changes from klone

[ml_measurement_error_public.git] / simulations / robustness_check_notes.md

diff --git a/simulations/robustness_check_notes.md b/simulations/robustness_check_notes.md
index 0a287a71e3c6d8be79647c53112a4e40a44095eb..ac7e88fac4ab937b5e274de557606eb702702310 100644 (file)
--- a/simulations/robustness_check_notes.md
+++ b/simulations/robustness_check_notes.md
@@ -2,13 +2,38 @@
 
 Tests how robust the MLE method for independent variables with differential error is when the model for $X$ is less precise. In the main paper, we include $Z$ on the right-hand-side of the `truth_formula`. 
 In this robustness check, the `truth_formula` is an intercept-only model.
 
 Tests how robust the MLE method for independent variables with differential error is when the model for $X$ is less precise. In the main paper, we include $Z$ on the right-hand-side of the `truth_formula`. 
 In this robustness check, the `truth_formula` is an intercept-only model.

-The stats are in the list named `robustness_1` in the `.RDS` file.
-
+The stats are in the list named `robustness_1` in the `.RDS` 

 # robustness\_1\_dv.RDS
 
 # robustness\_1\_dv.RDS
 

-Like `robustness\_1.RDS` but with a less precise model for $w_pred$.  In the main paper, we included $Z$ in the `outcome_formula`. In this robustness check, we do not.
+Like `robustness\_1.RDS` but with a less precise model for $w_pred$.  In the main paper, we included $Z$ in the `proxy_formula`. In this robustness check, we do not.

 
 # robustness_2.RDS
 
 
 # robustness_2.RDS
 

-This is just example 1 with varying levels of classifier accuracy. 
+This is just example 1 with varying levels of classifier accuracy indicated by the `prediction_accuracy` variable.. 
+
+# robustness_2_dv.RDS
+
+Example 3 with varying levels of classifier accuracy indicated by the `prediction_accuracy` variable.
+
+# robustness_3.RDS
+
+Example 1 with varying levels of skewness in the classified variable. The variable `Px` is the baserate of $X$ and controls the skewness of $X$.
+It probably makes more sense to report the mean of $X$ instead of `Px` in the supplement.
+
+# robustness_3_dv.RDS
+
+Example 3 with varying levels of skewness in the classified variable. The variable `B0` is the intercept of the main model and controls the skewness of $Y$.
+It probably makes more sense to report the mean of $Y$ instead of B0 in the supplement. 
+
+# robustness_4.RDS
+
+Example 2 with varying amounts of differential error. The variable `y_bias` controls the amount of differential error.
+It probably makes more sense to report the corrleation between $Y$ and $X-~$, or the difference in accuracy from when when $Y=1$ to $Y=0$ in the supplement instead of `y_bias`.
+
+# robustness_4_dv.RDS
+
+Example 4 with varying amounts of bias. The variable `z_bias` controls the amount of differential error.
+It probably makes more sense to report the corrleation between $Z$ and $Y-W$, or the difference in accuracy from when when $Z=1$ to $Z=0$ in the supplement instead of `z_bias`.
+
+