1

##### Exams and Grading / Re: Final Exam Grade Distribution

« Last post by**ateyf**on

*May 11, 2017, 11:06:15 PM*»

Is there a way that we can find out how we did on the exam?

1

Is there a way that we can find out how we did on the exam?

2

Hi all,

The grade distribution for the final exam is attached below:

<Class 12:30pm>

- Material Question: mean 57.1, s/d 8.1, max 73, min 37

- Essay Question: mean 20.9, s/d 1.5, max 24, min 17

- Total: mean 78.0, s/d 8.9, max 96, min 54

<Class 3:25pm>

- Material Question: mean 62.2, s/d 8.8, max 73, min 37

- Essay Question: mean 21.7, s/d 1.3, max 24, min 19

- Total: mean 83.9, s/d 9.1, max 96, min 57

The grade distribution for the final exam is attached below:

<Class 12:30pm>

- Material Question: mean 57.1, s/d 8.1, max 73, min 37

- Essay Question: mean 20.9, s/d 1.5, max 24, min 17

- Total: mean 78.0, s/d 8.9, max 96, min 54

<Class 3:25pm>

- Material Question: mean 62.2, s/d 8.8, max 73, min 37

- Essay Question: mean 21.7, s/d 1.3, max 24, min 19

- Total: mean 83.9, s/d 9.1, max 96, min 57

3

Hi, are we able to pick up our project papers?

Thanks!

Thanks!

4

Hi all,

Below are the grade distributions for the project:

1. Class 12:30pm

Mean: 64.4

SD: 14.9

Max: 95

Min: 32

2. Class 3:25pm

Mean: 63.9

SD: 11.1

Max: 91

Min: 40

Best,

Alexis

5

Hi Jason, there is no specific "formula" for a good essay response, but speaking generally, your essay should be sufficiently detailed (especially for certain parts - some of these will be determined by the certain topic or question you are given) and specific (don't just say that you will be careful of sample selection - how? what will you look for to determine if your sample is selected? Is it likely your sample will be selected? What will you do if your sample is selected?).

When I was preparing for my final, I found it helpful to rewrite Notes 20 into a structured list. Then I would come up with a few topics / research questions, and practice briefly writing through how I could go about researching it, keeping in mind my rewritten Notes 20. This way, you are able to recognize things you would not have given much thought to (sample selection may be a big issue, I need to be sure I touch on points X, Y, and Z relating to it) and helps you organize your thinking into a clear structure. Some points, depending on the topic, will need more focus than others. For example, some topics will need a much, much, much more careful evaluation of the casual model behind a particular economic process than others (for example, compare the causal model here https://economics.mit.edu/files/12736 - in 'Conceptual Underpinnings' and the causal model here https://economics.mit.edu/files/11556, both papers written by the same economist). It really depends on what topic you are given, though if you are unsure, you should err on the side of giving more detail (remember you have a time limit too). Some areas of your response will need a lot of discussion regardless of the field - interpretations of your results (examining endogeneity, fixing your model, conducting robustness checks) is very important and shouldn't be neglected in any essay response.

6

For the second example (#4), this is much more tricky, as the true functional form of the regression is nonlinear. Essentially here, in order to solve for the bias, we need to transform a linear equation into a nonlinear one, so comparing omitted variables isn't going to be strictly identical to how we would do it for the linear case. In this case, I view this as asking "what is the missing specification?"

7

Hi cchang46, I believe this just means that the "variable" itself is log(x_{1})-x_{1}, but the whole term to be added to get the true model takes into account the coefficient B1. It is the same as saying x_{1} is the variable, not B_{1}X_{1}. A much more simple version to see is:

True model: y = B_{0} + B_{1}X_{1} + B_{2}X_{2} + B_{3}X_{2}^{2} + u

Model you run: y = B_{0} + B_{1}X_{1} + B_{2}X_{2} + u

Then B_{0} + B_{1}X_{1} + B_{2}X_{2} + B_{3}X_{2}^{2} + u = B_{0} + B_{1}X_{1} + B_{2}X_{2} + u + *A* (solve for *A*)

*A *= B_{3}X_{2}^{2}

"A" is the whole term you need to include; you omitted X_{2}^{2}.

True model: y = B

Model you run: y = B

Then B

"A" is the whole term you need to include; you omitted X

8

Hello!

I have a question on how to determine the omitted variable. On page 5, we need to add the term B1(log(x1)-x1) for the first case of misspecification in order to get the true model.

Please correct me if I am wrong, but in class the omitted variable was determined to be log(x1)-x1.

This leaves out B1 when writing the omitted variable--why is that? Is this normal practice?

How then would we specify the omitted variable in the example directly after? In the model this was an addition of [(e^(B1x1+B2x2))-B1x1-B2x2]. Do we omit the betas here as well?

Thank you!

I have a question on how to determine the omitted variable. On page 5, we need to add the term B1(log(x1)-x1) for the first case of misspecification in order to get the true model.

Please correct me if I am wrong, but in class the omitted variable was determined to be log(x1)-x1.

This leaves out B1 when writing the omitted variable--why is that? Is this normal practice?

How then would we specify the omitted variable in the example directly after? In the model this was an addition of [(e^(B1x1+B2x2))-B1x1-B2x2]. Do we omit the betas here as well?

Thank you!

9

We will need to write an essay for the final. What kinds of things do you expect to see in a good essay? What points should be focused on?

10

Hey guys,

In preparation for your upcoming final I will be holding an extra study group tomorrow (5/6) from 7-9 pm in Carlson 1A!

Happy studying!

Laura

In preparation for your upcoming final I will be holding an extra study group tomorrow (5/6) from 7-9 pm in Carlson 1A!

Happy studying!

Laura