Re: cite for using linear regression instead of logistic regression

[Joe Ward]

David --   Logistic Regression is more appealing to some folks since it maps the Predicted values into the range 0-1.   If you do a least-squares regression predicting a 0-1 dependent variable, the predicted values may not be mapped into 0-1 (e.g. some predicted values may be < 0 and some may be > 1.   However, for "practical" decision-making such as "selection", "classification" the results will be the same.   Since you brought up the question, I'm sure that the "logistic regression" folks can enlighten us on the "practical" advantages of "logistic regression".   -- Joe   ******************************** Joe Ward 167 East Arrowhead Dr. San Antonio, TX 78228-2402 Home phone: 210-433-6575 Home fax: 210-433-2828 Email: [EMAIL PROTECTED] http://www.ijoa.org/joeward/wardindex.html -------------------------------- Health Careers High School 4646 Hamilton Wolfe San Antonio, TX 78229 Phone: 210-617-5400 Fax: 210-617-5423 ********************************     ----- Original Message ----- From: "David Duffy" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Sunday, March 18, 2001 8:41 PM Subject: Re: cite for using linear regression instead of logistic regression > Scheltema, Karen <[EMAIL PROTECTED]> wrote: > > > I've read several times on this listserve comments from people that when > > p(y) is not extreme, a logistic regression model can be estimated by a > > linear regression model. > > Some references cited by Harvey (1982):  also BF&H > > Harvey WR (1982).  Least squares analysis of discrete data.  J Anim Sci > 54: 1067-1071. > > Cochran WG (1940).  The analysis of variance when experimental errors follow > the Poisson or binomial laws.  Ann Math Statis 11: 335. > > Cochran WG (1943).  Analysis of variance for percentages based on > unequal numbers. JASA 38:287. > > Li JCR (1964).  Introduction to statistical inference I.  Ann Arbor: Edwards. > > -- > | David Duffy.                                                     ,-_|\ > | email: [EMAIL PROTECTED]  ph: INT+61+7+3362-0217 fax: -0101    /     * > | Epidemiology Unit, The Queensland Institute of Medical Research \_,-._/ > | 300 Herston Rd, Brisbane, Queensland 4029, Australia                 v > > > ================================================================= > Instructions for joining and leaving this list and remarks about > the problem of INAPPROPRIATE MESSAGES are available at >                   http://jse.stat.ncsu.edu/ > =================================================================