Content Disclaimer
Copyright @2014.
All Rights Reserved.
StatsToDo : Concordance (Agreements) Explained

Links : Home Index (Subjects) Contact StatsToDo

Related link :
General :
Cronbach's Alpha Program Page
Sample Size for Estimating a Cronbach's Alpha Program Page
Sample Size for Estimating a Cronbach's Alpha Explained and Tables Page
Sample Size for Comparing Two Cronbach's Alphas Program Page
Nominal : Kappa for Nominal Data Program Page
Binary : Kuder Richardson Coefficient for Binary Data Program Page
Ordinal :
Kappa (Cohen and Fleiss) for Ordinal Data Program Page
Kendall's W for Ranks Program Page
Continuous :
Agreement Program Page
Intraclass Correlation Program Page

Introduction Cronbach Alpha Nominal Binary Ordinal Continuous References
Concordance, in the context of statistics, is to study agreements between judges, scales, and measurements. This page provides support for all concordance programs provided by StatsToDo.

The information on this page is organized in separate panels according to the numerical nature of the measurements, and consists of the following

A commonly used test for agreement, which can be applied to all data, and cope with multiple groups, is Cronbach's Alpha. StatsToDo provides 3 programs and 1 detailed explanations for Cronbach's Alpha. They are

Nominal Data : where the numbers used are labels or names, and the scales are unrelated. Typical nominal numbers are 0=no illness, 1=appendicitis, 2=urinary infection, or 1=Caucasian, 2=Asian, 3=African, and so on. StatsToDo provides 1 program evaluating concordance for nominal data, Kappa

Binary Data : where the numbers used are 0/1, representing binary outcomes of no/yes, false/true, male/female, negative/positive, and so on. StatsToDo provides 1 program evaluating concordance for binary data, Kuder-Richardson Coefficient

Ordinal Data : where the number and whole integers represents ranks or order, so that 3>2>1, but the distance between the numbers are not specified, 3-2 is not necessarily the same as 2-1. The simplest ordinal scale is a 3 point scale of 0=no, 1=perhaps, 2=yes. The most commonly used ordinal scale is Likert, where 0=strongly disagree, 1=disagree, 2=neutral, 3=agree, and 4=strongly agreee. StatsToDo provides 3 programs evaluating concordance for ordinal data

Continuous Data : where the numbers are continuous (or near continuous) measurements that are normally distributed. StatsToDo provides 3 programs evaluating concordance for continuous data

Details for each program and procedure are discussed in their individual panels

StatsToDo Home Page    Contact StatsToDo