PAM 5100
Multiple Regression Analysis: Estimation
Spring 2016
Matt Hall
Introduction
Often the case in our bivariate model that our
assumption that E[U | X]=0 is unreasonable
The core problem being that there are other variables in U
that are correlate

PAM 5100
Bivariate Linear Regression Model
Spring 2016
Matt Hall
Introduction
Typically interested in assessing relationships between
variables
And, want to draw inferences about this relationship in
population (unobserved) from a sample (observed)
Goal i

Name: []
PAM 5100
PS#3
PROBLEM SET #3
Please answer all questions completely. Follow the directions closely regarding how to present
results. Please type the answers to the questions in this document, and submit the complete, single
document (in Word form

Derivation of OLS estimators using the method of moments technique
Bivariate model relates random variable to an explanatory random variable :
= 0 + 1 +
We use the method of moments to derive estimates of parameters, calling them 0 and 1 . Since we have

PAM 5100
Intro to Statistical Inference & Hypothesis Testing
Spring 2016
Matt Hall
Sampling
We dont actually observe PDFs and CDFs for the
random variables we are interested in.
In the colored ball example, say we did not know the
distribution of colors a

PAM 5100
Summary Statistics
Spring 2016
Matt Hall
What is Statistics?
1
What is Statistics?
In practice, statistics is just a set of tools
But, for getting formal, two types:
1.
2.
Descriptive
Inferential
Distinction between population and sample
Des Moin

PAM 5100
Multiple Regression Analysis: Inference
Spring 2016
Matt Hall
Introduction
Thus far, we have seen that under the Gauss-Markov
Assumptions, OLS is unbiased and we can solve for the
variance of the OLS estimators
Care not just about the expected va

PAM 5100
Intro to Probability
Spring 2016
Matt Hall
Probability and Random Variables
Probability: measure of the likelihood that an event will
occur (P)
Experiment/trial: process my which observation is made
Outcome: result obtained during experiment
Even

PAM 5100
Regression Assumption Violations
Spring 2016
Matt Hall
Introduction
Recall the 6 assumptions of the Classical Linear Model
1.
2.
3.
4.
5.
6.
The population model is linear in parameters
Random sample of n observations cfw_ #$, #&, #(, # : = 1,

PAM 5100
Binary Variables and Interactions
Spring 2016
Matt Hall
Introduction
Often the case that our variables are qualitative in nature that is,
they describe something about the world that is not quantifiably
defined.
Gender
Race
Marital status
Educati

Name: []
PAM 5100
PS#2
PROBLEM SET #2
Please answer all questions completely. To report work from Stata, cut and paste relevant parts of log file into
this document, clearly type the answers in the same document (please emphasize your answers in some way)

Name: []
PAM 5100
PS#4
PROBLEM SET #4
Please answer all questions completely. Follow the directions closely regarding how to present
results. Please type the answers to the questions in this document, and submit the complete, single
document (in Word form