Scribd
Upload
41 views8 pages

Econometrics: Amado Peiró (Universitat de València)

This document discusses hypothesis testing in econometrics models. It introduces different types of hypothesis tests that can be performed on econometrics models, including general tests on linear restrictions with multiple parameters, tests on single linear restrictions with one parameter using t-statistics, tests of the significance of an overall model, and tests for structural breaks using the Chow test. The tests are used to evaluate restrictions on parameters in an econometrics model and make conclusions about whether to reject or fail to reject the null hypothesis based on t-statistics, F-statistics, or p-values.

Uploaded by

kati
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
41 views8 pages

Econometrics: Amado Peiró (Universitat de València)

This document discusses hypothesis testing in econometrics models. It introduces different types of hypothesis tests that can be performed on econometrics models, including general tests on linear restrictions with multiple parameters, tests on single linear restrictions with one parameter using t-statistics, tests of the significance of an overall model, and tests for structural breaks using the Chow test. The tests are used to evaluate restrictions on parameters in an econometrics model and make conclusions about whether to reject or fail to reject the null hypothesis based on t-statistics, F-statistics, or p-values.

Uploaded by

kati
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 8

ECONOMETRICS

UNIT 6

HYPOTHESIS TESTING

Amado Peiró (Universitat de València)

1
6.1 Introduction

Once the econometric model has been specified and estimated,


very different questions can be asked under the light of this model.
These questions are frequently formulated as hypothesis tests.
In what follows, we will see the most important tests in a
regression model. In all these tests, the null hypothesis consists in
one or more linear restrictions about the parameters in our model,
and they will be done with a Student’s t-statistic or with a F-
statistic.

2
6.2 General tests on any number of linear restrictions

The null hypothesis includes linear restrictions, with :

3
6.3 Test on one single linear restriction with one single
parameter

Tests on one single linear restriction may be done, of course, with an


statistic, but they are usually done with t-statistics:

4
6.4 Test of the significance of the model
Let us consider the model . To test the joint
significance of the model:

▶ The model IS NOT significant

▶ The model IS significant


5
6.5 Test for structural break (Chow test)

Let us consider two samples from two models with the same variables but
possibly different coefficients,

I: Sample size:
II: Sample size:
ALL: Sample size:

▶ No Break

▶ Break

6
Appendix 1

true true
Accept  ✓ Type 2 error =
Reject Type 1 error = ✓

In the classical way, to carry out a statistical test:


i) we state the hypotheses
ii) we obtain the test statistic under the null hypothesis
iii)we compare our statistic with its theoretical distribution under the null

Nowadays many usual tests are also carried out by using P-values, as it is
explained in what follows.

7
Appendix 2

For a test statistic, the P-value is the probability of


obtaining a statistic so extreme, when the null hypothesis is
true. If is our significance level, the rule of decision that we
should apply with F-tests or two-tail t-tests is:

P-value >  is NOT rejected


P-value <  is rejected
Evidently, the P-value is comprised between 0 and 1.
Loosely speaking, low values of the P-value point to
evidence against the null, and very low values point to
strong evidence against the null.

You might also like