RobertNiles.com
About Robert
Reporting Help
Finding Data on the Internet
Questions for Candidates
Stats Lessons
Mean
Median
Percent
Per capita
Standard Deviation
Margin of Error
Data Analysis
Sample Sizes
Stats Tests
Help Board
Bookstore


The Cartoon Guide to Statistics

This isn't some dumb-downed whitewash. It's a must-read for a beginner taking his or her first steps toward understanding stats.
More information
-->

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Multiple regression for ordinal variables?
Message posted by Jinn-Yuh Guh on October 6, 2000 at 12:00 AM (ET)

Dear all:
How to perform a Multiple regression for ordinal variables?
Thanks


READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Multiple regression for ordinal variables?
Message posted by nancy diehl on October 6, 2000 at 12:00 AM (ET)

By ordinal, I will assume you mean categorical variables like Freshman, Sophomore, Junior, Senior, or ranks in the military, for example. In cases like these
you are dealing with qualitative not quantitative varibles so they are handled a bit differently in regression. If you are dealing with only two levels of your
categorical variable, then just code one level with 0 and the other with 1. Proceed with your analysis as you would any regression where the regression coefficient
is tested for significance. However, if you are deal with more than two levels, then you will create what is referred to as dummy variables, or inidicator variables.
If you have m levels you will need m-1 columns that are coded as 0s or 1s depending upon if that category level is present or not. For example, if you are looking at
geographical areas and there are 4, A, B, C, D. Then you will need three columns that will be coded as follows:
1 0 0 to represent Area A
0 1 0 to represent Area B
0 0 1 to represent Area C
0 0 0 to represent Area D

These three dummy variables are then evaluated in the regression to see which is relevant.


Re: Multiple regression for ordinal variables?
Message posted by JG on October 6, 2000 at 12:00 AM (ET)

Is your dependent variable continuous and your error term normally distributed with constant variance ? If not, you will have to transform your dependent variable appropriately.



Your $5 contribution helps cover part the $500 annual cost of keeping this site online.

Niles Online Sites:RobertNiles.comTheme Park InsiderViolinist.com

RobertNiles.com™, the site, content and services 咀opyright 1996-2002, Robert Niles.
All rights reserved. Questions? Comments? Read my Privacy Policy, or E-mail me!