Interval Estimation in the Classical Normal Linear Regression Model. 1. Introduction

Similar documents
Lecture 4 Hypothesis Testing

Introduction to Dummy Variable Regressors. 1. An Example of Dummy Variable Regressors

ECON 351* -- Note 23: Tests for Coefficient Differences: Examples Introduction. Sample data: A random sample of 534 paid employees.

Tests of Single Linear Coefficient Restrictions: t-tests and F-tests. 1. Basic Rules. 2. Testing Single Linear Coefficient Restrictions

ECONOMICS 351*-A Mid-Term Exam -- Fall Term 2000 Page 1 of 13 pages. QUEEN'S UNIVERSITY AT KINGSTON Department of Economics

Biostatistics 360 F&t Tests and Intervals in Regression 1

Economics 130. Lecture 4 Simple Linear Regression Continued

STAT 3008 Applied Regression Analysis

a. (All your answers should be in the letter!

x i1 =1 for all i (the constant ).

[The following data appear in Wooldridge Q2.3.] The table below contains the ACT score and college GPA for eight college students.

Chapter 2 - The Simple Linear Regression Model S =0. e i is a random error. S β2 β. This is a minimization problem. Solution is a calculus exercise.

e i is a random error

Chapter 5: Hypothesis Tests, Confidence Intervals & Gauss-Markov Result

β0 + β1xi and want to estimate the unknown

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4)

Properties of Least Squares

Statistics for Economics & Business

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction

1. Inference on Regression Parameters a. Finding Mean, s.d and covariance amongst estimates. 2. Confidence Intervals and Working Hotelling Bands

Interpreting Slope Coefficients in Multiple Linear Regression Models: An Example

Chapter 13: Multiple Regression

Module Contact: Dr Susan Long, ECO Copyright of the University of East Anglia Version 1

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity

Department of Quantitative Methods & Information Systems. Time Series and Their Components QMIS 320. Chapter 6

Basic Business Statistics, 10/e

Chapter 11: Simple Linear Regression and Correlation

Marginal Effects of Explanatory Variables: Constant or Variable? 1. Constant Marginal Effects of Explanatory Variables: A Starting Point

Econ Statistical Properties of the OLS estimator. Sanjaya DeSilva

Statistics II Final Exam 26/6/18

Professor Chris Murray. Midterm Exam

β0 + β1xi. You are interested in estimating the unknown parameters β

Comparison of Regression Lines

Linear Regression Analysis: Terminology and Notation

Chapter 14 Simple Linear Regression

18. SIMPLE LINEAR REGRESSION III

Learning Objectives for Chapter 11

Question 1 carries a weight of 25%; question 2 carries 20%; question 3 carries 25%; and question 4 carries 30%.

Correlation and Regression

28. SIMPLE LINEAR REGRESSION III

Statistics for Managers Using Microsoft Excel/SPSS Chapter 13 The Simple Linear Regression Model and Correlation

β0 + β1xi. You are interested in estimating the unknown parameters β

Problem of Estimation. Ordinary Least Squares (OLS) Ordinary Least Squares Method. Basic Econometrics in Transportation. Bivariate Regression Analysis

where I = (n x n) diagonal identity matrix with diagonal elements = 1 and off-diagonal elements = 0; and σ 2 e = variance of (Y X).

# c i. INFERENCE FOR CONTRASTS (Chapter 4) It's unbiased: Recall: A contrast is a linear combination of effects with coefficients summing to zero:

17 - LINEAR REGRESSION II

Statistics for Business and Economics

Scatter Plot x

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

Negative Binomial Regression

Resource Allocation and Decision Analysis (ECON 8010) Spring 2014 Foundations of Regression Analysis

Now we relax this assumption and allow that the error variance depends on the independent variables, i.e., heteroskedasticity

ANSWERS CHAPTER 9. TIO 9.2: If the values are the same, the difference is 0, therefore the null hypothesis cannot be rejected.

ANSWERS. Problem 1. and the moment generating function (mgf) by. defined for any real t. Use this to show that E( U) var( U)

ECONOMICS 351* -- Stata 10 Tutorial 6. Stata 10 Tutorial 6

Introduction to Regression

Marginal Effects in Probit Models: Interpretation and Testing. 1. Interpreting Probit Coefficients

DO NOT OPEN THE QUESTION PAPER UNTIL INSTRUCTED TO DO SO BY THE CHIEF INVIGILATOR. Introductory Econometrics 1 hour 30 minutes

Reminder: Nested models. Lecture 9: Interactions, Quadratic terms and Splines. Effect Modification. Model 1

The Geometry of Logit and Probit

STATISTICS QUESTIONS. Step by Step Solutions.

CHAPTER 8. Exercise Solutions

Correlation and Regression. Correlation 9.1. Correlation. Chapter 9

x = , so that calculated

Diagnostics in Poisson Regression. Models - Residual Analysis

Econ107 Applied Econometrics Topic 9: Heteroskedasticity (Studenmund, Chapter 10)

Lecture 6: Introduction to Linear Regression

Here is the rationale: If X and y have a strong positive relationship to one another, then ( x x) will tend to be positive when ( y y)

The SAS program I used to obtain the analyses for my answers is given below.

Lecture Notes for STATISTICAL METHODS FOR BUSINESS II BMGT 212. Chapters 14, 15 & 16. Professor Ahmadi, Ph.D. Department of Management

x yi In chapter 14, we want to perform inference (i.e. calculate confidence intervals and perform tests of significance) in this setting.

Chapter 14 Simple Linear Regression Page 1. Introduction to regression analysis 14-2

Statistics MINITAB - Lab 2

Outline. 9. Heteroskedasticity Cross Sectional Analysis. Homoskedastic Case

The Ordinary Least Squares (OLS) Estimator

Statistics for Managers Using Microsoft Excel/SPSS Chapter 14 Multiple Regression Models

4 Analysis of Variance (ANOVA) 5 ANOVA. 5.1 Introduction. 5.2 Fixed Effects ANOVA

ECONOMETRICS - FINAL EXAM, 3rd YEAR (GECO & GADE)

More metrics on cartesian products

Systems of Equations (SUR, GMM, and 3SLS)

Statistical Inference. 2.3 Summary Statistics Measures of Center and Spread. parameters ( population characteristics )

Chapter 3. Two-Variable Regression Model: The Problem of Estimation

Y = β 0 + β 1 X 1 + β 2 X β k X k + ε

Lecture 9: Linear regression: centering, hypothesis testing, multiple covariates, and confounding

Predictive Analytics : QM901.1x Prof U Dinesh Kumar, IIMB. All Rights Reserved, Indian Institute of Management Bangalore

Lecture 9: Linear regression: centering, hypothesis testing, multiple covariates, and confounding

University of California at Berkeley Fall Introductory Applied Econometrics Final examination

If we apply least squares to the transformed data we obtain. which yields the generalized least squares estimator of β, i.e.,

experimenteel en correlationeel onderzoek

January Examinations 2015

Linear Approximation with Regularization and Moving Least Squares

Maximum Likelihood Estimation of Binary Dependent Variables Models: Probit and Logit. 1. General Formulation of Binary Dependent Variables Models

Answers Problem Set 2 Chem 314A Williamsen Spring 2000

Durban Watson for Testing the Lack-of-Fit of Polynomial Regression Models without Replications

First Year Examination Department of Statistics, University of Florida

STAT 3340 Assignment 1 solutions. 1. Find the equation of the line which passes through the points (1,1) and (4,5).

/ n ) are compared. The logic is: if the two

Lecture 3 Stat102, Spring 2007

LOGIT ANALYSIS. A.K. VASISHT Indian Agricultural Statistics Research Institute, Library Avenue, New Delhi

Statistics Chapter 4

Transcription:

ECONOMICS 35* -- NOTE 7 ECON 35* -- NOTE 7 Interval Estmaton n the Classcal Normal Lnear Regresson Model Ths note outlnes the basc elements of nterval estmaton n the Classcal Normal Lnear Regresson Model (the CNLRM). Interval estmaton --.e., the constructon of confdence ntervals for unknown populaton parameters -- s one of the two alternatve approaches to statstcal nference; the other s hypothess testng.. Introducton We have prevously derved pont estmators of all the unknown populaton parameters n the Classcal Normal Lnear Regresson Model (CNLRM) for whch the populaton regresson equaton, or PRE, s Y = β + β X + u where u s d as N(0, σ ) ( =,...,N). () The unknown parameters of the PRE are and () the regresson coeffcents β and β () the error varance σ. The pont estmators of these unknown populaton parameters are and () the unbased OLS regresson coeffcent estmators β ˆ and β ˆ () the unbased error varance estmator Assume that we have computed the pont estmates β $, β $, and σ $ of the unknown parameters for a gven set of sample data (Y, X ), =,..., N. ˆσ. ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page of 6 pages

ECONOMICS 35* -- NOTE 7 We therefore begn wth the followng OLS sample regresson equaton (or OLS-SRE): Y = β $ + β $ X + u $ ( =,...,N). () where $ ( )( ) = = xy X X Y Y β = x ( X X) β$ = Y β$ X = OLS estmate of β ; σ$ $ = u N = unbased OLS estmate of σ σˆ σˆ Vâr(ˆ β ) = = ; x ( X X) σˆ σˆ sê(ˆ β ) = Vâr(ˆ β ) = = x ; x Vâr(ˆ σˆ X β ) = = ; N ( ) x N X X σˆ X ) = Vâr(ˆ β) = N x σˆ X sê(ˆ β. OLS estmate of β ; Under the assumptons of the Classcal Normal Lnear Regresson Model (CNLRM) -- ncludng n partcular the normalty assumpton A9 -- the sample t-statstcs for ˆβ and β ˆ each have the t-dstrbuton wth (N ) degrees of freedom:.e., t( β$ $ β β β β ) Var $ ( β $ ) se$( β $ = = ) ~ t[n ]; t ( $ β$ $ β β β β ) Var $ ( β $ ) se$( β $ = = ) ~ t[n ]. ; ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page of 6 pages

ECONOMICS 35* -- NOTE 7. Interval Estmaton: Some Basc Ideas. General Form of a Confdence Interval A two-sded confdence nterval for the slope coeffcent β takes the general form Pr(β $ L β β $ ) = Pr(β $ U $δ β β $ + $δ ) = (3) where = the sgnfcance level (0 < < ), = the confdence level (or confdence coeffcent), δ $ = a postvely-valued sample statstc, β $ = β $ $δ L = the lower confdence lmt, β $ = β $ + $δ = the upper confdence lmt. U The nterval [β $, β $ ] = [β $ L U $δ, β $ + $δ ] s called the two-sded ( )-level confdence nterval, or two-sded 00( ) percent confdence nterval, for the slope coeffcent β. ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 3 of 6 pages

ECONOMICS 35* -- NOTE 7. Interpretaton of Confdence Intervals Pr(β $ L β β $ ) = Pr(β $ U $δ β β $ + $δ ) = (3). The confdence nterval [β $, β $ ] s a random nterval. L U The confdence lmts β $ = $β L δ $ and β $ = β $ U + $δ are random varables (or sample statstcs) that vary n value from one sample to another because the values of β $ and δ $ vary from sample to sample. But for any one sample of data and the correspondng estmates of β $ and $δ, the confdence lmts β $ = β $ $δ and β $ = β $ + $δ are smply fxed L U numbers,.e., they take fxed values. Therefore, any one confdence nterval calculated for a partcular sample of data s a fxed -- meanng nonrandom -- nterval.. The correct nterpretaton of the confdence nterval [β $, β $ L U ] s based on the concept of repeated samplng. In repeated samples of the same sze from the same populaton, 00( ) percent of the confdence ntervals constructed usng the formulas for ˆβ L and ˆβ wll contan the true value of the populaton parameter β. U For example, f the confdence level = 0.95, 95 percent of the confdence ntervals computed usng repeated samples of the same sze from the same populaton wll contan the true value of β. But any one confdence nterval for β, based on one sample of data, may or may not contan the true value of β. Snce the true value of β s unknown, we do not know whether that value does or does not le nsde any one confdence nterval. ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 4 of 6 pages

ECONOMICS 35* -- NOTE 7 3. Confdence Intervals for the Regresson Coeffcents β and β 3. Confdence Interval for β : Dervaton A two-step dervaton: Step : Start wth a probablty statement formulated n terms of t( ˆβ ), the t-statstc for β ˆ. Ths probablty statement mplctly defnes the two-sded ( )- level confdence nterval for β. Step : Re-arrange ths probablty statement to obtan an equvalent probablty statement formulated n terms of β rather than t(β ˆ ). The resultant probablty statement explctly defnes the two-sded ( )-level confdence nterval for β. Step : The two-sded ( )-level confdence nterval for β s mplctly defned by the probablty statement where ( t [N ] t(ˆ β ) t [N ] ) = Pr / (4) / = the confdence level attached to the confdence nterval; = the sgnfcance level, where 0 < < ; t / [N ] = the crtcal value of the t-dstrbuton wth N degrees of freedom at the / (or 00/ percent) sgnfcance level; and t(β $ ) s the t-statstc for β $ gven by ˆ β β β t(ˆ β ) = =. (5) Vâr(ˆ β ) sê(ˆ β ) ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 5 of 6 pages

ECONOMICS 35* -- NOTE 7 The t(β ˆ ) statstc has the t[n ] dstrbuton area = / area = area = / 0.45 0.4 0.35 0.3 0.5 0. 0.5 0. 0.05 0-5 -4-3 - - 0 3 4 5 t / t / Pr(t < t / ) = / Pr( t / t t / ) = Pr(t > t / ) = / ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 6 of 6 pages

ECONOMICS 35* -- NOTE 7 Step : Express the double nequalty nsde the brackets n probablty statement (4) n terms of β rather than t(β $ ). ( t [N ] t(ˆ β ) t [N ] ) = Pr / (4) / () Substtute n the double nequalty t [N ] t(ˆ β ) t / [N / the expresson for t(β $ ) gven n (5) above: ] β$ β t [ N ] t N / se$( β $ [ ]. (6.) / ) () Multply the double nequalty (6.) by the postve number se $( β $ ) > 0: ˆ / sê(ˆ β ) β β t / sê(ˆ β ). (6.) t (3) Subtract β $ from both sdes of nequalty (6.): β $ t se $( β $ ) β β $ + t se $( $ ). (6.3) β / / (4) Multply all terms n nequalty (6.3) by, rememberng to reverse the drecton of the nequaltes: β$ t se$( β $ ) β β$ t se $( $ ). (6.4) + β / / ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 7 of 6 pages

ECONOMICS 35* -- NOTE 7 RESULT: The probablty statement (4) can be wrtten as ( t [N ]sê(ˆ β ) β + t [N ]sê(ˆ β )) = Pr. (7) The two-sded ( )-level confdence nterval for β can therefore be wrtten as β$ [ ] $( $ ) $ [ ] $( $ t/ N se β β β + t N se β ) or more compactly as β$ ± t [ N ] se$( β $ ) or [ β ˆ t [N ] sê(ˆ ), β ˆ t [N ]sê(ˆ )] β + β where at the ( ) confdence level, or 00( ) percent confdence level, β$ $ [ ] $( $ L = β t N se β ) = the lower 00( ) percent confdence lmt for β and β$ $ [ ] $( $ U = β + t N se β ) = the upper 00( ) percent confdence lmt for β ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 8 of 6 pages

ECONOMICS 35* -- NOTE 7 Two-sded ( )-level confdence nterval for β s centered around β ˆ tal area = / area = tal area = / 0.45 0.4 0.35 0.3 0.5 0. 0.5 0. 0.05 0 β $ β ˆ β $ L U = t [N ]sê(ˆ ) = the lower ( )-level confdence lmt for β L β = + t [N ]sê(ˆ ) = the upper ( )-level confdence lmt for β U β ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 9 of 6 pages

ECONOMICS 35* -- NOTE 7 3. Confdence Interval for β : Dervaton The confdence nterval (or nterval estmator) for the ntercept coeffcent β s derved, nterpreted, and constructed n exactly the same way as the confdence nterval for the slope coeffcent β.. The two-sded ( )-level confdence nterval for β s mplctly defned by the probablty statement where ( [N ] t(ˆ β ) t [N ] ) = Pr (8) t / / = the confdence level attached to the confdence nterval; = the sgnfcance level, where 0 < < ; t / [N ] = the crtcal value of the t-dstrbuton wth (N ) degrees of freedom at the / (or 00(/) percent) sgnfcance level; and t(β $ ) s the t-statstc for β $ gven by t( β $ ) = β$ β Var $ ( β $ ) = β$ β. (9) se$( β $ ). The double nequalty nsde the brackets n probablty statement (8) can be expressed n terms of β rather than t(β $ ), usng a dervaton analogous to that used n dervng the confdence nterval for β. ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 0 of 6 pages

ECONOMICS 35* -- NOTE 7 RESULT: The probablty statement (8) can be wrtten as ( t [N ]sê(ˆ β ) β + t [N ]sê(ˆ β )) = Pr. (0) The two-sded ( )-level confdence nterval for β can therefore be wrtten as β$ [ ] $( $ ) $ [ ] $( $ t / N se β β β + t N se β ) or more compactly as β$ ± t [ N ] se$( β $ ) or [ β ˆ t [N ] sê(ˆ ), β ˆ t [N ]sê(ˆ )] β + β where at the ( ) confdence level, or 00( ) percent confdence level, β$ $ [ ] $( $ L = β t N se β ) = the lower 00( ) percent confdence lmt for β and β$ $ [ ] $( $ U = β + t N se β ) = the upper 00( ) percent confdence lmt for β ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page of 6 pages

ECONOMICS 35* -- NOTE 7 3.3 Procedure for Computng Confdence Intervals Consder the problem of computng a confdence nterval for the slope coeffcent β. Recall that the two-sded ( )-level confdence nterval for β s gven by the double nequalty β$ t [ N ] se$( β $ ) β β$ + t [ N ] se$( β $ ). Step : After estmatng the PRE () by OLS, retreve from the estmaton results the OLS estmate β $ of β and the estmated standard error se $( β $ ). Step : Select the value of the confdence level ( ), whch amounts to selectng the value of. Although the choce of confdence level s essentally arbtrary, the values most commonly used n practce are: = 0.0 ( ) = 0.99,.e., the 00( ) = 00(0.99) = 99 percent confdence level; = 0.05 ( ) = 0.95,.e., the 00( ) = 00(0.95) = 95 percent confdence level; = 0.0 ( ) = 0.90,.e., the 00( ) = 00(0.90) = 90 percent confdence level. Step 3: Obtan the value of t [ N ], the / crtcal value of the t-dstrbuton wth N degrees of freedom, ether from statstcal tables of the t- dstrbuton or from a computer software program. Step 4: Use the values of β $, se$( β $ ), and t [ N ] to compute the upper and lower 00( ) percent confdence lmts for β : = + t [N ]sê( ˆ ) = upper 00( )% confdence lmt for β ; U β = t [N ]sê( ˆ ) = lower 00( )% confdence lmt for β. L β ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page of 6 pages

ECONOMICS 35* -- NOTE 7 4. Determnants of the Confdence Intervals for β and β Consder for example the two-sded 00( )% confdence nterval for β : or β$ [ ] $( $ ) $ [ ] $( $ t N se β β β + t N se β ) [ β $ [ ] $( β$ ), β $ [ ] $( β$ t N se + t N se ) ] The two-sded confdence nterval for β s wder () the greater the value of se$( β $ ), the estmated standard error of β $,.e., the less precse s the estmate of β; () the greater the crtcal value [N ],.e., the greater the chosen value of t the confdence level ( ) for the gven sample sze N. Explanaton: Gven sample sze N, the value of t [N ] s negatvely related to the value of, and so s postvely related to the value of ( ). Example: Suppose sample sze N = 30, so that the degrees-of-freedom N- = 8. Then from a table of percentage ponts for the t-dstrbuton, we obtan the followng values of t [ N ] = t [ 8] for dfferent values of : = 0.0 ( ) = 0.99: / = 0.005 and t 0.005 [8] =.763; = 0.0 ( ) = 0.98: / = 0.0 and t 0.0 [8] =.467; = 0.05 ( ) = 0.95: / = 0.05 and t 0.05 [8] =.048; = 0.0 ( ) = 0.90: / = 0.05 and t 0.05 [8] =.70. Note that hgher values of ( ) --.e., hgher confdence levels -- correspond to hgher crtcal values of t [ 8]. ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 3 of 6 pages

ECONOMICS 35* -- NOTE 7 5. Two-Sded 00( )% Confdence Intervals for β : Examples Two-Sded 00( ) Percent Confdence Interval for β j : Formulas In general, the two-sded 00(-) percent confdence nterval for regresson coeffcent β j s: where [ t [N k]sê( ), + t [N k]sê( ) ] j j j = + t [N k]sê( ) = upper 00( ) % confdence lmt for β ju j j = t [N k]sê(ˆ β ) = lower 00( ) % confdence lmt for β jl j j j DATA: auto.dta A sample of 74 cars sold n North Amerca n 978. MODEL: prce = β + β weght + u ( =,..., N) N = 74 ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 4 of 6 pages

ECONOMICS 35* -- NOTE 7 Compute the two-sded 95% confdence nterval for β. regress prce weght Source SS df MS Number of obs = 74 ---------+------------------------------ F(, 7) = 9.4 Model 8433937 8433937 Prob > F = 0.0000 Resdual 45083459 7 66548.04 R-squared = 0.90 ---------+------------------------------ Adj R-squared = 0.80 Total 635065396 73 869955.97 Root MSE = 50.3 ------------------------------------------------------------------------------ prce Coef. Std. Err. t P> t [95% Conf. Interval] ---------+-------------------------------------------------------------------- weght.044063.376834 5.44 0.000.9858.79568 _cons -6.707353 74.43-0.006 0.995-347.89 334.475 ------------------------------------------------------------------------------. dsplay nvttal(7, 0.05).9934635 $β =.044 se$( β $ ) = 0.37683 N = 74 k = N k = 74 = 7 = 0.95 = 0.95 = 0.05 / = 0.05/ = 0.05 t / [N ] = t 0.05 [7] =.9935 t / [N ]s (ˆ β ) = t0.05[n ] ê(ˆ β ) =.9935(0.37683) = 0.75 ê s = + t [N ]sê( ˆ ) =.044 + 0.75 =.7953 =.795 U β = t [N ]sê( ˆ ) =.044 0.75 =.989 =.93 L β Result: The two-sded 95% confdence nterval for β s [.93,.795]. ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 5 of 6 pages

ECONOMICS 35* -- NOTE 7 Compute the two-sded 99% confdence nterval for β. regress prce weght, level(99) Source SS df MS Number of obs = 74 -------------+------------------------------ F(, 7) = 9.4 Model 8433937 8433937 Prob > F = 0.0000 Resdual 45083459 7 66548.04 R-squared = 0.90 -------------+------------------------------ Adj R-squared = 0.80 Total 635065396 73 869955.97 Root MSE = 50.3 ------------------------------------------------------------------------------ prce Coef. Std. Err. t P> t [99% Conf. Interval] -------------+---------------------------------------------------------------- weght.044063.376834 5.4 0.000.04705 3.04 _cons -6.707353 74.43-0.0 0.995-34.074 300.659 ------------------------------------------------------------------------------. dsplay nvttal(7, 0.005).645859. scalar bu99 = _b[weght] +.6459*_se[weght]. scalar bl99 = _b[weght] -.6459*_se[weght]. scalar lst bu99 bl99 bu99 = 3.048 bl99 =.046997 $β =.044 se$( β $ ) = 0.37683 N = 74 k = N k = 74 = 7 = 0.99 = 0.99 = 0.0 / = 0.0/ = 0.005 t / [N ] = t 0.005 [7] =.6459 t / [N ]s (ˆ β ) = t0.005[n ] ê(ˆ β ) =.6459(0.37683) = 0.99705 ê s β ˆ = + t [N ]sê( ˆ ) =.044 + 0. 99705 = 3.045 = 3.04 U 0.005 β β ˆ = t [N ]sê( ˆ ) =.044 0. 99705 =.04705 =.047 L 0.005 β Result: The two-sded 99% confdence nterval for β s [.047, 3.04]. ECON 35* -- Note 7: Interval Estmaton n the CNLRM Page 6 of 6 pages