DSE 2011 Q38

[Siddhant]

[akanksha]

hey the ans is a)include 1 dummy variable in the regression model
otherwise it will lead to duumy variable trap
dummy variable is used for quality variable and assumes value 0(for benchmark category) and 1 (for other category)
duumy variable trap states that if there  is a common imtercept term and if qualtitative variable has m categories..then include only m-1 dummy variables..
here there are 2 categories( boys and girls,here we are essentially comparing gender differences in immunization )
so include m-1 ie
1 category (a)

[xyz123]

It should be a)

According to me equation should be:

Y = b1 + b2D1 + b3X3 + b4X4 + b5X5

[Akshay Jain]

the above question states that after controlling for other variables we want to get hold of gender differences so there shud be no interaction dummy in the model as if we include the interaction dummy then we cant intrpret the coefficient by controlling the other variables. Further there are 2 categories so 1 dummy shud be included and the logic is that the coefficient of dummy variable, when we include the intercept term in the model, gives the difference in dependent variable across the categories (which is the requarement of the ques)
so ans is option a