Monthly Archives: May 2022

Content

Towards contextualized social simulation: Complementary use of Multi-Fuzzy Cognitive Maps and MABS

Leave a comment

By Oswaldo Terán1 and Jose Aguilar2

1Escuela de Ciencias Empresariales, Universidad Católica del Norte, Coquimbo, Chile, and CESIMO, Universidad de Los Andes, Mérid.

2CEMISID, Universidad de Los Andes, Merida, Venezuela; GIDITIC, Universidad EAFIT, Medellin, Colombia; and Universidad de Alcala, Dpto. Automatica, Alcala de Henares, Spain.

Abstract.

This work suggests to complementarily use Multi-Fuzzy Cognitive Maps (MFCM) and Multi-agent Based Simulation (MABS) for social simulation studies, to overcome deficiencies of MABS for contextually understanding social systems, including difficulties for considering the historical and political domains of the systems, variation of social constructs such as goals and interest, as well as modeler’s perspective and assumptions. MFCM are a construction much closer than MABS to natural language and narratives, used to model systems appropriately conceptualized, with support of data and/or experts in the modeled domains. Diverse domains of interest can be included in a MFCM, permitting to incorporate the history and context of the system, explicitly represent and vary agents’ social constructs, as well as take into account modeling assumptions and  perspectives.  We briefly describe possible forms of complementarily use these modeling paradigms, and exemplifies the importance of the approach by considering its relevance to investigate othering and polarization.

1. Introduction

In order to understand better issues such as othering and polarization, there is a claim in social simulation for research that includes the important domains of history, politics and game of power, as well as for greater use of social science data, make more explicit and conscious about the models’ assumptions, and be more cautious in relation to the interpretation of the simulations’ results (Edmonds et al., 2020). We describe a possible form of dealing with these difficulties: combining Multi-Agent based Simulation (MABS) and Multi Fuzzy Cognitive Maps (MFCM) (or other forms of cognitive maps), suggesting new forms of dealing with complexity of social behavior. By using MFCM an alternative modeling perspective to MABS is introduced, which facilitates expressing the context of the model, and the modelers’ assumptions, as suggested in Terán (2004).  We will consider as a case studying othering and polarization, given the difficulties for modeling it via MABS (Edmonds et al., 2020). Our proposal permits to explicitly represent social constructs such as goals, interest and influence of powerful actors on, e.g., people’s othering and polarization, and so in better contextualizing the simulated model. Variations of social constructs (e.g., goals, othering, polarization, interests) can be characterized and modeled by using MFCM.

Combined use of MABS and Fuzzy Cognitive Maps (FCM) (MFCM, multi FCM, are an extension of FCM, see the Annex) has already been suggested, see for example Giabbanelli (2017). MABS develop models at the micro level, while FCM and MFCM permits us to create models at the macro or contextual level; the idea is to use one to complement the other, i.e., to generate rich feedback between them and enhance the modeling process. Additionally, Giabbanelli propose FCM as a representation closer than MABS to natural language, allowing more participatory models, and better representation of the decision making process. Giabbanelli recommend forms of combining these two modeling approaches, highlighting key questions modelers must be careful about. In this line, we also propose a combined usage of a MFCM and MABS to overcome deficiencies of MABS modelling in Social Simulation.

Initially (in section 2) we offer a description of human societies from a broad view point,  which recognizes their deep complexities and clarifies the need for better contextualizing simulation models, allowing modeling of diverse agents’ constructs, and making explicit modelers’ assumptions and perspectives. Afterwards (in section 3), we briefly review the drawbacks of MABS for modeling some of these deep complexities. Then (in section 4), MFCM are briefly described, supported on a brief technical account in the Annex. Following (in section 5), we suggest to complementarily use MABS and MFCM for having a more comprehensive representation of human societies and their context, e.g., to better model problems such as othering and polarization. MFCM will model context and give a conceptual mark for MABS (allowing to model variation of context, e.g., changes of agents’ interests or goals, making explicit modelers’ perspective and assumptions, among other advantages), which, in turn, can be used to explore in detail specific configurations or scenarios of interest suggested by the MFCM.  Finally (in section 6), some conclusions are given.

2. (A wide view of) Human societies and influence of communication media on actual culture

As humans and primates, we recognise the social groups within which we develop as people (e.g., family, the community where we grow up, partners at the school or at work) as part of our “large home”, in which its members develop a common identity, with strong rational and emotional links. Other groups beyond these close ones are “naturally” estrangers for us and its members “instinctively” seen as others. In large civilizations such as western society, we extend somewhat these limits to include nations, in certain respects. In groups we develop perspectives, follow certain myths and rites, and have common interests, viewpoints about problems, solutions for these, and give meaning to our life. Traditionally, human societies evolve from within groups by direct face to face interaction of its members, with diverse perspectives, goals, interest, and any other social construct with respect to other groups. Nowadays this evolution mainly from natural interaction has been importantly altered in some societies, especially western and western influenced societies, where social media has introduced a new form of communication and grouping: virtual grouping. Virtual grouping consists in the creation of groups, either formally or informally, by using the internet, and social networks such as Facebook, Instagram, etc. In this process, we access certain media sites, while discarding others, in accordance with our preferences, which in turn depends on our way of thinking and preferences created in social, both virtual and direct (face to face), interaction. Currently, social media, and traditional media (TV, newspapers, etc.) have a strong influence on our culture, impacting on ours myths, rites, perspectives, forms of life, goals, interests, opinion, reasoning, emotions, and othering.

Characteristics of virtual communication, e.g., anonymity, significantly impacts on all these constructs, which, differently from natural interaction, have a strong potential to promote polarization, because of several reasons, e.g., given that virtual environments usually create less reflexive groups, and emotional communication is poorer or lack deepness. Virtual interaction is poorer than direct social interaction: the lack of physical contact strongly reduces our emotional and reflexive connection. Virtual social interaction is “colder” than direct social interaction; e.g., lack of visual contact stops communication of many emotions that are transmitted via gestures, and prevents the call for attention from the other that visual contact and gestures demands.

Even more, many times sources and veracity of information, comments, ideas, and whatever is in social media, are not clear. Even more, fake news are common in social media, what generate false beliefs, and behavior of people influenced and somewhat controlled by those who promote fake news. Fake news can in this sense generate polarisation, as some groups in the society prefer certain media, and other groups choose a different one. As these media may promote different perspectives following interest of powerful actors (e.g., political parties), conflicting perspectives are induced in the different groups, what in turn generates polarization. Social media are highly sensitive to manipulation by powerful actors worldwide, including governments (because of, e.g., their geopolitical interests and strategies), corporations (in accordance with their economic goals), religious groups, political parties, among many others. Different groups of interests influence in direct and indirect, visible and hidden, forms the media, following a wide diversity of strategies, e.g., those of business marketing, which are supported by knowledge of people (e.g., psychology, sociology, games theory, etc.). Thus, the media can create and contribute to create visions of the word, or perspectives, in accordance with the interest of powerful international or national actors. For more about all this, see, e.g, Terán and Aguilar (2018).

As a consequence, people following media that promotes a world view, related with some powerful actor(s) (e.g., a political party or a group of governments) virtually group around media that support this world view, while other people do the same in relation to other media and powerful actor(s), who promote(s) a different perspective, which many times is in conflict with the first one. Thus, grouping following the media sometimes promotes groups with conflicting perspectives, goals, interests, etc.,  which generates polarization. We can find examples of this in diverse regions and countries of the world. The media has important responsibility for polarization in a diversity of issues such as regional integration in Europe, war in Ukraine, migrations from Middle East or Africa to Europe, etc. Consequently, media manipulation sometimes allow powerful actors to influence and somewhat control perspectives and social behavior. Even more, the influence of social media on people is sometimes stronger than the influence of direct social interaction. All these introduce deep complex issues in social human interaction and behavior. This is why we have chosen polarization as the case study for his essay.

Consequently, to comprehend actual human behavior, and in particular polarization, it is necessary to appropriately take into account the social context, what permits to understand better the actual complexity of social interaction, e.g., how powerful international, national, and local actors’ influence on media affects people perspectives, goals, interest, and polarization, as well as their strategies and actions in doing so. Contextualized modeling will help in determining social constructs (goals, interests, etc.) in certain situations, and their variation from situation to situation. For this, we suggest complementing MABS with MFCM. For more about the consequences of virtual interaction, see for example Prensky (2001a, 2001b). Prensky (2009) has also suggested forms to overcome such consequences: to promote digital wisdom. MABS and MFCM models will help in defining forms of dealing with the problems of high exposure to social networks, in line with Prensky’s concerns.

3. Weakness of the MABS approach for modeling context

Edmonds et al. (2020) recognize that MABS models assume a “state of the world” or “state of nature” that does not include the historical context of the agent, e.g., in such a way that they explicitly present goals, interests, etc., and pursue them via political actions, sometimes exerting power over others. For instance, the agents can not change their goals, interests or desires during the simulation, to show certain evolution, as a consequence of reflection and experience at the level of desires, allowing cognitive variations. The models are strongly limited in relation to representing the context of the social interaction, which in part determines variation of important factors of agents’ behavior, e.g., goals. This, to a good extent, is due to lack of representation of the agents’ context. For the same reason, it is difficult to represent modelers assumptions and perspectives, which might also be influenced by social media and powerful actors, as explained above.

The Special Issue of the Social Science Computer Review. (Volume 38, Issue 4, August 2020, see for instance Edmonds et al. (2020) and Rocco and Wander (2020)), presents several models aiming at dealing with some of these drawbacks of MABS, specifically, to relate models to social science data, be more aware about the models’ assumptions, and be more cautious in relation to the interpretation of the simulations’ results. However, in these works diverse difficulties are not addressed, e.g.,  having appropriate representation of the context in order to explicitly consider diverse constructs, e.g., goals and interests, as well as having a wide representation of modelers perspectives and assumptions so that diverse perspectives can be addressed and compared, among other important matters.

MABS represent social interaction, i.e., the interaction in a group, where the agent’s goal, and other social constructs are assumed given, not variable, and to understand the context where they appear is not of interest or is out of reach (too difficult). However, as explained above, agents are in diverse social groups, not only in the simulated one, and so goals, interests, and beliefs in the modeled group are shaped in accordance to their interactions in diverse groups, and the influence of multiple, virtual and natural groups in which they participate. In order to represent variations of such elements, the context must be taken into account, as well as to elaborate models from narratives. MFCM is naturally close to narratives, as it is elaborated from conceptual frameworks. In this sense, MFCM might represent an intermediate step towards MABS models. In a MFCM and in the steps towards elaborating the MABS, modelers’ perspectives and assumptions can be made explicit. In addition, MABS presents limitations to determine the conditions for which a certain behavior or tendency occurs (Terán, 2001; Terán et al. 2001), i.e., for making strong inferences and theorem proving of tendencies for subsets of the theory of the simulation, which could more easily be performed in the MFCM. Hopefully, exploring configurations of the MFCM the proof could be carried out indirectly, in a higher level than in MABS, as has already been suggested in previous papers (Aguilar et al., 2020; Perozo et al., 2013).

4. Multi-Fuzzy Cognitive Maps (MFCM)

We suggest conceptual or cognitive maps as a more flexible form than MABS to represent context of a social situation, and in particular MFCM, as implemented by Aguilar and others (see e.g., Kosko, 1986; Aguilar 2005, 2013, 2016; Aguilar et al., 2016, 2020; Contreras and Aguilar, 2010; and Sánchez et al., 2019; Puerto et al., 2019). A brief description of Fuzzy and Multi-fuzzy cognitive maps, following Sánchez et al. (2019), is given in the Annex.

Fuzzy cognitive maps help us in describing the context via qualitative (e.g., very low, low, medium, high, too high) and quantitative variables, as indicated in the annex. The system is represented by the network of concepts (variables) interrelated via weights (also given by variables). The high level of the MFCM paradigm, differently from a MABS, permits us to explicit different elements of the models such as the agents constructs (goals, interests, etc.), as well as the modelers assumptions and perspectives, as suggested in Terán (2004). MFCM will facilitate to explicit the accumulated set of assumptions (“abstraction-assumptions, design-assumptions, inference-assumptions, analysis-assumptions, interpretation-assumptions and application-assumptions”, as these are summarized in Terán, 2004).

In a MFCM, a particular situation of the system is given by a specific configuration of the weights (see, e.g., Sánchez et al., 2019). Suppose we are dealing with a model similar to that elaborated in Sánchez et al. (2019) to study the quality of opinion in a community. Sánchez et al. examine the capabilities of the MFCM for knowledge description and extraction about opinions presented in a certain topic, allowing the assessment of the quality of public opinion. Special attention is offered to the influence of the media on public opinion. The evolution of the concepts and relationships is presented. Concepts define the relevant aspects from which public opinion emerges, covering diverse domains, for instance, the social, technological and psycho-biological ones. The MFCM permits to identify the media preferred by the public in order to better understand several issues, including the high esteem that the new communication media hold.

In line with this, let us assume that we want to understand the quality of public opinion in a community of Europe about diverse issues during 2022. This network, with a certain configuration of the weights, but unspecified concepts, represents a social system with a certain structure (as the weights are given) that is in some sense general, as the values of the concepts can still vary. Variation of the concepts represents different scenarios of the social system (with a given structure, defined by the weights); e.g., the model of a European community considered in relation to three scenarios regarding the state of public opinion in relation to specific issues: 1: climate change, 2: situation of tourism in the community, 3: secondary effects of the COVID-19 vaccines. The weights of the network are determined by using a variety of scenarios; i.e., the network is trained with several scenarios, for which all possible values of the concepts are known. Once the network is trained, it can be used to infer unknown specific values of the concepts for other scenarios (following Aguilar et al. 2020; Sánchez et al. 2019; Terán y Aguilar, 2018); e.g., the state of public opinion in relation to the involvement of EU in the war in Ukraine. Even more, by exploring an appropriate set of scenarios, proofs about the state of certain concepts can be developed; e.g., that a majority of people in the community is against direct EU involvement in the war in Ukraine. The proof could be carried out for a subset of the possible configurations of a domain, several domains, or part of a domain, e.g., for the psycho-biological domain. Additionally, having an appropriate elaboration of the model would allow evaluating how polarized is the opinion of the community in relation to the involvement of EU in that war.

Diverse configurations of the MFCM can represent different modelers’ perspectives and assumptions, as well as various agents’ constructs, such as goals, interests, etc., allowing to deal with the above described drawbacks of MABS to cope with complexity of social systems.

5. Combined use of MABS and cognitive maps

The combined usage will give at least two levels of modeling: the inner, defined by the agents’ interaction concreted in a MABS, and the outer or contextual one, given by the MFCM. These will be the two last levels in the description given in 5.1. Interaction between these models occurs as the modeler interprets each model outputs and feedbacks the other. Ideally, we would have direct automatic feedback between these models.

5.1 Levels of description of the System

In order to contextually model a social system and investigate problems such as polarization, we suggest below five levels of description of the system. The first three levels are not directly associated to computational models, while levels four and five are descriptions that assist development of the computational models: MFCM and MABS, respectively. Each level gives context to the following one (the first gives context to the second, etc.). A lower level (e.g., 1. in relation to 2.) of description corresponds to a more general language, as suggested in Terán (2004). Each level must take into account the previous levels, especially the immediately superior level, which gives the most immediate context to it. This description is in line with the suggestions given in Terán (Idem). Each description makes certain assumptions and is shaped by the modeler’s perspective, which in part is coming from those actors given information to build the model. Assumptions and perspectives introduced in level of modeling i, i = 1, …, 6, can be called Assumptions-given in (i) and Perspectives-given in (i). At levels of description j, Assumption i = 1, 2, …, i are accumulated, and can be called Assumptions(j), as well as holistic Perspective(j) based on Perspectives-given in (i), i = 1, 2, .., i.  These assumptions and perspectives correspond to those defined in Terán (2004) as “abstraction-assumptions, design-assumptions, inference-assumptions, analysis-assumptions, interpretation-assumptions and application-assumptions”.

  1. Describe in natural language the system, including its relevant history, with emphasis in culture (practices, costumes, etc.) and behavior of individuals and groups relevant to the object of the study, e.g., from a historical-ontological perspective. Here, how the system has reached its actual situation is explained. This will give a global view of the society and the general form of behavior, problematics, conflicts, etc.
  2. Describe the diverse relevant domains given context to the system of interest in accordance with the study, e.g., political, economical, dominant actors, etc., and the relationships among them. Concrete specifications of these domains sets scenarios for the real system, i.e., possible configurations of it.
  3. Describe the particular social group of interest as part of the society explained above, in 1., and the domains given in 2., showing its particularities, e.g., culturally, in terms of interests, situation and social interaction of this group in relation with other groups in the whole society, in accordance to the problematic addressed in the study.
  4. Elaborate a cognitive map of the situation of the social group of interest, following the description given in 2 and 3. This is a description to be represented in a computational language, such as the MFCM tool developed by Contreras and Aguilar (2010).
  5. Describe the MABS model. The MABS model is then represented in a simulation language.
  6. The computational MFCM (or other cognitive map) and the MABS developed following 4. and 5. are then used to generate the virtual outputs and simulation study.

5.2 Possible combined uses of MFCM and MABS.

The MFCM (in general a cognitive or conceptual map) gives context to the MABS model (modeler’s assumptions and perspectives are added in the process, as indicated above), while the MABS model represents in detail the interaction of the agents’ considered in the MFCM for a specific scenario of this, as indicated in the levels of modeling given above. With this idea in mind, among the specific forms a combined usage of ABM and MFCM, we have:

i) Offering feedback from the MFCM to the MABS. A MABS in a certain configuration can be used to generate either directly or indirectly (e.g., with additional verification or manipulations) input for a simulation model. For example, in parallel to the model presented in Sánchez et al. (2019), where the domains social, biological-psychological, technological and state of the opinion are displayed, a MABS model can be developed to represent the interaction between social entities, such as people who receive information from the media, the media itself, and powerful actors who design the agenda setting of the media. This MABS model might use diverse methodologies, e.g., endorsements or BDI, to represent social interaction, or a higher level of interaction where actors share resources, on which they have interest, as in a System of Organized Action (see, e.g., SocLab models Terán and Sibertin-Blanc, 2020; Sibertin-Blanc et al., 2013). Constructs required as inputs for the model, e.g., goals, interests, values to define the endorsements schema, etc. could be deducted from the MFCM, as direct values of some concepts or functions (mathematical, logical, etc.) of the concepts. These operations could be defined by experts in the modeled domains (e.g., media owners, academics working in the area, etc.). Ideally, we would have an isomorphic relation between some variables of the MABS and variables of the MFCM – however, this is not the usual case –. In this process, the MABS is contextualized by the MFCM, whose modeling level also permits to identify modeler’s assumptions and perspectives. Also, in a narrative, and then in the MFCM, goal, interest, and other constructs can be explicitly represented, then varied and their consequence understood in order to feedback the MABS model.

ii) Giving feedback to the MFCM from the MABS. Inputs and outputs values of the MABS simulation can be used as an input to the MFCM, e.g., as a set of scenarios to train the network and determine a certain structure of the MFCM, or to determine a specific scenario/ configuration (where both, the weights and the concepts are known).

iii) Determining conditions of correspondence among the models. By simulating the MABS associated to certain scenarios of the MFCM, or, vice-verse, by determining the scenarios of MFCM related to certain MABS, the consistency among the two models and possible errors, omissions, etc. in one of the models can be detected, and then the corrections applied. Even more, this exercise can hopefully determine certain rules or conditions of correspondence among the MABS and the MFCM.

iv) Using a model to verify properties in the other model. Once certain correspondence among the models has been determined, we can use one of the models to help in determining properties of the other. For instance, a proof of a tendency in an MABS (this has been an important area of research, Terán, 2001; Terán et al. 2001; Edmonds et al., 2006) could be developed in a much easier form in the corresponding MFCM. For this, we need to characterize the set of configurations of the MFCM corresponding to the set of configurations of the MABS for which we want to perform the proof.

These possible combined uses of MFCM and MABS do not exhaust all potentials, and diverse other alternatives could appear in accordance with the needs in a particular study. Even more, automatic feedback between MFCM (or other cognitive or conceptual map) and MABS could be implemented in the future, to facilitate the mutual contributions between the two modeling approaches. This would cover modeling requirements the MABS in itself does not support at present.

5.3 A case: Modeling othering and polarization, the case of “children with virtually mediated culture”.

We outline a possible model that considers othering and polarization. In section 2 we described a society. In a society, as virtual groups become homogeneous in beliefs, motivations, intentions and behavior, certain sort of endogamy of ideas and opinions appear, constraining the variety and richness of perspectives from which people observe and judge others, making them generally less tolerant to others, more restrictive in accepting opinions and behavior of others, so less inclusive. This has diverse additional effects, for instance, increase of polarization between virtual groups regarding a diversity of themes. Problems such as polarization occur also in children with strong usage of virtual social networks (see, Prensky, 2001a, 2001b, 2009).

To investigate this problem and support the MABS, as a case, we suggest a MFCM with four levels (see Figure 1). The goal of the models (MFCM and MABS) would be to better understand the differences between the communities of children whose interaction is basically virtually mediated and the community of children whose interaction is face to face, or direct, people to people. In general, it is of interest to determine the state of othering and polarization for diverse configurations of the MFCM. As we explained above, there are clear differences between virtual and face to face interaction, consequently the upper layers in Figure 1 are the two possible niches of cultural acquisition (costumes, points of view, etc.), during life of people (ontogenesis), namely, the virtual mediated culture and the direct, face to face, cultural acquisition. These two layers involve interaction among diverse actors (e.g., people, media and powerful actors are present in layer 2). Layer 2 represents technological actors, while layer 1 represents social interaction, but both of them might involve other elements, if required. The third level represent those biological aspects related with behavior, which are created via culture: the psycho-biological level. Both levels, 1 and 2, affect the third layer, as emotions, reasoning, etc., are founded on people interaction and have a cultural base. Constructs of behavior such as goals, interests, desires, polarization, etc., appear and can be explicitly represented at this level. This third level, in turn, impacts on the overall state of the community, e.g., on the auto-generative capacity of the society, finally affecting global society/community’s othering and polarization, as our emotions, reasoning, etc., impact on our view point, on othering, etc. These last are variables defined in terms of the previous levels. In this model, the definition of concepts such as “othering” and “polarization” is crucial, and indicates basic modeling assumptions and perspective. Finally, the overall state of society/community impacts back on the cultural niches (layers 1 and 2).

In a specific situation, the whole interaction (1) of the society or community is divided between the two niches given by layers 1 and 2, a proportion of interaction frequency occurs as virtual communication, and the compliment (one minus the proportion of virtual interaction), occurs as direct, face to face, contact. This is represented in Figure 3 by the variable “Proportion of virtual interaction type”. Changes of this variable allows us to explore diverse configurations or scenarios of interaction, ranging from total virtual interaction (null direct contact) (the variable takes the value 1), to null virtual interaction (total direct contact) (the variable takes the value 0).

The four levels of MFCM

Figure 1. The four levels of the MFCM. The two alternative niches structuring the psycho-biology of people are at the top of the process. The overall state represents general measures such as the auto-generative character of a social system, and attitudes including othering and polarization.

Example of possible variables in some levels (Figure 1) are:

i) Face to face interaction, and ii) virtual interaction or technological: The next variables are candidates to be at these levels. Degree of:

  •  Coherence of the interaction (possible state: good, etc.);
  •  Identification of the others in the interaction (good or clear, etc.);
  •  Richness of the interaction (high or good, etc.);
  •  Truthfulness of the messages (fairness) (e.g., good: messages and communication are fair);
  •  Openness of the community (e.g., high: usually people is open to interact with others);
  •  Speed of the interaction (e.g.: low, medium, …);
  •  Intentioned influence and control of the communication by powerful actors (e.g., high, medium, low, ..);

iii) Psycho-biological level: 

  •  Reflection (state: good means that people question their experiences, and observed phenomena);
  • Closeness of interpretations, attitudes, desires, intentions, and plans (a high value means that people’s interpretations, etc., are not very different);
  •  Emotion and mood;
  •  Empathy;
  •  Addiction to virtual interaction;
  • Goal;
  •  Interest;
  •  Immediatism (propensity to do things quickly and constantly change focus of reasoning).

iv) Overall state of people and society:

  •  Auto-generative capacity of the society;
  •  Capacity of people to reflect about social situations (and autonomously look for solutions);
  •  Othering;
  •  Polarization; 

Concepts at each layer impacts concepts at the other layers. E.g., concepts of level three have a strong impact on concepts of the fourth layer, such as “polarization”, and “auto-generative character of the society”.

As indicated above, to understand the dynamics of the MFCM we can develop a wide range of scenarios, for instance, varying the switch “Proportion of virtual interaction” in the interval [0, 1], to explore a set of scenarios for which the degree of virtual interaction increases from 1 to 0, as the proportion of direct interaction decreases from 1 to 0 (the real case corresponds to an intermediate value between 1 and 0). These experiments will help us in understanding better the consequences of virtual mediated culture. Even when the outline of the model presented here might need some adjustments and improvements, the present proposal keeps its potential to reach this goal.

The MFCM will be useful to deal with many issues and questions of interest, for instance:

  • How social networks affect basic social attitudes such as: i) critical rationality (people’s habit for questioning and explaining their experience (issues/phenomena in their life)), ii) tolerance, iii) compromise with public well-being, and iv) othering and polarization
  •  How social networks affect social feelings, such as empathy.

The MABS model will be elaborated in accordance with the description of the MFCM indicated above. In particular, different values of the social constructs at levels 1 and 2 (e.g., goals and interests of the actors), and the corresponding state of layer 3 (e.g., Polarization), imply diverse MABS models.

The whole network of concepts, the particular network of concepts at each level, and the definition of each concept, offers a perspective of the modelers. Different modelers can develop these elements of the model differently. Assumptions can be identified also at each level. Both, perspectives and assumptions come from the modelers as well as from the theories, consult to experts to create the model, etc.. An specific model is not part of this essay, but rather a subject of future work.

6. Conclusion

Social simulation has been widely recognized as an alternative to study social systems, using diverse modeling tools, including MABS, which, however, present some limitations, like any other research tool. One of the deficiencies of MABS is their limitations to contextually modeling social systems; e.g., to suitably include the historical and political contexts or domains; difficulties to represent variation of agents’ constructs, e.g., goals and interest; and drawbacks to made explicit modeler’s assumptions and perspectives. In this paper, we have suggested to mutually complement MABS and MFCM, to overcome MABS drawbacks, to potentiate the usefulness of MABS to represent social systems.

We argue that the high level of the MFCM paradigm permits us to express different elements of the models such as the agents’ constructs (goals, interests, etc.), as well as the modelers assumptions and perspectives, as suggested in Terán (2004). Thus, MFCM facilitates the identification of the accumulated set of assumptions during the modeling process. Even more, diverse configurations of a MFCM can represent diverse modelers’ perspectives and assumptions, as well as diverse agents constructs, such as goals, interests, etc., allowing to deal with the above described complexity. This permits us to more realistically elaborate models of a wide diversity of social problems, e.g., polarization, and consequences of the influence of social networks in culture.

Among the forms MFCM and MABS complement each other we have identified the followings:  mutual feedbacking of variables and concepts between the MFCM and the MABS, determining conditions of correspondence among the models, what facilitate other modeling needs, e.g., using a model to verify properties in the other model (e.g., proofs required in a MABS could be carried out in a corresponding MFCM).

A case study was outlined to exemplify the problematic that can be addressed and the advantages of using MFCM to complement MABS: Modeling othering and polarization, the case of children with virtually mediated culture. Combined use of MFCM and MABS in this case will contribute to understand better the problems created by the high use of digital interaction, especially social networks, as described by Prensky (2001a, 2001b, 2009), given that virtual interaction has strong influence on our culture, impacting on ours myths, rites, perspectives, goals, interests, opinion, othering and polarization, etc. Characteristics of virtual communication, e.g., anonymity, significantly impacts on all these constructs, which, differently from natural interaction, have a strong potential to promote certain tendencies of such constructs, e.g., polarization or our opinions, because of several reasons, for instance, given that virtual environments usually create less reflexive groups, while emotional communication is poorer or lack deepness. It is difficult to represent all these dynamics in a MABS, but it can be alternatively expressed in a MFCM.

The purpose of the work was to give an outline of the proposal; future work will be conducted to wholly develop concrete study cases with complementary MABS and MFCM models.

References

Aguilar, J. (2005). A Survey about Fuzzy Cognitive Maps Papers. International journal of computational cognition 3 (2), pp. 27-33.

Aguilar, J. (2013) Different Dynamic Causal Relationship Approaches for Cognitive Maps”, Applied Soft Computing, Elsevier, 13(1), pp. 271–282.

Aguilar, J. (2016) “Multilayer Cognitive Maps in the Resolution of Problems using the FCM Designer Tool”, Applied Artificial Intelligence, 30 (7), pp. 720-743.

Aguilar J., Hidalgo J., Osuna F., Perez N.(2016) Multilayer Cognitive Maps to Model Problems”, Proc. IEEE World Congress on Computational Intelligence,  pp. 1547-1553, 2016.

Aguilar Jose, Yolmer Romero y OswaldoTerán (2020). “Analysis of the effect on the marketing of the sport product from the “par Conditio” principle in Latin American football and baseball competitions”. Submitted to International Journal of Knowledge-Based and Intelligent Engineering Systems.

Contreras, J. Aguilar, J. (2010). “The FCM Designer Tool”. Fuzzy Cognitive Maps: Advances in Theory, Methodologies, Tools and Application (Ed. M. Glykas), Springer, pp. 71-88, 2010.

Edmonds, B. and Hales, D. and Lessard-Phillips, L.(2020). Simulation Models of Ethnocentrism and Diversity: An Introduction to the Special Issue. Social Science Computer Review. Volume 38 Issue 4, August 2020, 359–364, https://journals.sagepub.com/toc/ssce/38/4

Edmonds Bruce, Oswaldo Terán y Grary Polhill (2006). “To the Outer Limits and Beyond –characterising the envelope of social simulation trajectories”, Proceedings of theThe First World Congress on Social Simulation WCSS”, Kyoto, 21-25 Agosto, 2006.

Giabbanelli Philippe, Steven Gray and Payam Aminpour (2017), Combining fuzzy cognitive maps with agent-based modeling: Frameworks and pitfalls of a powerful hybrid modeling approach to understand human-environment interactions, Environmental Modeling and Software, September 2017,  95:320-325 DOI:10.1016/j.envsoft.2017.06.040

Kosko, B. (1986). Fuzzy Cognitive Maps. International Journal of Man-Machine Studies 24, pp. 65-75.

Perozo Niriaska, Jose Aguilar, Oswaldo Terán y Heidi Molina (2013). A Verification Method for MASOES, IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS—PART B: CYBERNETICS, Vol 43, num 1, February, pp. 64-76,  ISBN: 1083-4419. DOI: 10.1109/TSMCB.2012.2199106

Prensky, Marc (2001a). Digital Natives, Digital Immigrants Part 1, On the Horizon, Vol. 9, No. 5, September, pp 1-6. DOI: 10.1108/10748120110424816

Prensky, Marc (2001b). Do They Really Think Differently? Digital Natives, Digital Immigrants Part 2, On the Horizon, Vol. 9, No. 6, October 2001, pp 1-6. DOI: 10.1108/10748120110424843

Prensky, Marc (2009). “H. Sapiens Digital: From Digital Immigrants and Digital Natives to Digital Wisdom,” Innovate: Journal of Online Education: Vol. 5 : Issue 3, Article 1. Available at: https://nsuworks.nova.edu/innovate/vol5/iss3/1

Puerto E., Aguilar J., Chávez D., López C. (2019) Using Multilayer Fuzzy Cognitive Maps to Diagnose Autism Spectrum Disorder, Applied Soft Computing Journal, 75, pp. 58–71.

Rocco Paolillo, and Wander Jager (2020). Simulating Acculturation Dynamics Between Migrants and Locals in Relation to Network Formation, Social Science Computer Review. Volume 38 Issue 4, August 2020, pp. 365–386. https://journals.sagepub.com/toc/ssce/38/4

Sánchez Hebert, Jose Aguilar, Oswaldo Terán, José Gutiérrez de Mesa (2019). “Modeling the process of shaping the public opinion through Multilevel Fuzzy Cognitive Maps”, Applied Soft Computing, Volume 85. https://doi.org/10.1016/j.asoc.2019.105756.

Sibertin-Blanc, C., Roggero, P., Adreit, F., Baldet, B., Chapron, P., El-Gemayel, J., Mailliard, M., and Sandri, S. (2013). “SocLab: A Framework for the Modeling, Simulation and Analysis of Power in Social Organizations”, Journal of Artificial Societies and Social Simulation (JASSS), 16(4). http://jasss.soc.surrey.ac.uk/

Terán Oswaldo (2001). Emergent Tendencies in Multi-Agent Based Simulations Using Constraint-Based Methods to Effect Practical Proofs Over Finite Subsets of Simulation Outcomes, Doctoral Thesis, Centre for Policy Modelling, Manchester Metropolitan University, 2001.

Terán Oswaldo  (2004). Understanding MABS and Social Simulation: Switching Between Languages in a Hierarchy of Levels, Journal of Artificial Societies and Social Simulation vol. 7, no. 4. http://jasss.soc.surrey.ac.uk/7/4/5.html

Terán Oswaldo, Edmonds Bruce and Steve Wallis (2001) “Mapping the Envelope of Social Simulation Trajectories”, Lecture Notes in Artificial Intelligence (Subserie de Lecture Notes in Computer Science): MABS, Volume 1979, pp. 229-243, Springer, Alemania, 2001

Terán Oswaldo y Jose Aguilar (2018). “Modelo del proceso de influencia de los medios de comunicación social en la opinión pública”, EDUCERE, 21 (71), Universidad de Los Andes, Mérida, Venezuela. https://www.redalyc.org/toc.oa?id=356&numero=56002

Terán Oswaldo and Christophe Sibertin-Blanc (2020). Impact on Cooperation of altruism, tenacity, and the need of a resource,  accepted in IEEE Transactions on Computational Social Systems. DOI: 10.1109/TCSS.2021.3136821.

Annex. Fuzzy Cognitive Maps (FCM) and Muti-Fuzzy Cognitive Maps (MFCM).

Cognitive map theory is based on symbolic representation for the description of a system. It uses information, knowledge and experience, to describe particular domains using concepts (variables, states, inputs, outputs), and the relationships between them (Aguilar 2005, 2013, 2016). Cognitive maps can be understood as directed graphs, whose arcs represent causal connections between the nodes (concepts), used to denote knowledge. An arc with a positive sign (alternatively, negative sign), going from node X to node Y means that X (causally) increases (alternatively, decreases) Y. Cognitive maps are graphically represented: concepts are connected by arcs through a connection matrix. In the connection matrix, the i-nth line represents the weight of the arc connections directed outside of the  concept. The i-nth column lists the arcs directed toward , i.e., those affecting .

The conceptual development of FCMs rests on the definition and dynamic of concepts and relationships created by the theory of fuzzy sets (Kosko, 1986). FCM can describe any system using a causality-based model (that indicates positive or negative relationships), which takes fuzzy values and is dynamic (i.e., the effect of a change in one concept/node affects other nodes, which then affect further nodes). This structure establishes the forward and backward propagation of causality (Aguilar, 2005, 2013, 2016). Thus, the concepts and relations can be represented as fuzzy variables (expressed in linguistic terms), such as “Almost Always”, “Always”, “Normally”, “Some (see Figure 2).

The value of a concept depends on its previous iterations, following the equation (1):

Screenshot 2022-05-23 at 15.10.24

Cm(i+1) stands for the value of the concept in the next iteration after the iteration i, N indicates the number of concepts, wm,k represents the value of the causal relationship between the concept Ck and the concept Cm, and S(y) is a function used to normalize the value of the concept.

ot-fig2

Figure 2. Example of an FCM (taken from Sánchez et al., 2019).

MFCM is an extension of the FCM. It is a FCM with several layers where each layer represents a set of concepts that define a specific domain of a system. To construct a MFCM, the previous equation for calculating the current status of the concepts of a FCM is modified, to describe the relationships between different layers (Aguilar, 2016):

Where F(p) is the input function generated by the relationships among different layers, and p is the set of concepts of the other layers that impact this concept. Thus, the update function of the concepts has two parts. The first part, the classic, calculates the value of  concept in iteration  based on the values of concepts in the previous iteration . All these concepts belong to the same layer where the “m” concept belongs. The second part is the result of the causal relationship between the concepts in different levels of the MFCM.

Terán, O. & Aguilar, J. (2022) Towards contextualized social simulation: Complementary use of Multi-Fuzzy Cognitive Maps and MABS. Review of Artificial Societies and Social Simulation, 25th May 2022. https://rofasss.org/2022/05/25/MFCM-MABS

© The authors under the Creative Commons’ Attribution-NoDerivs (CC BY-ND) Licence (v4.0)

Content

Discussions on Qualitative & Quantitative Data in the Context of Agent-Based Social Simulation

Leave a comment

By Peer-Olaf Siebers, in collaboration with Kwabena Amponsah, James Hey, Edmund Chattoe-Brown and Melania Borit

Motivation

1.1: Some time ago, I had several discussions with my PhD students Kwabena Amponsah and James Hey (we are all computer scientists, with a research interest in multi-agent systems) on the topic of qualitative vs. quantitative data in the context of Agent-Based Social Simulation (ABSS). Our original goal was to better understand the role of qualitative vs. quantitative data in the life cycle of an ABSS study. But as you will see later, we conquered more ground during our discussions.

1.2: The trigger for these discussions came from numerous earlier discussions within the RAT task force (Sebastian Achter, Melania Borit, Edmund Chattoe-Brown, and Peer-Olaf Siebers) on the topic, while we were developing the Rigour and Transparency – Reporting Standard (RAT-RS). The RAT-RS is a tool to improve the documentation of data use in Agent-Based Modelling (Achter et al 2022). During our RAT-RS discussions we made the observation that when using the terms “qualitative data” and “quantitative data” in different phases of the ABM simulation study life cycle these could be interpreted in different ways, and we felt difficult to clearly state the definition/role of these different types of data in the different contexts that the individual phases within the life cycle represent. This was aggravated by the competing understandings of the terminology within different domains (from social and natural sciences) that form the field of social simulation.

1.3: As the ABSS community is a multi-disciplinary one, often doing interdisciplinary research, we thought that we should share the outcome of our discussions with the community. To demonstrate the different views that exist within the topic area, we ask some of our friends from the social simulation community to comment on our philosophical discussions. And we were lucky enough to get our RAT-RS colleagues Edmund Chattoe-Brown and Melania Borit on board who provided critical feedback and their own view of things. In the following we provide a summary of the overall discussion Each of the following paragraph contains summaries of the initial discussion outcomes, representing the computer scientists’ views, followed by some thoughts provided by our two friends from the social simulation community (Borit’s in {} brackets and italic and Chattoe-Brown’s in [] brackets and bold), both commenting on the initial discussion outcomes of the computer scientists. To see the diverse backgrounds of all the contributors and perhaps to better understand their way of thinking and their arguments, I have added some short biographies of all contributors at the end of this Research Note. To support further (public) discussions I have numbered the individual paragraphs to make it easier to refer back to them. 

Terminology

2.1: As a starting point for our discussions I searched the internet for some terminology related to the topic of “data”. Following is a list of initial definitions of relevant terms [1]. First, the terms qualitative data and quantitative data, as defined by the Australian Bureau of Statistics: “Qualitative data are measures of ‘types’ and may be represented by a name, symbol, or a number code. They are data about categorical variables (e.g. what type). Quantitative data are measures of values or counts and are expressed as numbers. They are data about numeric variables (e.g. how many; how much; or how often).” (Australian Bureau of Statistics 2022) [Maybe don’t let a statistics unit define qualitative research? This has a topic that is very alien to us but argues “properly” about the role of different methods (Helitzer-Allen and Kendall 1992). “Proper” qualitative researchers would fiercely dispute this. It is “quantitative imperialism”.].

2.2: What might also help for this discussion is to better understand the terms qualitative data analysis and quantitative data analysis. Qualitative data analysis refers to “the processes and procedures that are used to analyse the data and provide some level of explanation, understanding, or interpretation” (Skinner et al 2021). [This is a much less contentious claim for qualitative data – and makes the discussion of the Australian Bureau of Statistics look like a distraction but a really “low grade” source in peer review terms. A very good one is Strauss (1987).] These methods include content analysis, narrative analysis, discourse analysis, framework analysis, and grounded theory and the goal is to identify common patterns. {These data analysis methods connect to different types of qualitative research: phenomenology, ethnography, narrative inquiry, case study research, or grounded theory. The goal of such research is not always to identify patterns – see (Miles and Huberman 1994): e.g., making metaphors, seeing plausibility, making contrasts/comparisons.} [In my opinion some of these alleged methods are just empire building or hot air. Do you actually need them for your argument?] These types of analysis must therefore use qualitative inputs, broadening the definition to include raw text, discourse and conceptual frameworks.

2.3 When it comes to quantitative data analysis “you are expected to turn raw numbers into meaningful data through the application of rational and critical thinking. Quantitative data analysis may include the calculation of frequencies of variables and differences between variables.” (Business Research Methodology 2022) {One does the same in qualitative data analysis – turns raw words (or pictures etc.) into meaningful data through the application of interpretation based on rational and critical thinking. In quantitative data analysis you usually apply mathematical/statistical models to analyse the data.}. While the output of quantitative data analysis can be used directly as input to a simulation model, the output of qualitative data analysis still needs to be translated into behavioural rules to be useful (either manually or through machine learning algorithms). {What is meant by “translated” in this specific context? Do we need this kind of translation only for qualitative data or also for quantitative data? Is there a difference between translation methods of qualitative and quantitative data?} [That seems pretty contentious too. It is what is often done, true, but I don’t think it is a logical requirement. I guess you could train a neural net using “cases” or design some other simple “cognitive architecture” from the data. Would this (Becker 1953), for example, best be modelled as “rules” or as some kind of adaptive process? But of course you have to be careful that “rule” is not defined so broadly that everything is one or it is true by definition. I wonder what the “rules” are in this: Chattoe-Brown (2009).]

2.4: Finally, let’s have a quick look at the difference between “data” and “evidence”. For this, we found the following distinction by Wilkinson (2022) helpful: “… whilst data can exist on its own, even though it is essentially meaningless without context, evidence, on the other hand, has to be evidence of or for something. Evidence only exists when there is an opinion, a viewpoint or an argument.”

Hypothesis

3.1: The RAT-RS divides the simulation life cycle into five phases, in terms of data use: model aim and context, conceptualisation, operationalisation, experimentation, and evaluation (Siebers et al 2019). We started our discussion by considering the following hypothesis: The outcome of qualitative data analysis is only useful for the purpose of conceptualisation and as a basis for producing quantitative data. It does not have any other roles within the ABM simulation study life cycle. {Maybe this hypothesis in itself has to be discussed. Is it so that you use only numbers in the operationalisation phase? One can write NetLogo code directly from qualitative data, without numbers.} [Is this inevitable given the way ABM works? Agents have agency and therefore can decide to do things (and we can only access this by talking to them, probably “open ended”). A statistical pattern – time series or correlation – has no agency and therefore cannot be accessed “qualitatively” – though we also sometimes mean by “qualitative” eyeballing two time series rather than using some formal measure of tracking. I guess that use would >>not<< be relevant here.]

Discussion

4.1: One could argue that qualitative data analysis provides causes for behaviour (and indications about their importance (ranking); perhaps also the likelihood of occurrence) as well as key themes that are important to be considered in a model. All sounds very useful for the conceptual modelling phase. The difficulty might be to model the impact (how do we know we model it correctly and at the right level), if that is not easily translatable into a quantitative value but requires some more (behavioural) mechanistic structures to represent the impact of behaviours. [And, of course, there is a debate in psychology (with some evidence on both sides) about the extent to which people are able to give subjective accounts we can trust (see Hastorf and Cantril (1954).] This might also provide issues when it comes to calibration – how does one calibrate qualitative data? {Triangulation.} One random idea we had was that perhaps fuzzy logic could help with this. More brainstorming and internet research is required to confirm that this idea is feasible and useful. [A more challenging example might be ethnographic observation of a “neighbourhood” in understanding crime. This is not about accessing the cognitive content of agents but may well still contribute to a well specified model. It is interesting how many famous models – Schelling, Zaller-Deffuant – actually have no “real” environment.]

4.2: One could also argue that what starts (or what we refer to initially) as qualitative data always ends up as quantitative data, as whatever comes out of the computer are numbers. {This is not necessarily true. Check  the work on qualitative outputs using Grounded Theory by Neumann (2015).} Of course this is a question related to the conceptual viewpoint. [Not convinced. It sounds like all sociology is actually physics because all people are really atoms. Formally, everything in computers is numbers because it has to be but that isn’t the same as saying that data structures or whatever don’t constitute a usable and coherent level of description: We “meet” and “his” opinion changes “mine” and vice versa. Somewhere, that is all binary but you can read the higher level code that you can understand as “social influence” (whatever you may think of the assumptions). Be clear whether this (like the “rules” claim) is a matter of definition – in which case it may not be useful (even if people are atoms we have no idea of how to solve the “atomic physics” behind the Prisoner’s Dilemma) or an empirical one (in which case some models may just prove it false). This (Beltratti et al 1996) contains no “rules” and no “numbers” (except in the trivial sense that all programming does).]

4.3: Also, an algorithm is expressed in code and can only be processed numerically, so it can only deliver quantitative data as output. These can perhaps be translated into qualitative concepts later. A way of doing this via the use of grounded theory is proposed in Neumann and Lotzmann (2016). {This refers to the same idea as my previous comment.} [Maybe it is “safest” to discuss this with rules because everyone knows those are used in ABM. Would it make sense to describe the outcome of a non trivial set of rules – accessed for example like this: Gladwin (1989) – as either “quantitative” or “numbers?”]

4.4: But is it true that data delivered as output is always quantitative? Let’s consider, for example, a consumer marketing scenario, where we define stereotypes (shopping enthusiast; solution demander; service seeker; disinterested shopper; internet shopper) that can change over time during a simulation run (Siebers et al 2010). These stereotypes are defined by likelihoods (likelihood to buy, wait, ask for help, and ask for refund). So, during a simulation run an agent could change its stereotype (e.g. from shopping enthusiast to disinterested shopper), influenced by the opinion of others and their own previous experience. So, at the beginning of the simulation run the agent can have a different stereotype compared to the end. Of course we could enumerate the five different stereotypes, and claim that the outcome is numeric, but the meaning of the outcome would be something qualitative – the stereotype related to that number. To me this would be a qualitative outcome, while the number of people that change from one stereotype to another would be a quantitative outcome. They would come in a tandem. {So, maybe the problem is that we don’t yet have the right ways of expressing or visualising qualitative output?} [This is an interesting and grounded example but could it be easily knocked down because everything is “hard coded” and therefore quantifiable? You may go from one shopper type to another – and what happens depends on other assumptions about social influence and so on – but you can’t “invent” your own type. Compare something like El Farol (Arthur 1994) where agents arguably really can “invent” unique strategies (though I grant these are limited to being expressed in a specified “grammar”).]

4.5: In order to define someone’s stereotype we would use numerical values (likelihood = proportion). However, stereotypes refer to nominal data (which refers to data that is used for naming or labelling variables, without any quantitative value). The stereotype itself would be nominal, while the way one would derive the stereotype would be numerical. Figure 1 illustrates a case in which the agent moves from the disinterested stereotype to the enthusiast stereotype. [Is there a potential confusion here between how you tell an agent is a type – parameters in the code just say so – and how you say a real person is a type? Everything you say about the code still sounds “quantitative” because all the “ingredients” are.]

Figure 1: Enthusiastic and Disinterested agent stereotypes

4.6: Let’s consider a second example, related to the same scenario: The dynamics over time to get from an enthusiastic shopper (perhaps via phases) to a disinterested shopper. This is represented as a graph where the x-axis represents time and the y-axis stereotypes (categorical data). If you want to take a quantitative perspective on the outcome you would look at a specific point in time (state of the system) but to take a qualitative perspective of the outcome, you would look at the pattern that the curve represents over the entire simulation runtime. [Although does this shade into the “eyeballing” sense of qualitative rather than the “built from subjective accounts” sense? Another way to think of this issue is to imagine “experts” as a source of data. We might build an ABM based on an expert perception of say, how a crime gang operates. That would be qualitative but not just individual rules: For example, if someone challenges the boss to a fight and loses they die or leave. This means the boss often has no competent potential successors.]

4.7: So, the inputs (parameters, attributes) to get the outcome are numeric, but the outcome itself in the latter case is not. The outcome only makes sense once it’s put into the qualitative context. And then we could say that the simulation produces some qualitative outputs. So, does the fact that data needs to be seen in a context make it evidence, i.e. do we only have quantitative and qualitative evidence on the output side? [Still worried that you may not be happy equating qualitative interview data with qualitative eyeballing of graphs. Mixes up data collection and analysis? And unlike qualitative interviews you don’t have to eyeball time series. But the argument of qualitative research is you can’t find out some things any other way because, to run a survey say, or an experiment, you already have to have a pretty good grasp of the phenomenon.]

4.8: If one runs a marketing campaign that will increase the number of enthusiastic shoppers. This can be seen as qualitative data as it is descriptive of how the system works rather than providing specific values describing the performance of a system. You could also equally express this algebraic terms which would make it quantitative data. So, it might be useful to categorise quantitative data to make the outcome easier to understand. [I don’t think this argument is definitely wrong – though I think it may be ambiguous about what “qualitative” means – but I think it really needs stripping down and tightening. I’m not completely convinced as a new reader that I’m getting at the nub of the argument. Maybe just one example in detail and not two in passing?]

Outcome

5.1: How we understand things and how the computer processes things are two different things. So, in fact qualitative data is useful for the conceptualisation and for describing experimentation and evaluation output, and needs to be translated into numerical data or algebraic constructs for the operationalisation. Therefore, we can reject our initial hypothesis, as we found more places where qualitative data can be useful. [Yes, and that might form the basis for a “general” definition of qualitative that was not tied to one part of the research process but you would have to be clear that’s what you were aiming at and not just accidentally blurring two different “senses” of qualitative.]

5.2: In the end of the discussion we picked up the idea of using Fuzzy Logic. Could perhaps fuzzy logic be used to describe qualitative output, as it describes a degree of membership to different categories? An interesting paper to look at in this context would be Sugeno and Yasukawa (1993). Also, a random idea that was mentioned is if there is potential in using “fuzzy logic in reverse”, i.e. taking something that is fuzzy, making it crisp for the simulation, and making it fuzzy again for presenting the result. However, we decided to save this topic for another discussion. [Devil will be in the detail. Depends on exactly what assumptions the method makes. Devil’s advocate: What if qualitative research is only needed for specification – not calibration or validation – but it doesn’t follow from this that that use is “really” quantitative? How intellectually unappealing is that situation and why?]

Conclusion

6.1: The purpose of this Research Note is really to stimulate you to think about, talk about, and share your ideas and opinions on the topic! What we present here is a philosophical impromptu discussion of our individual understanding of the topic, rather than a scientific debate that is backed up by literature. We still thought it is worthwhile to share this with you, as you might stumble across similar questions. Also, we don’t think we have found the perfect answers to the questions yet. So we would like to invite you to join the discussion and leave some comments in the chat, stating your point of view on this topic. [Is the danger of discussing these data types “philosophically”? I don’t know if it is realistic to use examples directly from social simulation but for sure examples can be used from social science generally. So here is a “quantitative” argument from quantitative data: “The view that cultural capital is transmitted from parents to their children is strongly supported in the case of pupils’ cultural activities. This component of pupils’ cultural capital varies by social class, but this variation is entirely mediated by parental cultural capital.” (Sullivan 2001). As well as the obvious “numbers” (social class by a generally agreed scheme) there is also a constructed “measure” of cultural capital based on questions like “how many books do you read in a week?” Here is an example of qualitative data from which you might reason: “I might not get into Westbury cos it’s siblings and how far away you live and I haven’t got any siblings there and I live a little way out so I might have to go on a waiting list … I might go to Sutton Boys’ instead cos all my mates are going there.” (excerpt from Reay 2002). As long as this was not just a unique response (but was supported by several other interviews) one would add to one’s “theory” of school choice: 1) Awareness of the impact of the selection system (there is no point in applying here whatever I may want) and 2) The role of networks in choice: This might be the best school for me educationally but I won’t go because I will be lonely.]

Biographies of the authors

Peer-Olaf Siebers is an Assistant Professor at the School of Computer Science, University of Nottingham, UK. His main research interest is the application of Computer Simulation and Artificial Intelligence to study human-centric and coupled human-natural systems. He is a strong advocate of Object-Oriented Agent-Based Social Simulation and is advancing the methodological foundations. It is a novel and highly interdisciplinary research field, involving disciplines like Social Science, Economics, Psychology, Geography, Operations Research, and Computer Science.

Kwabena Amponsah is a Research Software Engineer working for the Digital Research Service, University of Nottingham, UK. He completed his PhD in Computer Science at Nottingham in 2019 by developing a framework for evaluating the impact of communication on performance in large-scale distributed urban simulations.

James Hey is a PhD student at the School of Computer Science, University of Nottingham, UK. In his PhD he investigates the topic of surrogate optimisation for resource intensive agent based simulation of domestic energy retrofit uptake with environmentally conscious agents. James holds a Bachelor degree in Economics as well as a Master degree in Computer Science.

Edmund Chattoe-Brown is a lecturer in Sociology, School of Media, Communication and Sociology, University of Leicester, UK. His career has been interdisciplinary (including Politics, Philosophy, Economics, Artificial Intelligence, Medicine, Law and Anthropology), focusing on the value of research methods (particularly Agent-Based Modelling) in generating warranted social knowledge. His aim has been to make models both more usable generally and particularly more empirical (because the most rigorous social scientists tend to be empirical). The results of his interests have been published in 17 different peer reviewed journals across the sciences to date. He was funded by the project “Towards Realistic Computational Models Of Social Influence Dynamics” (ES/S015159/1) by the ESRC via ORA Round 5.

Melania Borit is an interdisciplinary researcher and the leader of the CRAFT Lab – Knowledge Integration and Blue Futures at UiT The Arctic University of Norway. She has a passion for knowledge integration and a wide range of interconnected research interests: social simulation, agent-based modelling; research methodology; Artificial Intelligence ethics; pedagogy and didactics in higher education, games and game-based learning; culture and fisheries management, seafood traceability; critical futures studies.

References

Achter S, Borit M, Chattoe-Brown E, and Siebers PO (2022) RAT-RS: a reporting standard for improving the documentation of data use in agent-based modelling. International Journal of Social Research Methodology, DOI: 10.1080/13645579.2022.2049511

Australian Bureau of Statistics (2022) Statistical Language > Qualitative and Quantitative data. https://www.abs.gov.au/websitedbs/D3310114.nsf/Home/Statistical+Language (last accessed 05/05/2022)

Arthur WB (1994) Inductive reasoning and bounded rationality. The American Economic Review, 84(2), pp.406-411. https://www.jstor.org/stable/pdf/2117868.pdf

Becker HS (1953). Becoming a marihuana user. American Journal of Sociology, 59(3), pp.235-242. https://www.degruyter.com/document/doi/10.7208/9780226339849/pdf

Beltratti A, Margarita S, and Terna P (1996) Neural Networks for Economic and Financial Modelling. International Thomson Computer Press.

Business Research Methodology (2022) Quantitative Data Analysis. https://research-methodology.net/research-methods/data-analysis/quantitative-data-analysis/ (last accessed 05/05/2022)

Chattoe-Brown E (2009) The social transmission of choice: a simulation with applications to hegemonic discourse. Mind & Society, 8(2), pp.193-207. DOI: 10.1007/s11299-009-0060-7

Gladwin CH (1989) Ethnographic Decision Tree Modeling. SAGE Publications.

Hastorf AH and Cantril H (1954) They saw a game; a case study. The Journal of Abnormal and Social Psychology, 49(1), pp.129–134.

Helitzer-Allen DL and Kendall C (1992) Explaining differences between qualitative and quantitative data: a study of chemoprophylaxis during pregnancy. Health Education Quarterly, 19(1), pp.41-54. DOI: 10.1177%2F109019819201900104

Miles MB and Huberman AM (1994) . Qualitative Data Analysis: An Expanded Sourcebook. Sage

Neumann M (2015) Grounded simulation. Journal of Artificial Societies and Social Simulation, 18(1)9. DOI: 10.18564/jasss.2560

Neumann M and Lotzmann U (2016) Simulation and interpretation: a research note on utilizing qualitative research in agent based simulation. International Journal of Swarm Intelligence and Evolutionary Computing 5/1.

Reay D (2002) Shaun’s Story: Troubling discourses of white working-class masculinities. Gender and Education, 14(3), pp.221-234. DOI: 10.1080/0954025022000010695

Siebers PO, Achter S, Palaretti Bernardo C, Borit M, and Chattoe-Brown E (2019) First steps towards RAT: a protocol for documenting data use in the agent-based modeling process (Extended Abstract). Social Simulation Conference 2019 (SSC 2019), 23-27 Sep, Mainz, Germany.

Siebers PO, Aickelin U, Celia H and Clegg C (2010) Simulating customer experience and word-of-mouth in retail: a case study. Simulation: Transactions of the Society for Modeling and Simulation International, 86(1) pp. 5-30. DOI: 10.1177%2F0037549708101575

Skinner J, Edwards A and Smith AC (2021) Qualitative Research in Sport Management – 2e, p171. Routledge.

Strauss AL (1987). Qualitative Analysis for Social Scientists. Cambridge University Press.

Sugeno M and Yasukawa T (1993) A fuzzy-logic-based approach to qualitative modeling. IEEE Transactions on Fuzzy Systems, 1(1), pp.7-31.

Sullivan A (2001) Cultural capital and educational attainment. Sociology 35(4), pp.893-912. DOI: 10.1017/S0038038501008938

Wilkinson D (2022) What’s the difference between data and evidence? Evidence-based practice. https://oxford-review.com/data-v-evidence/ (last accessed 05/05/2022)

Notes

[1] An updated set of the terminology, defined by the RAT task force in 2022, is available as part of the RAT-RS in Achter et al (2022) Appendix A1.

Peer-Olaf Siebers, Kwabena Amponsah, James Hey, Edmund Chattoe-Brown and Melania Borit (2022) Discussions on Qualitative & Quantitative Data in the Context of Agent-Based Social Simulation. Review of Artificial Societies and Social Simulation, 16th May 2022. https://rofasss.org/2022/05/16/Q&Q-data-in-ABM

© The authors under the Creative Commons’ Attribution-NoDerivs (CC BY-ND) Licence (v4.0)