Scribd
Upload
28 views4 pages

Multiple Regression Analysis With Qualitative Information

Uploaded by

selena.shi.jw
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
28 views4 pages

Multiple Regression Analysis With Qualitative Information

Uploaded by

selena.shi.jw
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 4

BECO2001 Econometrics I, 2024/2025-1

MULTIPLE REGRESSION ANALYSIS: QUALITATIVE INFORMATION


(Textbook: Chapter 7, Sections 7.1, Section 7.2, Section 7.3, Section 7.4)

The dependent variable can be influenced by qualitative regressors, such as gender


(male, female) and geographical region (Asia, America, …). A qualitative variable is
indicated by the presence or absence of an attribute, quantified by the construction of

presence of attribute
a dummy 
D 1, absence of attribute
variable:

0,

that classifies data into mutually exclusive categories.

 D = 0 defines the reference/base/benchmark category.

 For a qualitative variable with M categories, introduce only (M - 1) dummy


variables in regression with an intercept term to avoid falling into the dummy
variable trap, that is, the situation of perfect multicollinearity.

Intercept Dummy Variable


An intercept dummy variable modifies the intercept parameter of the econometric
model, and represents an intercept shift of the PRF.

Consider the model of hourly wage determination:


𝑤𝑎𝑔𝑒𝑖 = 𝛽0 + 𝛿0𝑓𝑒𝑚𝑎𝑙𝑒𝑖 + 𝛽1𝑒𝑑𝑢𝑐𝑖 +
𝑢𝑖, where femalei = 0 when the person is a male, and
femalei = 0 when the person is a female

𝐸(𝑤𝑎𝑔𝑒𝑖|𝑓𝑒𝑚𝑎𝑙𝑒𝑖 = 0, 𝑒𝑑𝑢𝑐𝑖) = 𝛽0 + 𝛽1𝑒𝑑𝑢𝑐𝑖


𝐸(𝑤𝑎𝑔𝑒𝑖|𝑓𝑒𝑚𝑎𝑙𝑒𝑖 = 1, 𝑒𝑑𝑢𝑐𝑖) = (𝛽0 + 𝛿0) + 𝛽1𝑒𝑑𝑢𝑐𝑖
[male]
[female]

Male is the benchmark category. Female is the intercept dummy variable, and the
coefficient 𝛿0 attached to the dummy variable is called the differential intercept
coefficient. Educ is the quantitative regressor, called a covariate or a control
variable.

𝛿0 = 𝐸(𝑤𝑎𝑔𝑒𝑖|𝑓𝑒𝑚𝑎𝑙𝑒𝑖 = 1, 𝑒𝑑𝑢𝑐𝑖) − 𝐸(𝑤𝑎𝑔𝑒𝑖|𝑓𝑒𝑚𝑎𝑙𝑒𝑖 = 0, 𝑒𝑑𝑢𝑐𝑖)


If 𝛿0 is negative, then women earn less than men on average holding education fixed,
and there is discrimination against women in the labor market.
1
Dummy Variables for Multiple Categories
Four categories: married men, married women, single men (base), single women
Dummy variables: marrmale, marrfem, singfem

Interactions between Dummy Variables

2
Slope Dummy Variable
A slope dummy variable (the product of a dummy variable and a quantitative
variable) modifies the slope parameter of the econometric model, and represents a
slope shift of the PRF.

log(𝑤𝑎𝑔𝑒𝑖) = 𝛽0 + 𝛿0𝑓𝑒𝑚𝑎𝑙𝑒𝑖 + 𝛽1𝑒𝑑𝑢𝑐𝑖 + 𝛿1𝑓𝑒𝑚𝑎𝑙𝑒𝑖 ∗ 𝑒𝑑𝑢𝑐𝑖


Consider the extended model of hourly wage determination:

+ 𝑢𝑖,

𝐸(𝑤𝑎𝑔𝑒𝑖|𝑓𝑒𝑚𝑎𝑙𝑒𝑖 = 0, 𝑒𝑑𝑢𝑐𝑖) = 𝛽0 + 𝛽1𝑒𝑑𝑢𝑐𝑖


𝐸(𝑤𝑎𝑔𝑒𝑖|𝑓𝑒𝑚𝑎𝑙𝑒𝑖 = 1, 𝑒𝑑𝑢𝑐𝑖) = (𝛽0 + 𝛿0) + (𝛽1 + 𝛿1)𝑒𝑑𝑢𝑐𝑖
[male]

[female]

Female*educ is the slope dummy variable, and the attached coefficient 𝛿1 is called
the differential slope coefficient.

3
Hypothesis testing
𝐻0: 𝛿0 = 0 : Women earn the same level of wage as men [t-test]
𝐻0: 𝛿1 = 0 : Return to education is the same for men and women [t-test]
𝐻0: 𝛿0 = 0 and 𝛿1 = 0 : Average wages are the same for men and women [F-
test]

Example: Wage equation

You might also like