wrapper function structure

[manncz]

A wrapper function called dRCT() should include the following inputs:

1. required:
   • either outcome vector, covariate matrix, and treatment assignment vector
   • or data and formula
2. optional
   • vector of auxiliary predictions or indication of the variable name in the data frame given
   • vector or numeric for treatment probability and if given assumes Bernoulli randomization
   • blocking variable.
     - If none given, assumes no block/paired design.
     - If given, overrides treatment probability and assumes complete randomization within each block and gives a warning
   • if neither of the two above given, assumes complete randomization of individual units ("this isn't implemented yet")
   • clustering variable (if not pairs - "this isn't implemented yet message")
   • weighting variable (if not pairs - "this isn't implemented yet message")
   • interpolation function
     - if not given, use default for the type of randomization given
     - if given, check if it is compatible with the type of randomization and if not, throw an error

The wrapper should check if all needed inputs are there and if they are consistent. It will give a message if the randomization design is not yet supported.

the method function should renamed to loop.paired and loop.bernoulli

[manncz]

Commit 8430722 to wrapper branch implements most of above.

I have not renamed the method functions for backward compatibility with analyses previously ran - need to figure out a workaround for this.

A couple notes on what remains to be done:

• Much more checking of inputs
• Documentation
• Suppressing annoying dplyr messages