File Exchange

image thumbnail

mReliability

version 1.1.0.0 (202 KB) by Jeffrey Girard
MATLAB Functions for Computing Inter-Observer Reliability

1 Download

Updated 15 May 2018

GitHub view license on GitHub

MATLAB Functions for Computing Inter-Observer Reliability
Indexes for Categorical Measurements
-Aickin's alpha coefficient
-Bennett et al.'s S score
-Cohen's kappa coefficient
-Gwet's gamma coefficient
-Krippendorff's alpha coefficient
-Scott's pi coefficient
Indexes for Dimensional Measurements
-Intraclass correlation coefficient

See https://github.com/jmgirard/mReliability for more information.

Cite As

Jeffrey Girard (2020). mReliability (https://github.com/jmgirard/mReliability), GitHub. Retrieved .

Comments and Ratings (3)

Jeffrey Girard

Hi Carlos,
Thanks for your interest in mReliability. It might be easier to troubleshoot on GitHub Issues than here, but I am not getting any errors when I use your example:
>> CODES=rand(40,14)>0.5
>> mSPECIFIC(CODES, [0, 1]);
Specific agreement for category 0 = 0.506
Specific agreement for category 1 = 0.481
>> mGAMMA(CODES, [0, 1], 'identity')
ans = -0.0120

carlos lopez

Hello:
Thank you for your submission. I have had problems with routine mSPECIFIC, for inputs a CODES matrix of size 40x14 and entries 0 and 1. CODES might be similar to
CODES=rand(40,14)>0.5;
The error appears at line
denominator = sum(r_ik .* (r_i - 1));
and it appears that the solution might be to change it to
denominator = sum(r_ik .* ((r_i - 1)*ones(1,q)));
However, I am not confident about this.

I have also problems with routine mGAMMA. With the same input as before, the line
pihat = transpose(repmat(1 / n, 1, n, 1) * (r_ik ./ (r_i * ones(1, q))));
fails. One solution apparently is to change it to
pihat = transpose(repmat(1 / n, 1, n) * (r_ik ./ (r_i * ones(1, q))));
Please, let us know if this is correct. As a suggestion it will be nice to have some numerical example at hand.

carlos lopez

I have also problems with routine mGAMMA. With the same input as before, the line
pihat = transpose(repmat(1 / n, 1, n, 1) * (r_ik ./ (r_i * ones(1, q))));
fails. One solution apparently is to change it to
pihat = transpose(repmat(1 / n, 1, n) * (r_ik ./ (r_i * ones(1, q))));
Please, let us know if this is correct.

Updates

1.1.0.0

Added Aickin's alpha coefficient

MATLAB Release Compatibility
Created with R2015b
Compatible with any release
Platform Compatibility
Windows macOS Linux