13.Qualitative07.Coincidence analysis (CNA) - sporedata/researchdesigneR GitHub Wiki

1. Use cases: in which situations should I use this method?

Coincidence Analysis (CNA), of which QCA (Qualitative Comparative Analysis) is a subtype, is often used in studies with small sample sizes, but where researchers still want to investigate the association between certain co-existing variables and outcomes (known as conditions in the QCA literature).
Healthcare policy analysis - see Covid-19 pandemic by the "real-time" monitoring: the Tunisian case and lessons for global epidemics in the context of 3PM strategies
Marketing applied to healthcare - see A Qualitative and Quantitative Comparative Analysis of Commercial and Independent Online Information for Hip Surgery: A Bias in Online Information Targeting Patients?
As an aid in meta-analyses, for example Conventional weight loss interventions across the different BMI obesity classes: A systematic review and quantitative comparative analysis
As part of mixed methods studies - see Inclusion of intimate partner violence-related content within undergraduate health care professional curriculum: mixed methods study of academics' attitudes and beliefs
Implementation studies - see Qualitative Comparative Analysis: A Mixed-Method Tool for Complex Implementation Questions

2. Input: what kind of data does the method require?

At least one outcome (condition) and a few predictors.

3. Algorithm: how does the method work?

Model mechanics

Describing in words

Crisp CNA
- The technique relies on finding possible combinations of variables (i.e., order is not important) that are often associated with a given outcome
- Crisp CNA has a standardized set of steps
  1. Crisp set calibration (variable recategorization) for categorical and continuous variables, using a mix of theory-driven and data analysis.
  2. Testing of necessity relations, which is measured by counting the frequency of the occurrence of the outcome given the presence of a specific predictor, as a percentage of all all outcomes events. This analysis is aided by Venn diagrams.
  3. Analysis of sufficiency relationss, which includes complex, parsimonious and intermediate solutions. These include the analysis of all combinations of predictors leading to the outcome.
  4. Plotting of results, often including Venn diagrams
Fuzzy CNA
- In contrast with crisp CNA, each of the variables is not Boolean (i.e., having a true/false value), but instead can have continuous values ranging between 0 and 1.
- The two calibration methods include direct and transformational assignment.
Multivalue CNA, where the outcome (condition) has more than two categories
Temporal CNA, where the condition (outcome) measurement is repeated over time.

Describing in images

Describing with code

Breaking down equations

Reporting guidelines for Methods

Data science packages

QCA
Introduction to the CNA method and package [3]
The package cutpointr provides tools to determine optimal cutpoints and may assist in the calibration of crisp datasets.

Suggested companion methods

Visualization

Learning materials

Books

Articles

4. Output: how do I interpret this method's results?

Mock conclusions or most frequent format for conclusions reached at the end of a typical analysis.

Tables, plots, and their interpretation

CRISP QCA models
- Necessity tables
- Sufficiency tables
- Venn diagrams

Reporting guidelines for Results

5. SporeData-specific

Templates

Data science functions

References

[1] Baumgartner M. Inferring Causal Complexity. Sociological methods & research. 2009 Aug;38(1):71-101.

[2] Baumgartner M, Ambühl M. Causal modeling with multi-value and fuzzy-set Coincidence Analysis. Political Science Research and Methods. 2020 Jul;8(3):526-42.