Home
Subjects
Create
Search
Log in
Sign up
Upgrade to remove ads
Only $2.99/month
Social Science
Economics
Econometrics
ECO 252 - FINAL
STUDY
Flashcards
Learn
Write
Spell
Test
PLAY
Match
Gravity
Terms in this set (14)
Ordinary Least Squares Method
A statistical technique which attempts to find a function which most closely approximates the data ("best-fit" measure). With OLS, we want to minimize "e" variables and find out alpha and beta, thus giving us significance. From OLS, we get the regression equation (prediction equation).
Endogeneity (definition, consequences, detect, fix)
Some or all x's in a model are correlated with "e"
- Consequences: p-values are too high, coefficients ARE biased
- Detect: observe p-values, critical thinking
- Fix: collect more data, make e term as random as possible
Multicollinearity (definition, consequences, detect, fix)
When 2 or more x's are correlated
- Consequences: p-values are too high, coefficients MAY be biased
- Detect: Use SPSS to observe VIF's; VIF > 5 = problem!
- Fix: eliminate coefficients with high VIF and run regression again
- "What substitutes can you come up with for the variable with high VIF" ... THEN I come up with a few that MAY NOT correlate with other x's
Omitted Variable Bias
When omitted variables (in e term) correlate with one or more x's
- For test, we would have to make up variables that may correlate with any original x's
- If any of the ones I make up correlate with original x's, we have endogeneity
Reverse Causality
When y affects some or all of the x's
- EX: do final exam scores (y) affect current GPA, tardiness, hours of study, sex (all x's)?
- If y does affect any x's, we have endogeneity
Adj. R2 (interpretation)
_% of changes in (y) are explained by the regression
R2 (interpretation)
_% of changes in (y) are explained by the regression
Is Adj. R2 or R2 a better goodness-of-fit measure? Why and why not?
Adj. R2 is a better measure because it increases with significance (lower p-values) and eliminating multicollinearity
How do you identify the better model?
The better model has a higher Adj. R2, lower Significance F., and a lower Standard Error
Interpretation of dummy variable (and what is it?)
The additional effect of being __ compared to __ is #, ALL ELSE EQUAL
Dummy -> non-numerical data
What is considered a high VIF; what do VIF's mean?
If the VIF's of two X variables are > 5, it suggests that the variables are correlated and we lose significance
Interpretation of significance (p-values)
EX: p-value = 0.043 < 0.05, so the coefficient is significant. It is not 0.
Interpretation of numerical x's
EX: for every one hour increase in studying time, final exam scores are predicted to increase by 1.05, ALL ELSE EQUAL
Assumptions for Multiple Regression Analysis
1) y depends on x's and e linearly
2) e terms are normally distributed with mean = 0
3) x's and e are uncorrelated
4) x's are also uncorrelated
5) homoscedasticity; y has constant variance throughout
PROBLEM: heteroscedasticity
You might also like...
Week 10 & 11: Standard Multiple Regressi…
103 terms
Unit 2 vocab
22 terms
Stats Final
25 terms
Statistics Vocab Part 2
25 terms
Other sets by this creator
ARH 103 - Final Exam
40 terms
ARH 103 - Exam 5
70 terms
ARH 103 - Exam 4
79 terms
FIN 326 - Exam 2 Concepts
45 terms
Other Quizlet sets
CLOA IB Psych
36 terms
BIO 2100 2B Nervous System
40 terms
Head to Toe Assessment Exam
11 terms
EDPY 303 final exam study notes
92 terms