Questions 9-15 Laptop access and GPA. Based on a random sample of 500 high-school students without laptops, a non-profit organization randomly lends laptops to 250 of them for the fall semester. A researcher wants to investigate whether being lent a laptop affects a student's grades, and collects data on the students' GPA for the Fall semester.D Question 9 4 pts For each student, the researcher has data on the Fall semester GPA (GPA), the treatment X (equals to 1 if the students is lent a laptop, and 0 otherwise), and pretreatment characteristics FEMALE (equals to 1 if the student is a female, and 0 if male), and INCOME (monthly family income in dollars) The researcher estimates the following regression by OLS: GPA; = Bo + B1Xi + 32FEMALEi + B3INCOMEi + Wi Why do we add FEMALE and INCOME into the regression if we know that laptops are randomly lent to students? (Check all that apply) Additional regressors typically reduce the standard error of B1 . O If the randomization is not perfect, and female students and low-income students are more likely to be lent laptops, then B1 may be biased and inconsistent if we exclude FEMALE and INCOME. If B1 is very different from the B, we obtain when we do not include FEMALE and INCOME, then this would suggest that there may have been failures in the randomization process. O None of the above