Chapter 6

Linear Regression
with Multiple
Regressors

© Pearson Education Limited 2015

Outline

1. Omitted variable bias

2. Causality and regression analysis
3. Multiple regression and OLS
4. Measures of fit
5. Sampling distribution of the OLS estimator

© Pearson Education Limited 2015

6-2
Omitted Variable Bias
(SW Section 6.1)

The error u arises because of factors, or

variables, that influence Y but are not
included in the regression function. There
are always omitted variables.
 
Sometimes, the omission of those variables
can lead to bias in the OLS estimator.

© Pearson Education Limited 2015

6-3
Omitted variable bias, ctd.

The bias in the OLS estimator that occurs as a result

of an omitted factor, or variable, is called omitted
variable bias. For omitted variable bias to occur, the
omitted variable “Z” must satisfy two conditions:
The two conditions for omitted variable bias
1. Z is a determinant of Y (i.e. Z is part of u); and
2. Z is correlated with the regressor X (i.e. corr(Z,X)
≠ 0)
 
Both conditions must hold for the omission of Z to
result in omitted variable bias.

© Pearson Education Limited 2015

6-4
Omitted variable bias, ctd.

In the test score example:

1. English language ability (whether the student has
English as a second language) plausibly affects
standardized test scores: Z is a determinant of
Y.
2. Immigrant communities tend to be less affluent
and thus have smaller school budgets and higher
STR: Z is correlated with X.
Accordingly, is biased. What is the direction of
this bias?
– What does common sense suggest?
– If common sense fails you, there is a formula…
© Pearson Education Limited 2015
6-5
Omitted variable bias, ctd.

A formula for omitted variable bias: recall the

equation,
1 n

n i1
vi
– β1 = =
 n  1 2
 n  s X
 
where vi = (Xi – X )ui ≈ (Xi – μX)ui. Under Least
Squares Assumption #1,
E[(Xi – μX)ui] = cov(Xi,ui) = 0.
 
But what if E[(Xi – μX)ui] = cov(Xi,ui) = σXu ≠ 0?
© Pearson Education Limited 2015
6-6
Omitted variable bias, ctd.

Under LSA #2 and #3 (that is, even if LSA #1 is not true),

1 n

n i1
( X i  X )u i
– β1 =
1 n

n i1 i
( X  X ) 2

 Xu
p
 2
X
  u    Xu   u 
=
     =
    Xu
,
   X   X u  X
where ρXu = corr(X,u). If assumption #1 is correct, then ρXu =
0, but if not we have….
© Pearson Education Limited 2015
6-7
The omitted variable bias formula:
p  u 
 1 +     Xu
X

• If an omitted variable Z is both:

1. a determinant of Y (that is, it is contained in u); and
2. correlated with X,
then ρXu ≠ 0 and the OLS estimator is biased and is not
consistent.
•  
• For example, districts with few ESL students (1) do better on
standardized tests and (2) have smaller classes (bigger
budgets), so ignoring the effect of having many ESL
students factor would result in overstating the class size
effect. Is this is actually going on in the CA data?

© Pearson Education Limited 2015

6-8
• Districts with fewer English Learners have higher test scores
• Districts with lower percent EL (PctEL) have smaller classes
• Among districts with comparable PctEL, the effect of class
size is small (recall overall “test score gap” = 7.4)
© Pearson Education Limited 2015
6-9
Causality and regression analysis

• The test score/STR/fraction English Learners

example shows that, if an omitted variable
satisfies the two conditions for omitted variable
bias, then the OLS estimator in the regression
omitting that variable is biased and inconsistent.
So, even if n is large, ̂1 will not be close to β1.

• This raises a deeper question: how do we define

β1? That is, what precisely do we want to estimate
when we run a regression?

© Pearson Education Limited 2015

6-10
What precisely do we want to estimate
when we run a regression?

There are (at least) three possible answers to this

question:
1. We want to estimate the slope of a line through a
scatterplot as a simple summary of the data to
which we attach no substantive meaning.
This can be useful at times, but isn’t very interesting
intellectually and isn’t what this course is about.

© Pearson Education Limited 2015

6-11
2. We want to make forecasts, or predictions, of the
value of Y for an entity not in the data set, for
which we know the value of X.

Forecasting is an important job for economists,

and excellent forecasts are possible using
regression methods without needing to know
causal effects. We will return to forecasting later
in the course.

© Pearson Education Limited 2015

6-12
3. We want to estimate the causal effect on Y of a
change in X.

This is why we are interested in the class size

effect. Suppose the school board decided to cut
class size by 2 students per class. What would be
the effect on test scores? This is a causal question
(what is the causal effect on test scores of STR?)
so we need to estimate this causal effect. Except
when we discuss forecasting, the aim of this
course is the estimation of causal effects using
regression methods.

© Pearson Education Limited 2015

6-13
What, precisely, is a causal effect?

• “Causality” is a complex concept!

• In this course, we take a practical approach

to defining causality:
 
A causal effect is defined to be the
effect measured in an ideal
randomized controlled experiment.

© Pearson Education Limited 2015

5-14
Ideal Randomized Controlled Experiment

• Ideal: subjects all follow the treatment protocol –

perfect compliance, no errors in reporting, etc.!
• Randomized: subjects from the population of
interest are randomly assigned to a treatment or
control group (so there are no confounding
factors)
• Controlled: having a control group permits
measuring the differential effect of the treatment
• Experiment: the treatment is assigned as part of
the experiment: the subjects have no choice, so
there is no “reverse causality” in which subjects
choose the treatment they think will work best.
© Pearson Education Limited 2015
6-15
Back to class size:

Imagine an ideal randomized controlled experiment for

measuring the effect on Test Score of reducing STR…

• In that experiment, students would be randomly assigned to

classes, which would have different sizes.

• Because they are randomly assigned, all student

characteristics (and thus ui) would be distributed
independently of STRi.

• Thus, E(ui|STRi) = 0 – that is, LSA #1 holds in a randomized

controlled experiment.

© Pearson Education Limited 2015

6-16
How does our observational data differ
from this ideal?

• The treatment is not randomly assigned

• Consider PctEL – percent English learners – in the
district. It plausibly satisfies the two criteria for
omitted variable bias: Z = PctEL is:
1. a determinant of Y; and
2. correlated with the regressor X.
• Thus, the “control” and “treatment” groups differ
in a systematic way, so corr(STR,PctEL) ≠ 0

© Pearson Education Limited 2015

6-17
• Randomization + control group means that any
differences between the treatment and control
groups are random – not systematically related to
the treatment
• We can eliminate the difference in PctEL between
the large (control) and small (treatment) groups
by examining the effect of class size among
districts with the same PctEL.
– If the only systematic difference between the large and
small class size groups is in PctEL, then we are back to
the randomized controlled experiment – within each
PctEL group.
– This is one way to “control” for the effect of PctEL when
estimating the effect of STR.

© Pearson Education Limited 2015

6-18
Return to omitted variable bias

Three ways to overcome omitted variable bias

1. Run a randomized controlled experiment in which treatment
(STR) is randomly assigned: then PctEL is still a
determinant of TestScore, but PctEL is uncorrelated with
STR. (This solution to OV bias is rarely feasible.)
2. Adopt the “cross tabulation” approach, with finer gradations
of STR and PctEL – within each group, all classes have the
same PctEL, so we control for PctEL (But soon you will run
out of data, and what about other determinants like family
income and parental education?)
3. Use a regression in which the omitted variable (PctEL) is no
longer omitted: include PctEL as an additional regressor in a
multiple regression.

© Pearson Education Limited 2015

6-19
The Population Multiple Regression Model
(SW Section 6.2)
• Consider the case of two regressors:
Yi = β0 + β1X1i + β2X2i + ui, i = 1,…,n
• Y is the dependent variable
• X1, X2 are the two independent variables (regressors)
• (Yi, X1i, X2i) denote the ith observation on Y, X1, and X2.
• β0 = unknown population intercept
• β1 = effect on Y of a change in X1, holding X2 constant
• β2 = effect on Y of a change in X2, holding X1 constant
• ui = the regression error (omitted factors)

© Pearson Education Limited 2015

6-20
Interpretation of coefficients in multiple
regression

Yi = β0 + β1X1i + β2X2i + ui, i = 1,…,n

Consider changing X1 by ΔX1 while holding X2
constant:
Population regression line before the change:
 
Y = β0 + β1X1 + β2X2
 
Population regression line, after the change:
 
Y + ΔY = β0 + β1(X1 + ΔX1) + β2X2

© Pearson Education Limited 2015

6-21
Before: Y = β0 + β1(X1 + ΔX1) + 2X2
 
After: Y + ΔY = β0 + β1(X1 + ΔX1) + β2X2
 
Difference: ΔY = β1ΔX1
So:
Y
β1 =
X 1 , holding X2 constant
 
Y
β2 = , holding X1 constant
X 2

 
β0 = predicted value of Y when X1 = X2 = 0.

© Pearson Education Limited 2015

6-22
The OLS Estimator in Multiple Regression
(SW Section 6.3)

• With two regressors, the OLS estimator solves:

n
min b ,b ,b
0 1 2
 i 0 1 1i 2 2i
[Y  (b  b X  b X )]2

i1

• The OLS estimator minimizes the average squared

difference between the actual values of Yi and the
prediction (predicted value) based on the
estimated line.
• This minimization problem is solved using calculus
• This yields the OLS estimators of β0 and β1 .

© Pearson Education Limited 2015

6-23
Example: the California test score data

Regression of TestScore against STR:

 
Test Score = 698.9 – 2.28×STR
 
Now include percent English Learners in the district
(PctEL):
  Test Score = 686.0 – 1.10×STR – 0.65PctEL
•  What happens to the coefficient on STR?
• What (STR, PctEL) = 0.19)

© Pearson Education Limited 2015

6-24
Multiple regression in STATA

reg testscr str pctel, robust;

 
Regression with robust standard errors Number of obs = 420
F( 2, 417) = 223.82
Prob > F = 0.0000
R-squared = 0.4264
Root MSE = 14.464
 
------------------------------------------------------------------------------
| Robust
testscr | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
str | -1.101296 .4328472 -2.54 0.011 -1.95213 -.2504616
pctel | -.6497768 .0310318 -20.94 0.000 -.710775 -.5887786
_cons | 686.0322 8.728224 78.60 0.000 668.8754 703.189
------------------------------------------------------------------------------
 

Test Score = 686.0 – 1.10×STR – 0.65PctEL

More on this printout later…

© Pearson Education Limited 2015

6-25
Measures of Fit for Multiple Regression
(SW Section 6.4)

Actual = predicted + residual: Yi = Yˆ + uˆi

i
 
SER = std. deviation of uˆi (with d.f. correction)
 
RMSE = std. deviation of uˆi (without d.f. correction)
 
R2 = fraction of variance of Y explained by X
 
R 2 = “adjusted R2” = R2 with a degrees-of-freedom correction
that adjusts for estimation uncertainty; R 2 < R2

© Pearson Education Limited 2015

6-26
SER and RMSE

As in regression with a single regressor, the

SER and the RMSE are measures of the
spread of the Ys around the regression line:
n
1
SER = 
n  k  1 i 1
ˆ
ui
2

 
1 n 2
RMSE = 
n i 1
uˆi

© Pearson Education Limited 2015

6-27
R2 and R 2 (adjusted R2)

The R2 is the fraction of the variance explained –

same definition as in regression with a single
regressor:
ESS SSR
R = 2
= 1 ,
TSS TSS
n n n
where ESS =  (Yi  Y )
ˆ ˆ 2
, SSR =  ˆ
u
i 1
, TSS
2
i =  (Yi

i1
Y ) 2
.
i 1

 The R2 always increases when you add another regressor

(why?) – a bit of a problem for a measure of “fit”

© Pearson Education Limited 2015

6-28
R2 and R 2 ctd.

The R 2 (the “adjusted R2”) corrects this problem by

“penalizing” you for including another regressor –
the R 2 does not necessarily increase when you add
another regressor.
 
 n  1  SSR
Adjusted R : R = 1 
2 2
 TSS
   n  k  1
Note that R 2 < R2, however if n is large the two will
be very close.

© Pearson Education Limited 2015

6-29
Measures of fit, ctd.

Test score example:

 
(1) Test Score = 698.9 – 2.28×STR,
R2 = .05, SER = 18.6
 
(2) Test Score = 686.0 – 1.10×STR – 0.65PctEL,
R2 = .426, R 2 = .424, SER = 14.5
 
• What – precisely – does this tell you about the fit of regression
(2) compared with regression (1)?
• Why are the R2 and the 2 so close in (2)?
R

© Pearson Education Limited 2015

6-30
The Least Squares Assumptions for
Multiple Regression (SW Section 6.5)

Yi = β0 + β1X1i + β2X2i + … + βkXki + ui, i = 1,…,n

 
1. The conditional distribution of u given the X’s has mean
zero, that is, E(ui|X1i = x1,…, Xki = xk) = 0.

2. (X1i,…,Xki,Yi), i =1,…,n, are i.i.d.

3. Large outliers are

4 unlikely: X1,…,
4 Xk, and Y 4have four
X
moments: E( 1i ) < ∞,…, E(X ki ) < ∞, E(Yi ) < ∞.

4. There is no perfect multicollinearity.

© Pearson Education Limited 2015

6-31
Assumption #1: the conditional mean of u
given the included Xs is zero.
E(u|X1 = x1,…, Xk = xk) = 0
 
• This has the same interpretation as in regression with a
single regressor.
• Failure of this condition leads to omitted variable bias,
specifically, if an omitted variable
1. belongs in the equation (so is in u) and
2. is correlated with an included X
• then this condition fails and there is OV bias.
• The best solution, if possible, is to include the omitted
variable in the regression.
• A second, related solution is to include a variable that
controls for the omitted variable (discussed in Ch. 7)
© Pearson Education Limited 2015
6-32
Assumption #2: (X1i,…,Xki,Yi), i =1,…,n, are i.i.d.
This is satisfied automatically if the data are collected
by simple random sampling.

Assumption #3: large outliers are rare (finite

fourth moments)
This is the same assumption as we had before for a
single regressor. As in the case of a single regressor,
OLS can be sensitive to large outliers, so you need to
check your data (scatterplots!) to make sure there are
no crazy values (typos or coding errors).

© Pearson Education Limited 2015

6-33
Assumption #4: There is no perfect multicollinearity
Perfect multicollinearity is when one of the regressors is an
exact linear function of the other regressors.

Example: Suppose you accidentally include STR twice:

regress testscr str str, robust
Regression with robust standard errors Number of obs = 420
F( 1, 418) = 19.26
Prob > F = 0.0000
R-squared = 0.0512
Root MSE = 18.581
-------------------------------------------------------------------------
| Robust
testscr | Coef. Std. Err. t P>|t| [95% Conf. Interval]
--------+----------------------------------------------------------------
str | -2.279808 .5194892 -4.39 0.000 -3.300945 -1.258671
str | (dropped)
_cons | 698.933 10.36436 67.44 0.000 678.5602 719.3057
-------------------------------------------------------------------------

© Pearson Education Limited 2015

6-34
Perfect multicollinearity is when one of the
regressors is an exact linear function of the
other regressors.
• In the previous regression, β1 is the effect on
TestScore of a unit change in STR, holding STR
constant (???)
• We will return to perfect (and imperfect)
multicollinearity shortly, with more examples…
•  
• With these least squares assumptions in hand,
we now can derive the sampling distribution of
̂1 , ̂2 ,…, ˆ k .

© Pearson Education Limited 2015

6-35
The Sampling Distribution of the OLS
Estimator (SW Section 6.6)

Under the four Least Squares Assumptions,

• The sampling distribution of ̂1 has mean β1
• var( ̂1 ) is inversely proportional to n.
• Other than its mean and variance, the exact
(finite-n) distribution of ̂1 is very complicated; but
for large n… p
– ̂1 is consistent:̂1  β1 (law of large numbers)

ˆ1  E (ˆ1 )
– is approximately distributed N(0,1) (CLT)
ˆ
var( ) 1
– These statements hold for ̂ ,…,ˆ k
1
Conceptually, there is nothing new here!
© Pearson Education Limited 2015
6-36
Multicollinearity, Perfect and Imperfect
(SW Section 6.7)

Perfect multicollinearity is when one of the regressors is an

exact linear function of the other regressors.
 
Some more examples of perfect multicollinearity
1. The example from before: you include STR twice,
2. Regress TestScore on a constant, D, and B, where: Di = 1
if STR ≤ 20, = 0 otherwise; Bi = 1 if STR >20, = 0
otherwise, so Bi = 1 – Di and there is perfect
multicollinearity.
3. Would there be perfect multicollinearity if the intercept
(constant) were excluded from this regression? This
example is a special case of…

© Pearson Education Limited 2015

6-37
The dummy variable trap

Suppose you have a set of multiple binary (dummy) variables, which

are mutually exclusive and exhaustive – that is, there are multiple
categories and every observation falls in one and only one category
(Freshmen, Sophomores, Juniors, Seniors, Other). If you include all
these dummy variables and a constant, you will have perfect
multicollinearity – this is sometimes called the dummy variable
trap.
• Why is there perfect multicollinearity here?
• Solutions to the dummy variable trap:
1. Omit one of the groups (e.g. Senior), or
2. Omit the intercept
• What are the implications of (1) or (2) for the interpretation of the
coefficients?

© Pearson Education Limited 2015

6-38
Perfect multicollinearity, ctd.

• Perfect multicollinearity usually reflects a mistake

in the definitions of the regressors, or an oddity in
the data
• If you have perfect multicollinearity, your
statistical software will let you know – either by
crashing or giving an error message or by
“dropping” one of the variables arbitrarily
• The solution to perfect multicollinearity is to
modify your list of regressors so that you no
longer have perfect multicollinearity.

© Pearson Education Limited 2015

6-39
Imperfect multicollinearity

Imperfect and perfect multicollinearity are quite different despite the

similarity of the names.
 
Imperfect multicollinearity occurs when two or more regressors are
very highly correlated.
• Why the term “multicollinearity”? If two regressors are very highly
correlated, then their scatterplot will pretty much look like a
straight line – they are “co-linear” – but unless the correlation is
exactly ±1, that collinearity is imperfect.

© Pearson Education Limited 2015

6-40
Imperfect multicollinearity, ctd.

Imperfect multicollinearity implies that one or more of the

regression coefficients will be imprecisely estimated.
• The idea: the coefficient on X1 is the effect of X1 holding X2
constant; but if X1 and X2 are highly correlated, there is very
little variation in X1 once X2 is held constant – so the data
don’t contain much information about what happens when X1
changes but X2 doesn’t. If so, the variance of the OLS
estimator of the coefficient on X1 will be large.

• Imperfect multicollinearity (correctly) results in large

standard errors for one or more of the OLS coefficients.
• The math? See SW, App. 6.2

 Next topic: hypothesis tests and confidence intervals…

© Pearson Education Limited 2015
6-41