L1 Flashcards

1
Q

What is a model?

A

Explanation of the world in an easy way with variables

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

How are y and x related?

A

Linearly

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What does u capture?

A

Everything that determines y, that is not x

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Name observable, and non-observable: y, x, u.

A

y and x observable. u unnobservable.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What are Betas?

A

Unobservable parameters -> we want to estimate

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What is B1 in y = B0 + B1 x + B2 x2 + u?

A

CAUSAL effect of x on y, ceteris paribus.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What is spurious correlation?

A

Correlation without causation.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What is the key assumption for causality?

A

Zero Conditional Mean Assumption -> E[u|x] = E[u] = 0

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Divide the systematic part from the idiosyncratic part.

A

Starting with E[y|x] = B0 + B1 x

y = E[y|x] (systematic) + u (idiosyncratic)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

For the population sample, yi equals?

A

yi = E[y|xi] + ui

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Symbol to represent estimated error:

A

ûi

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

ûi equals:

A

yi - ^yi

(Estimated error = real value - estimated value)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What is the goal of choosing ^B0 and ^B1?

A

Minimizing ûi (squared)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

^B1 equals:

A

^B1 = cov(x,y) / var(x)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

^B0 equals:

A

Intercept -> avg(y) - ^B1 avg(x)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Do the properties of OLS estimators always hold true?

17
Q

Explain the difference between errors and residuals

A

Errors (u) are never observed -> distance between the observations and the PRF

Residuals (û) are captured from data -> distance between observations and estimated regression function

18
Q

Difference between PRF and Estimated RF?

A

The first one is for the population (almost theoretical -> the real one), the other is the one we estimate.

19
Q

What does SST measure?

A

Total sample variation in the yi

20
Q

What does SSE measure?

A

Sample variation in the ^yi

21
Q

What does SSR measure?

A

Sample variation in the ^ui

22
Q

What is R-squared?

A

How much of the total variation can be explained by the model.

23
Q

R-squared formulas:

A

SSE / SST

1 - SSR / SST

Formula Sheet

24
Q

What does an higher R-squared mean?

A

Higher proportion of variation in yi is explained by variation in xi (as long as they are not correlated)

25
Types of Scaling variables:
Scaling y Scaling x Shifting y Shifting x
26
Describe Scaling y:
all coefficients are scaled y = B0 + B1 x + u c y = c B0 + c B1 x + c u
27
Describe Scaling x:
Slope coefficient is scaled y = B0 + B1 x + u y = B0 + B1/c x + u
28
Decribe Scaling both dependent and independent variables:
y = B0 + B1 x + u c y = c B0 + c B1/d x + c u
29
Describe shifting the dependent variable:
Intercept shifts, slope is unchanged. y = B0 + B1 x + u (y + c) = (B0 + c) + B1 x + u
30
Describe shifting the independent variable:
Intercept shifts, slope unchanged y = B0 + B1 x + u y = (B0 - B1 c) + B1 (x + c) + u
31
Do changes in scaling of shifting have any effect on significance or interpretation?
No
32
Usefulness of logs?
If y > 0, can mitigate skewness and heteroskedacity by reducing the influence of outliers
33
When not to take logs?
Variables measured in years, months, etc Proportions (rates)
34
Limitations of logs
Can't be used if variables takes 0 or negative values (function log is undefined there) If y is non-negative but may be equal to zero: log (1+y) may be a solution (assuming few zeros)
35
log(y)=x
y = e^x
36
What to do if log isn't working?
Quadratic terms
37
log(wage) = B0 + B1 educ + B2 (educ^2) + u What is B1?
No longer the ceteris paribus effect of educ on log(wage)