review diff-in-diff

Author: matheusfacure

causal-inference-for-the-brave-and-true/06-Grouped-and-Dummy-Regression.ipynb

@@ -684,7 +684,7 @@
    "source": [
     "In this case, when the person hasn't completed 12th grade (dummy off), the average income is 19.9. Whe he or she has completed 12th grade (dummy on), the predicted value or the average income is 24.8449 (19.9405 + 4.9044). Hence, the dummy coefficient captures the difference in means, which is 4.9044 in our case.\n",
     "\n",
-    "More formality, when the dependent variable is binary, as is often the case with treatment indicators, regression captures the ATE perfectly. That is because regression is a linear approximation to the conditional expectation function \\\\(E[Y|X]\\\\) and, in this particular case, the CEF IS linear. Namely, we can define \\\\(E[Y_i|T_i=0]=\\alpha\\\\) and \\\\(E[Y_i|T_i=1] = \\alpha + \\beta\\\\), which leads to the following CEF\n",
+    "More formally, when the dependent variable is binary, as is often the case with treatment indicators, regression captures the ATE perfectly. That is because regression is a linear approximation to the conditional expectation function \\\\(E[Y|X]\\\\) and, in this particular case, the CEF IS linear. Namely, we can define \\\\(E[Y_i|T_i=0]=\\alpha\\\\) and \\\\(E[Y_i|T_i=1] = \\alpha + \\beta\\\\), which leads to the following CEF\n",
     "\n",
     "$\n",
     "E[Y_i|T_i] =  E[Y_i|T_i=0] + \\beta T_i = \\alpha + \\beta T_i\n",