Guide to
Fuzzy Cognitive Mapping
Tools and references to
Fuzzy Cognitive Mapping (FCM)
(Versión en español)
Fuzzy cognitive mapping is a critical tool in participatory research. As a method of data collection and analysis, fuzzy cognitive mapping provides a way to systematise different forms of knowledge- whether from individuals, community groups, health service providers, policy-makers, or from the best available published evidence. Analysing fuzzy cognitive maps offers insights into how different groups make sense of an issue and its contributing factors, including identifying priority areas for action, influence and intervention.
Below you will find resources to support researchers and other people interested in strengthening community voice in evidence-based planning using fuzzy cognitive mapping (FCM). We invite readers to follow the available links and to contact us for additional information. The content is organised into four main chapters: 1) an introduction to FCM and the process of creating maps, 2) resources to facilitate analysis, 3) tools for combining maps, and 4) examples of how we have used FCM in over 20 participatory research projects across eight countries.
If you are new to FCM, we recommend you watch this video and read these papers:
- Introductory video to Fuzzy cognitive mapping with stakeholders
- Cameron et al. 2014. Culturally Safe Epidemiology: Oxymoron or Scientific Imperative. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2962656/
- Andersson et al. 2019. Fuzzy cognitive mapping: An old tool with new uses in nursing research. https://doi.org/10.1111/jan.14192
What is the main topic of the maps?
Before creating maps, you should define the focus or the main question you want to explore. Ideally, this would include a dialogue with key stakeholders using other participatory techniques such as focus groups, interviews or conversations. Please take some time to think and get a clear idea of the main topic you want to consider in your map. These are questions that could help in this process:
- Guiding questions to initiate FCM (Dion, A). This is also a tool to determine whether the Weight of Evidence is an appropriate method for a project’s objectives.
Think carefully about what you want to do with the maps and plan accordingly. If the central “outcome” is too broad, the maps can get complicated to interpret and analyse. You might miss some important issues if the central outcome is too narrow.
Who should participate?
Based on your topic or outcome of interest, you need to identify the relevant sources of information, such as surveys or literature reviews, or relevant stakeholders to participate in the mapping sessions. Individuals or groups can create stakeholder maps. It is usually best to have groups of similar people create each map. For instance, they may be women, young men, community leaders, service providers, patients or others. One reason for this is that we often want to compare maps produced by different stakeholder groups. Think about what comparisons you might want to make when planning your mapping groups.
Importantly, in the group sessions, you need to make sure that all the participants contribute to the map. Small groups of no more than five people are easier to manage.
The mapping team
For each session, someone will need to facilitate the discussion and support the participants in drawing the map. A separate reporter will be responsible for taking notes about the discussion of the concepts and their relationships. The facilitators and reporters need careful training for their roles. It is better if the training includes field practice.
- Training materials for facilitators (Cockcroft, A, Omer, K, Ansari, U, Sarmiento, I) – Please, contact us if you are interested in training activities and materials.
You can create maps from at least three sources: stakeholder knowledge, literature reviews and data from interviews or surveys
Stakeholder maps
In this video, you will find instructions on how to facilitate mapping sessions with stakeholders.
Use this printable guide for your mapping session:
See these references for detailed explanations of mapping protocols (Andersson et al. 2019, Sarmiento et al. 2021)
It is also useful to follow procedures to ensure you have good-quality data. Mapping protocols should consider mechanisms to identify facilitation issues, misunderstanding of concepts, deviation of mapping steps etc.
- Checklist for quality assessment during data collection (Cockcroft, A, Omer, K, Ansari, U) – If you want to use this material, please contact us.
Software to draw maps
These are two free options that we have found very useful to work with maps on your computer.
yEd free desktop application (stand-alone)
You can use yEd to draw and digitise maps. It is very flexible to display maps, and you can import them into the application if they are in tabular format (adjacency matrices or edge lists).
Click here to watch an introductory video on yEd’s useful functionalities for FCM.
Follow this link to watch a video on how to digitise maps with yEd.
Maps from a literature review
Participatory research expands the idea of what counts as evidence, opening space for the experience and knowledge of stakeholders. FCM portrays qualitative and quantitative evidence from the literature in the same terms as stakeholder experience and beliefs, thus becoming a cornerstone of an innovative and systematic approach called the Weight of Evidence. In this approach, stakeholders interpret, expand on and prioritise evidence from literature reviews.
We also have a section with specific references dedicated to the description and application of the Weight of Evidence.
Maps from scoping reviews
This graph shows the process of summarising scoping reviews with FCM. To build literature maps, you will initially create a list of relationships (edge lists) identified in a standard scoping review.
Building literature maps
Maps from interviews and questionnaire data
We used questionnaire data to generate maps of a behavioural change model in dengue prevention in Mexico (Andersson et al. 2017) and cultural safety among medical trainees in Colombia (Pimentel et al. 2021). Please, follow the links to obtain more information about this.
Analysis and interpretation of the maps
There are multiple options for the analysis of the maps. We offer here some tools that you can use in the process. You can find a detailed explanation of four handy analysis tools here (Sarmiento et al., Under review). A general approach for the analysis of stakeholder maps would start with making the maps with individuals or groups, comparing the meaning of the concepts across multiple maps, identifying the total influence across nodes (transitive closure) and creating maps by groups of stakeholders.
Pattern correspondence (pattern matching) tables
It is an informative first step, which requires no specialised software to inform discussions about similarities and differences between stakeholder views. It is based on the principles of classification and thematic analysis to compare how the meaning of concepts is present or changes across maps.
Transitive closure
A module available on CIETmap freeware (see below) converts multiple individual associations in each map into a network of interlocking relations, a structure of causality rather than a collation of independent associations. This identifies direct and indirect perceived influences and influential intervention pathways that might not be apparent from the individual associations.
Reduction (condensation, aggregation and restriction)
It might be helpful to reduce the maps into categories to facilitate communication and discussion of their content. Reduction of the number of concepts in maps facilitates communication and discussion but, because of loss of information, should follow formal and transparent procedures. These involve a qualitative step to organising the nodes into categories and a mathematical procedure to calculate the net influence at the category level.
Social network analysis (SNA)
SNA of transitive closure maps uses measures of centrality based on the number and weights of outgoing and incoming arrows in the maps to identify the most prominent causes and intermediate outcomes. You can use yEd or Mental Modeler to calculate these measures.
Software to analyse the maps
CIETmap Beta 2.2 (FCM Module) for Windows.
This package allows you to calculate Fuzzy and Probabilistic Transitive Closure. It has additional tools to import yEd files, calculate average maps from multiple inputs and use basic functions to display your data.
CIETmap is a work in progress with additional capabilities yet to come.
Please, make sure you update the package after you have installed it. In the Tools menu, you will find the Update option.
You can use yEd to draw and digitise maps. It is very flexible to display maps, and you can import them into the application if they are in tabular format (adjacency matrices or edge lists). The application also has handy tools to calculate centrality measures.
Follow this link to see a video with a detailed explanation of how to digitise maps.
Other tools and procedures
Script to convert edge lists into adjacency matrices: Excel
Script to convert adjacency matrices into edge lists: Excel
Adjacency matrix into edge list: R script (contact us)
Spreadsheet for thematic analysis and adjusting the sign of relationships (contact us)
Spreadsheet to create category-level maps (contact us)
Spreadsheet to generate discourse analysis maps (contact us)
Script to convert yEd files (.TGF) into edge lists (in progress)
Combination of maps
Central to the use of fuzzy cognitive mapping in participatory research is the recognition that different stakeholders bring different knowledge, presenting opportunities to combine different viewpoints within or across stakeholder groups to better address shared concerns. For example, you can build average maps by stakeholder group. You can see a more detailed explanation of the procedures to combine maps here Dion et al., under review.
This image summarises the steps and tools to create stakeholder maps, which combines a qualitative step based on thematic analysis and pattern matching with a mathematical procedure to calculate the reconciled or updated weights of the combined map:
The Weight of Evidence
The combination of maps is the cornerstone of this method to contextualise published evidence in stakeholder knowledge. At the end of this page, you will find a section with detailed references about this.
Applications in developing theories of change, co-design and quality improvement strategies
We have applied FCM in more than 20 studies in eight countries since 2016. You can find a description of our experience here (Sarmiento et al., Under review).
In a section at the end of this page, you will find additional references with more details on FCM use in participatory research.
- Sarmiento et al. 2022. Maternal health and Indigenous traditional midwives in southern Mexico: contextualisation of a scoping review. https://doi.org/10.1136/bmjopen-2021-054542
- Ghadirian et al. 2021. Participatory Video Influenced Behaviour and Social Environment of Ghanaian Female Adolescent Learners: A Most Significant Change Evaluation. https://doi.org/10.1093/cdn/nzab051_018
- Dion. 2021. Chapter 2: The influence of social and economic exclusion on perinatal health outcomes and care experiences in Canada: An integrative review. https://escholarship.mcgill.ca/concern/theses/02871189q
- Dion. 2021. Chapter 6: How adolescent mothers interpret and prioritize evidence about perinatal child protection involvement: participatory contextualization of published evidence. https://escholarship.mcgill.ca/concern/theses/02871189q
- Dion et al. 2021. Evidence-based priorities of under-served pregnant and parenting adolescents: addressing inequities through a participatory contextualizing evidence syntheses. https://doi.org/10.1186/s12939-021-01458-7
- Pimentel et al. 2021. Impact of game jam learning about cultural safety in Colombian medical education: a randomised controlled trial. https://doi.org/10.1186/s12909-021-02545-7
- Sarmiento et al. 2021. Causes of short birth interval (kunika) in Bauchi State, Nigeria: systematizing local knowledge with fuzzy cognitive mapping. https://doi.org/10.1186/s12978-021-01066-2
- Belaid et al. 2021. Community views on short birth interval in northern Uganda: A participatory grounded theory. https://doi.org/10.1186/s12978-021-01144-5
- Belaid et al. 2021. Communities and service providers address access to perinatal care in postconflict Northern Uganda: socialising evidence for participatory action. https://doi.org/10.1136/fmch-2020-000610
- Pimentel et al. 2021. Game jams for cultural safety training in Colombian medical education: a pilot randomised controlled trial. https://doi.org/10.1136/bmjopen-2020-042892
- Tratt et al. 2020. Fuzzy cognitive mapping with Inuit women: What needs to change to improve cervical cancer screening in Nunavik, Northern Quebec? https://doi.org/10.1186/s12913-020-05399-9
- Andersson et al. 2017. The women made it work: fuzzy transitive closure of the results chain in a dengue prevention trial in Mexico. https://doi.org/10.1186/s12889-017-4301-0
You can follow this link to download the Guide to the Weight of evidence
- Dion et al. 2022. Weight of Evidence: Participatory Methods and Bayesian Updating to Contextualize Evidence Synthesis in Stakeholders’ Knowledge. https://doi.org/10.1177/15586898211037412
- Sarmiento et al. 2022. Maternal health and Indigenous traditional midwives in southern Mexico: contextualisation of a scoping review. https://doi.org/10.1136/bmjopen-2021-054542
- Dion et al. 2019. Grounding evidence in experience to support people-centered health services. https://doi.org/10.1007/s00038-018-1180-9
También hacemos disponibles las traducciones de algunas referencias para facilitar su uso en contextos hispanohablantes.
We are open to provide training to those interested in using FCM. Please, reach out to us if you want to know more about this posibility.
For trainees:
Please do not hesitate to contact us if you have any questions or would like support in adapting these tools to your work and context. We continue to advance our own work with fuzzy cognitive mapping and are always interested in hearing about your experiences using and/or adapting these tools.