Fox Module 13 Dummy variable regression


Fox Module 13 Dummy variable regression

Author
Message
NEAS
Supreme Being
Supreme Being (5.9K reputation)Supreme Being (5.9K reputation)Supreme Being (5.9K reputation)Supreme Being (5.9K reputation)Supreme Being (5.9K reputation)Supreme Being (5.9K reputation)Supreme Being (5.9K reputation)Supreme Being (5.9K reputation)Supreme Being (5.9K reputation)

Group: Administrators
Posts: 4.3K, Visits: 1.3K

Fox Module 13 Dummy variable regression

 


           Dichotomous factors

           Polytomous factors


 

 

Read Section 7.1, “Dichotomous factors,” on pages 120-124. Insurance class ratemaking uses dichotomous and polytomous factors more than quantitative explanatory variables, so the on-line course stresses this chapter of the textbook.

 

Illustration: Sex (male vs female) is a dichotomous factor. Age group and territory are polytomous factors.

 

Graph 7.1 on page 121 shows how omission of a dichotomous factor distorts a regression line.

 

Know equation 7.1 on page 121. We use this equation for analysis of variance as well.

 

The equations at the bottom of page 123 show how dichotomous factors affect the slope coefficient. The final exam tests the use of these factors in the regression equations.

 

Read Section 7.1, “Polytomous factors,” on pages 124-129. The pattern is the same as for dichotomous factors. Know equations 7.2 and 7.3 on the bottom of page 125.

 

F-tests have a null hypothesis that the coefficients are equal, not that they are zero. See equation 7.4 on the bottom of page 126 and the paragraph at the top of page 127. Know equations 7.5 and 7.6 on page 127.

 

Know the example on page 128-129. The example puts the pieces together, making the logic easier to follow.

 

Natural science studies use quantitative explanatory variables. Social science and actuarial work use factors, such as sex, smoking, marital status, territory, and type of vehicle.

 

Final exam questions may ask for the number of dummy variables and the relation of means and regression coefficients. The practice problems show the types of questions.

 


Attachments
CalLadyQED
Forum Guru
Forum Guru (66 reputation)Forum Guru (66 reputation)Forum Guru (66 reputation)Forum Guru (66 reputation)Forum Guru (66 reputation)Forum Guru (66 reputation)Forum Guru (66 reputation)Forum Guru (66 reputation)Forum Guru (66 reputation)

Group: Forum Members
Posts: 62, Visits: 2
Why is the null hypothesis really "that the coefficients are equal, not that they are zero"? Fox seems to be saying that the null hypothesis for the F-test is that gamma1 = gamma2 = ... = gamma(m-1) = 0. (Page 126, equation 7.4; and again on page 128.) Can you explain more why this is not true?
Michelle2010
Junior Member
Junior Member (19 reputation)Junior Member (19 reputation)Junior Member (19 reputation)Junior Member (19 reputation)Junior Member (19 reputation)Junior Member (19 reputation)Junior Member (19 reputation)Junior Member (19 reputation)Junior Member (19 reputation)

Group: Forum Members
Posts: 18, Visits: 1

NEAS:

Could you please explain the following sentence in Fox Module 13 Dummy Variable Regression: "the equations at the bottom of page 123 show how dichotomous factors affect the slope coefficient."

I'm confused by this statement because I thought that this section assumed common slopes (parallel regression surfaces), and that the the dichotomous factors only affecte the y-intercept.

Thanks for your time.

Michelle

[NEAS: Thank you for noticing the typo. It should say: "at the bottom of page 132 ..."]


Nicholas Chu
Forum Newbie
Forum Newbie (4 reputation)Forum Newbie (4 reputation)Forum Newbie (4 reputation)Forum Newbie (4 reputation)Forum Newbie (4 reputation)Forum Newbie (4 reputation)Forum Newbie (4 reputation)Forum Newbie (4 reputation)Forum Newbie (4 reputation)

Group: Forum Members
Posts: 4, Visits: 1

In the 1st equation of P. 128, SE(B1) and SE(B2) of Ducan's data are 0.000219 and 0.336.

However, on P. 107, SE(B1) and SE(B2) of the same data set are 0.098252 and 0.11967.

Why is the different?

Besides, I also want to know why adding the dummy regressors into the model will make the Estimated Standard Error (i.e. SE(A) to be 3.116 before adding D1 and D2; SE (A) to be 5.2275 after adding D1 and D2)

[NEAS: Adding dummy variables changes the standard errors. Fox shows how the standard error depends on all the explanatory variables.]


apgarrity
Junior Member
Junior Member (16 reputation)Junior Member (16 reputation)Junior Member (16 reputation)Junior Member (16 reputation)Junior Member (16 reputation)Junior Member (16 reputation)Junior Member (16 reputation)Junior Member (16 reputation)Junior Member (16 reputation)

Group: Awaiting Activation
Posts: 13, Visits: 393
Please update page and formula references for the newest (3rd edition of the book) thanks.
GO
Merge Selected
Merge into selected topic...



Merge into merge target...



Merge into a specific topic ID...





Reading This Topic


Login
Existing Account
Email Address:


Password:


Social Logins

  • Login with twitter
  • Login with twitter
Select a Forum....











































































































































































































































Neas-Seminars

Search