Earhart is an 8 step process, plus the output:
- Create the desired model based on the specified list of DVs and IVs;
- Create an Amelia object from the model specified in Step 1. Meld the imputations to get regression coefficients;
- Extract the imputations from the Amelia obejct;
- Use the extracted values from Step 3 to manually calcualte the mean across imputations (this is done by the Amelia::meld in Step 2, but it stays under the hood);
- Calcualte y-hat values using the means from Step 4 and the regression coefficients from Step 2;
- Calcualte sum-of-squares of the y-hat values from Step 5 and the observed/imputed y values;
- Calcualte the multiple R-squared effect size using the values from Step 6;
- Use a specified R-squared adjustment to correct for model size and observations (UNDER CONSTRUCTION)
END: Output the Multiple R-squared effect size.