00:01
Part 1.
00:03
Out of 990 students in the sample, the number that were never awarded a voucher are 468.
00:19
And you will find that number in r by summing all the cases.
00:32
This is the name of the dataset.
00:36
Dollar sign and select years equal.
00:47
Equal zero.
00:56
You are counting the total number of cases where select years is zero.
01:08
How many students had a voucher available for four years? so similarly, you find the total number of cases were selected years, double equal four.
01:28
And you will get 100.
01:30
8 students.
01:34
Then how many students actually attended a choice school for 4 years? counting all the cases where variable choice years equal equal 4.
01:50
You get 56.
01:58
Part 2, you run a simple regression of choice years on select years.
02:03
This is what i get.
02:05
I get their slope coefficient to be minus 1 .837.
02:11
And this one is significant.
02:14
The minus sign is not what i expect.
02:23
This regression is telling us that this is actually part three.
02:46
All right, so in part three, you run a simple regression of math percent on choice years.
02:56
And this is not what i expected because this result is telling me being able to attend choice years actually decreases math scores.
03:14
And the variable, the estimate on choice years, does not change when i add the demographic variables, black, hispanic, and female.
03:48
So i accidentally skipped part two.
03:53
Let's move down to the end in part two.
04:12
You are examining whether select ears is a good instrument for choice ears.
04:23
And the answer is yes.
04:34
This is a regression result.
04:58
This equation is telling us being selected for a voucher increases the chance of the students attending the school of choice.
05:10
And the relationship is strong, almost one to one, and also highly significant.
05:34
Select ears can be a good instrument, assuming that it does not correlate with the error terms in the choice years.
05:45
Now in the math score equation.
05:53
Let's continue with part 4.
06:00
Why might choice years be endogenous in the given? equation.
06:07
I could think of two reasons.
06:12
First one, omitted variable bias.
06:20
It is possible that we have not accounted for some factors that affect both the math scores and the number of years students attending the school of choices.
06:33
For example, intellectual ability.
06:44
And the second reason is reverse causality...