CausalQueries: Make, update, and query causal models

Start by making a model, refine it if needed, update with data if needed, then pose queries.

1. Input Model

2. Set Restrictions (Optional) 


              

              

                

                  3. Set Parameters (Optional)
                  
                

                

                

              

            

            

              

                
Current model

                

              

              

                
Model parameters

                

              

            

          

        

        

          

            

              

                

                  1. Complete Data Types
                  
                

                
Enter counts for fully observed data types (default 0).

                

              

              

                

                  2. Partial Data Types
                  
                

                
Select a strategy and enter counts for all implied data types.

                

                  
                  

                

                

              

            

            

              

                

                  3. Update Options
                  
                

                

                  

                    
                    
                  

                  

                    
                    
                  

                  
                  
                

              

              

                
Current data

                

              

              

                
Stan Summary

                

              

            

          

        

        

          

            

              

                

                  Input Queries
                  
                

                

                  
                  
                

                

                

                  
                  
                

                

              

              

                
Options

                

                  
                  

                    

                      
                    

                    

                      
                    

                    

                      
                    

                  

                

                
              

            

            

              

                
Query Plot

                

                
              

              

                
Query Results

                
                

              

            

          

        

        

          

            

              

                
Minimal Replication Code

                

                  
                

                

                  

                

                
              

            

          

        

        

          

            This tab gives intuition for how a case-level inference is made (Bayesian process tracing). 
            On the left you input the data you see along with your query; on the right we then show the set of 
            'causal types' that are (a) consistent with the data and (b) consistent with the data and the query. 
            The final inference is the probability of the latter divided by the probability of the former.
          

          

            

              

                
Case level data

                

              

              

                
Case level query

                

                  
                  
                

                
              

            

            

              

                
Results

                

                
Summary

                

              

              

                
Detailed Results Table

                

              

            

          

        

        

          

            

              

                
About the App

                
            
This "shiny" app lets you explore the CausalQueries package. The CausalQueries R package, maintained by Till Tietz, lets you declare binary causal models, update beliefs about causal types given data and calculate arbitrary estimands. Model definition is implemented via a dagitty style syntax. Updating is implemented in Stan.

          
              

              

                
Authors

                
            
Macartan Humphreys and Alan Jacobs are the authors of Integrated Inferences.

          
              

            

            

              

                
Background

                
            
For more background see Integrated Inferences, which provides an introduction to fundamental principles of causal inference and Bayesian updating and shows how these tools can be used to implement and justify inferences using within-case (process tracing) evidence, correlational patterns across many cases, or a mix of the two.

          
              

              

                
Resources

                
            
Learn more about CausalQueries and related resources at Integrated Inferences.

          
              

              

                

                  CausalQueries logo
                

              

            

          

        

      

    

    

    
For more about CausalQueries and the Integrated Inferences framework see resources.