Question

Need some help to start derive the general formula for leverage in a simple regression model using a single binary covariate? Consider a data set where the variable X takes only 0 or 1 values (e.g., Yes/No type, we code Yes = 1 and No = 0)

Answer 1

I know leverage is \[h _{ii} =\frac{ 1 }{ n }+\frac{ (X_i-X^-)^{2} }{ \sum_{k=1}^{n}(X_k-X^-)^2 }\] \[X^-\] is mean of X