and Statistics
Moshe Buchinsky
UCLA
Course Logistics
I Changsu Ko - Head TA
I Changsu Ko (head TA)
I Yun Feng
I Dong Ook Eun
I Kun Hu
I Jeonghwan Kim
I Lu Liu
I Pavel Andreyanov
Textbooks and Other Material
I Econometrics...
I Uses statistical methods to analyze economic data.
I aims to answer quantitative questions.
I Main tool: regression analysis
I We want to determine the causal effect of one variable
(X ) on another variable (Y ).
I Econ 41: statistical analysis of one variable.
I Econ 103: analysis of the relationship between two (or more)
variables.
Examples of questions of interest:
I Theory:
I How to interpret results in articles, or the results your
computer gives.
I To understand the caveats of regression analysis.
I Why it is essential to do good empirical work.
I How to do empirical analysis yourself:
I You will work with real datasets.
I You will use statistical software, namely STATA.
Review
I Moments (continued):
I Variance
X2 = Var (X ) = E (X X )2
P 2
i (xi X ) f (xi ) if X discrete
=
(x X )2 f (x)dx if X cont.
R
fX ,Y (x, y ) = Pr(X = x, Y = y ).
fY |X (y | x) = Pr(Y = y | X = x)
Pr(X = x, Y = y )
=
Pr(X = x)
fX ,Y (x, y )
=
fX (x)
2. Definitions - Two Random Variables
I Example: Men say they will vote for the Republican candidate
rather than the Democratic candidate in their districts by a margin
of 45 percent to 32 percent. The numbers are nearly reversed for
women, with 36 percent saying they will vote Republican and 43
percent saying they will vote Democratic. New York Times,
September 20, 2010
I Assume that there are 50% men and 50% women.
1 Democrat
1 male
X = Y = 2 Republican
2 female
3 other
2. Definitions - Two Random Variables
x1 = 1 x2 = 2
y1 = 1 .16 .215
y2 = 2 .225 .18
y3 = 3 .115 .105
2. Definitions - Two Random Variables
1 1
2. Definitions - Two Random Variables
I Independence:
fX ,Y (x, y ) = fX (x)fY (y )
I If two random variables are independent, then the conditional
distribution of each variable coincides with its marginal
distribution, that is
fY |X (y | x) = fY (y ), and
fX |Y (x | y ) = fX (x).
Properties of Expectations, Variance and Covariance
I E (a) = a
I E (aX + b) = aE (X ) + b
I E (X + Y ) = E (X ) + E (Y )
I Var (a) = 0
I Var (aX + b) = a2 Var (X )
I Cov (X , X ) = Var (X )Cov (X , Y ) = E (XY ) E (X )E (Y )
I Var (aX + bY ) = a2 Var (X ) + b 2 Var (Y ) + 2abCov (X , Y )
Properties of Expectations, Variance and Covariance
Z
p tm
W /m
1 0.8944 = 0.1056!