HAND OUT FOR CALCULATING MEASURES OF ASSOCIATION FOR NOMINAL AND ORDINAL DATA

HAND OUT FOR CALCULATING MEASURES OF ASSOCIATION FOR NOMINAL AND ORDINAL DATA

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
HAND OUT FOR CALCULATING MEASURES OF ASSOCIATION FOR NOMINAL AND ORDINAL DATA PRE MEASURES LAMBDA GAMMA PRE: Proportional Reduction of Error E1= Errors of prediction made when the independent variable is ignored. Example: If I ask you to predict the race of an individual drawn at random from the US population, the mode is white, but the rest of the population is the amount of error in your prediction. E2= Errors of prediction made when the prediction is based on the independent variable. Example: If we add location, we know there are higher percentages of minorities in particular locations CALCULATING LAMBDA where: E1= N total - N mode of dependent variable N total =Number in sample Compute percentages for table to get the data in a format that allows you to clearly assess the nature of the relationship. Calculate E1 E1 = N total
Background image of page 1
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: N mode Calculate E2 for however many column categories you have: E2 = [N (category total) N (category mode) ] + [N (category total) N (category mode) ] etc Lambda = [E1 E2] / E1 CALCULATING GAMMA This is a symmetrical measure of association for ordinal or dichotomous nominal variables. Be sure table is arranged in rank order across and down First count the number of same order pairs (N s ) (ie: HIGH/HIGH or LOW/LOW) Count number of inverse order pairs (N d ) (ie : HIGH/LOW or LOW/HIGH) Ns=(total high/high )*(total low/low ) See page 378 in your book Nd=(total high/low)*(total low/high ) also page 378 The result is read as the proportional reduction of error 1 2 1 E E E PRE 1 2 1 E E E Lambda categories all for category for e category N N E ) ( 2 mod Nd Ns Nd Ns Gamma...
View Full Document

Ask a homework question - tutors are online