Midterm Flashcards
variance
standard deviation^2
variance of the mean
standard dev^2 / n
SE of the mean
s / sqrt (n) OR sqrt (Stddev^2 / n)
SE of the mean difference
square root: (Sa^2/Na + Sb^2/Nb)
95% confidence interval implies p value is ____ and want |t| ______
95% confidence interval implies p value is less than .05 and want |t| greater than 1.96
t = x-bar - u / SE(x)
x-bar =
Ux=
SE =
x-bar = sample average Ux= population parameter of interest SE = standard error
Omitted Variable Bias
Correl x1, x2: +
B2: +
Bias =
Bias = +
Omitted Variable Bias
Correl x1, x2: -
B2: -
Bias =
Bias = +
Omitted Variable Bias
Correl x1, x2: +
B2: -
Bias =
Bias = -
Omitted Variable Bias
Correl x1, x2: -
B2: +
Bias =
Bias = -
Bias +
B1 -
over/under estimating bias?
underestimating
If the bias leads our regression coefficient to be larger in absolute value than it should be (i.e. if it moves us away from zero), we say that we are overstating (or overestimating) the effect of X on Y. This happens when B1 and the bias have the same sign
Bias +
B1 +
over/under estimating bias?
overestimating
If the bias leads our regression coefficient to be larger in absolute value than it should be (i.e. if it moves us away from zero), we say that we are overstating (or overestimating) the effect of X on Y. This happens when B1 and the bias have the same sign
Bias -
B1 -
over/under estimating bias?
overestimating
If the bias leads our regression coefficient to be larger in absolute value than it should be (i.e. if it moves us away from zero), we say that we are overstating (or overestimating) the effect of X on Y. This happens when B1 and the bias have the same sign
Bias -
B1 +
over/under estimating bias?
underestimating
If the bias leads our regression coefficient to be larger in absolute value than it should be (i.e. if it moves us away from zero), we say that we are overstating (or overestimating) the effect of X on Y. This happens when B1 and the bias have the same sign
Key Assumption 1
Xi: E(Ui|Xi) = 0
The conditional distribution of Ui given Xi has a mean of 0.
All factors contained in Ui are unrelated to Xi.
Key Assumption 2
(Xi, Yi) i = 1…..n. All are independent and identically distributed
No patterns in the data
Key Assumption 3
X and u have 4 finite movement.
No large outliers.
Key Assumption 4
no strong multicollinearity
linear-log
A 1% change in X is associated with a change in Y of .01*B1
log-linear
A 1 unit change in X is associated with a change in Y of 100%*B1
log-log
A 1% change in X is associated with a change in Y of B1%
odds
1 / 1-p
homoskedasticity vs. heteroskadisticity
variance of u|X = x is constant
variance of u|X =x is a function of x
logit(Pi)
ln (p/1-p) = nu
–> pi = 1 / 1+e^nu
