors is unique and allows revealing cause-effect chains, feedback loops, multicausalities, and side effects. The model demonstrates the complexity of nutritional behavior and increases understanding of the complex phenomenon. In addition, it can be used as a basis for several analyses in order to further enhance this understanding and to deduce promising starting points for modifying food consumption from a systems perspective.

Acknowledgments

The authors sincerely thank all experts (see table 2 and 3) for their valuable contributions to the development of the presented model. The authors express gratitude to Martina Metz, Dr. Katja Schneider, and the master students of Justus-Liebig-University Giessen, Germany, Daniela Brandt, Anne Brants, Nadine Bräuning, Tanja Diehl, Isabel Dörnberger, Ina Goeritz, Katrin Hank, Carina Harms, Kimberley Hoffmann, Mareike Inkemann, Jennifer Juli, Carolin Kastner, Jenny Klauß, Aline Kopp, Theresa Krippl, Nora Münkel, Stephanie Ryschka, Angela Sebastiany, Jennifer Wendorf, and Manuela Wölfel for joint preparation of the list of factors mirroring current literature, that was used as a first step for the final list of factors of the presented model.

Additionally, the authors would like to thank Dr. Wolfgang Weimer-Jehle for his methodological assistance, Nicole Hillebrand and Michaela Vaas for support in depicting the model, Martina Metz for discussions concerning the factors of the model, as well as Dr. Katja Schneider and Dr. Marianne Eisinger-Watzl for reviewing an earlier version of the manuscript.

Notes

^1. In this article, terms in italics indicate factors named in the model.