Update ICC functions

jmgirard · Nov 14, 2022 · 25489c0 · 25489c0
1 parent 1a18a5b
commit 25489c0
Show file tree

Hide file tree

Showing 6 changed files with 73 additions and 86 deletions.
diff --git a/DARMA.prj b/DARMA.prj
@@ -1,9 +1,9 @@
 <deployment-project plugin="plugin.ezdeploy" plugin-version="1.0">
-  <configuration file="D:\GitHub\DARMA\DARMA.prj" location="D:\GitHub\DARMA" name="DARMA" target="target.ezdeploy.standalone" target-name="Application Compiler">
+  <configuration file="C:\Users\j553g371\DARMA\DARMA.prj" location="C:\Users\j553g371\DARMA" name="DARMA" target="target.ezdeploy.standalone" target-name="Application Compiler">
     <param.appname>DARMA</param.appname>
     <param.icon />
     <param.icons />
-    <param.version>6.09</param.version>
+    <param.version>6.10</param.version>
     <param.authnamewatermark>Jeffrey M. Girard</param.authnamewatermark>
     <param.email>me@jmgirard.com</param.email>
     <param.company />
@@ -91,27 +91,14 @@
     <fileset.package />
     <fileset.depfun />
     <build-deliverables>
-      <file location="${PROJECT_ROOT}\DARMA\for_testing" name="splash.png" optional="false">D:\GitHub\DARMA\DARMA\for_testing\splash.png</file>
-      <file location="${PROJECT_ROOT}\DARMA\for_testing" name="DARMA.exe" optional="false">D:\GitHub\DARMA\DARMA\for_testing\DARMA.exe</file>
-      <file location="${PROJECT_ROOT}\DARMA\for_testing" name="readme.txt" optional="true">D:\GitHub\DARMA\DARMA\for_testing\readme.txt</file>
+      <file location="${PROJECT_ROOT}\DARMA\for_testing" name="readme.txt" optional="true">C:\Users\j553g371\DARMA\DARMA\for_testing\readme.txt</file>
+      <file location="${PROJECT_ROOT}\DARMA\for_testing" name="splash.png" optional="false">C:\Users\j553g371\DARMA\DARMA\for_testing\splash.png</file>
+      <file location="${PROJECT_ROOT}\DARMA\for_testing" name="DARMA.exe" optional="false">C:\Users\j553g371\DARMA\DARMA\for_testing\DARMA.exe</file>
     </build-deliverables>
     <workflow />
     <matlab>
-      <root>C:\Program Files\MATLAB\R2019b</root>
-      <toolboxes>
-        <toolbox name="matlabcoder" />
-        <toolbox name="gpucoder" />
-      </toolboxes>
-      <toolbox>
-        <matlabcoder>
-          <enabled>true</enabled>
-        </matlabcoder>
-      </toolbox>
-      <toolbox>
-        <gpucoder>
-          <enabled>true</enabled>
-        </gpucoder>
-      </toolbox>
+      <root>C:\Program Files\MATLAB\R2022b</root>
+      <toolboxes />
     </matlab>
     <platform>
       <unix>false</unix>

diff --git a/Functions/ICC_A_1.m b/Functions/ICC_A_1.m
@@ -1,4 +1,4 @@
-function [ICC,LB,UB] = ICC_A_1(DATA,ALPHA)
+function [ICC, LB, UB] = ICC_A_1(DATA, ALPHA)
 % Calculate the single rater agreement intraclass correlation coefficient
 %   [ICC,LB,UB] = ICC_A_1(DATA)
 %
@@ -16,31 +16,31 @@
 %
 %   Reference: McGraw, K. O., & Wong, S. P. (1996).
 %   Forming inferences about some intraclass correlation coefficients. 
-%   Psychological Methods, 1(1), 30–46.
+%   Psychological Methods, 1(1), 30–46.
 
 %% Remove any missing values
-[rowindex,~] = find(~isfinite(DATA));
-DATA(rowindex,:) = [];
+[rowindex, ~] = find(~isfinite(DATA));
+DATA(rowindex, :) = [];
 %% Calculate mean squares from two-way ANOVA
-[~,tbl,~] = anova2(DATA,1,'off');
-MSC = tbl{2,4};
-MSR = tbl{3,4};
-MSE = tbl{4,4};
+[~, tbl, ~] = anova2(DATA, 1, 'off');
+MSC = max([0, tbl{2, 4}]);
+MSR = max([0, tbl{3, 4}]);
+MSE = max([0, tbl{4, 4}]);
 %% Calculate single rater agreement ICC
-[n,k] = size(DATA);
-ICC = (MSR - MSE) / (MSR + MSE*(k - 1) + (k/n)*(MSC - MSE));
+[n, k] = size(DATA);
+ICC = (MSR - MSE) / (MSR + MSE * (k - 1) + (k / n) * (MSC - MSE));
 %% Calculate the confidence interval if requested
 if nargout > 1
     if nargin < 2
         ALPHA = 0.05;
     end
-    a  = (k*ICC) / (n*(1 - ICC));
-    b  = 1 + (k*ICC*(n - 1))/(n*(1 - ICC));
-    v  = ((a*MSC + b*MSE)^2) / (((a*MSC)^2)/(k - 1) + ((b*MSE)^2)/((n - 1)*(k - 1)));
-    FL = finv((1 - ALPHA/2),(n - 1),v);
-    FU = finv((1 - ALPHA/2),v,(n - 1));
-    LB = (n*(MSR - FL*MSE)) / (FL*(k*MSC + MSE*(k*n - k - n)) + n*MSR);
-    UB = (n*(FU*MSR - MSE)) / (k*MSC + MSE*(k*n - k - n) + n*FU*MSR);
+    a  = (k * ICC) / (n * (1 - ICC));
+    b  = 1 + (k * ICC * (n - 1)) / (n * (1 - ICC));
+    v  = ((a * MSC + b * MSE) ^ 2) / (((a * MSC) ^ 2) / (k - 1) + ((b * MSE) ^ 2) / ((n - 1) * (k - 1)));
+    FL = finv((1 - ALPHA / 2), (n - 1), v);
+    FU = finv((1 - ALPHA / 2), v, (n - 1));
+    LB = (n * (MSR - FL * MSE)) / (FL * (k * MSC + MSE * (k * n - k - n)) + n * MSR);
+    UB = (n * (FU * MSR - MSE)) / (k * MSC + MSE * (k * n - k - n) + n * FU * MSR);
 end
 
-end
+end
diff --git a/Functions/ICC_A_k.m b/Functions/ICC_A_k.m
@@ -1,6 +1,6 @@
-function [ICC,LB,UB] = ICC_A_k(DATA,ALPHA)
+function [ICC, LB, UB] = ICC_A_k(DATA, ALPHA)
 % Calculate the average rater agreement intraclass correlation coefficient
-%   [ICC,LB,UB] = ICC_A_k(DATA)
+%   [ICC, LB, UB] = ICC_A_k(DATA)
 %
 %   DATA is a numerical matrix of ratings (missing values = NaN).
 %   Each row is a single item and each column is a single rater.
@@ -16,31 +16,31 @@
 %
 %   Reference: McGraw, K. O., & Wong, S. P. (1996).
 %   Forming inferences about some intraclass correlation coefficients. 
-%   Psychological Methods, 1(1), 30–46.
+%   Psychological Methods, 1(1), 30–46.
 
 %% Remove any missing values
-[rowindex,~] = find(~isfinite(DATA));
-DATA(rowindex,:) = [];
+[rowindex, ~] = find(~isfinite(DATA));
+DATA(rowindex, :) = [];
 %% Calculate mean squares from two-way ANOVA
-[~,tbl,~] = anova2(DATA,1,'off');
-MSC = tbl{2,4};
-MSR = tbl{3,4};
-MSE = tbl{4,4};
+[~, tbl, ~] = anova2(DATA, 1, 'off');
+MSC = max([0 tbl{2, 4}]);
+MSR = max([0 tbl{3, 4}]);
+MSE = max([0 tbl{4, 4}]);
 %% Calculate average rater agreement ICC
-[n,k] = size(DATA);
-ICC = (MSR - MSE) / (MSR + (MSC - MSE)/n);
+[n, k] = size(DATA);
+ICC = (MSR - MSE) / (MSR + (MSC - MSE) / n);
 %% Calculate the confidence interval if requested
 if nargout > 1
     if nargin < 2
         ALPHA = 0.05;
     end
-    c  = ICC / (n*1 - ICC);
-    d  = 1 + (p*(n - 1)) / (n*(1 - ICC));
-    v  = ((c*MSC + d*MSE)^2) / (((c*MSC)^2)/(k - 1) + ((d*MSE)^2)/((n - 1)*(k - 1)));
-    FL = finv((1 - ALPHA/2),(n - 1),v);
-    FU = finv((1 - ALPHA/2),v,(n - 1));
-    LB = (n*(MSR - FL*MSE)) / (FL*(MSC - MSE) + n*MSR);
-    UB = (n*(FU*MSR - MSE)) / (MSC - MSE + n*FU*MSR);
+    c  = ICC / (n * (1 - ICC));
+    d  = 1 + (ICC * (n - 1)) / (n * (1 - ICC));
+    v  = ((c * MSC + d * MSE) ^ 2) / (((c * MSC) ^ 2) / (k - 1) + ((d * MSE) ^ 2) / ((n - 1) * (k - 1)));
+    FL = finv((1 - ALPHA / 2), (n - 1), v);
+    FU = finv((1 - ALPHA / 2), v, (n - 1));
+    LB = (n * (MSR - FL * MSE)) / (FL * (MSC - MSE) + n * MSR);
+    UB = (n * (FU * MSR - MSE)) / (MSC - MSE + n * FU * MSR);
 end
 
-end
+end
diff --git a/Functions/ICC_C_1.m b/Functions/ICC_C_1.m
@@ -1,6 +1,6 @@
-function [ICC,LB,UB] = ICC_C_1(DATA,ALPHA)
+function [ICC, LB, UB] = ICC_C_1(DATA, ALPHA)
 % Calculate the single rater consistency intraclass correlation coefficient
-%   [ICC,LB,UB] = ICC_C_1(DATA)
+%   [ICC, LB, UB] = ICC_C_1(DATA)
 %
 %   DATA is a numerical matrix of ratings (missing values = NaN).
 %   Each row is a single item and each column is a single rater.
@@ -16,27 +16,27 @@
 %
 %   Reference: McGraw, K. O., & Wong, S. P. (1996).
 %   Forming inferences about some intraclass correlation coefficients. 
-%   Psychological Methods, 1(1), 30–46.
+%   Psychological Methods, 1(1), 30–46.
 
 %% Remove any missing values
-[rowindex,~] = find(~isfinite(DATA));
-DATA(rowindex,:) = [];
+[rowindex, ~] = find(~isfinite(DATA));
+DATA(rowindex, :) = [];
 %% Calculate mean squares from two-way ANOVA
-[~,tbl,~] = anova2(DATA,1,'off');
-MSR = tbl{3,4};
-MSE = tbl{4,4};
+[~, tbl, ~] = anova2(DATA, 1, 'off');
+MSR = max([0, tbl{3, 4}]);
+MSE = max([0, tbl{4, 4}]);
 %% Calculate single rater consistency ICC
-[n,k] = size(DATA);
-ICC = (MSR - MSE) / (MSR + MSE*(k-1));
+[n, k] = size(DATA);
+ICC = (MSR - MSE) / (MSR + MSE * (k - 1));
 %% Calculate the confidence interval if requested
 if nargout > 1
     if nargin < 2
         ALPHA = 0.05;
     end
-    FL = (MSR/MSE) / finv((1-ALPHA/2),(n - 1),(n - 1)*(k - 1));
-    FU = (MSR/MSE) / finv((1-ALPHA/2),(n - 1)*(k - 1),(n - 1));
+    FL = (MSR / MSE) / finv((1 - ALPHA / 2), (n - 1), (n - 1) * (k - 1));
+    FU = (MSR / MSE) / finv((1 - ALPHA / 2), (n - 1) * (k - 1), (n - 1));
     LB = (FL - 1) / (FL + k - 1);
     UB = (FU - 1) / (FU + k - 1);
 end
 
-end
+end
diff --git a/Functions/ICC_C_k.m b/Functions/ICC_C_k.m
@@ -1,6 +1,6 @@
-function [ICC,LB,UB] = ICC_C_k(DATA,ALPHA)
+function [ICC, LB, UB] = ICC_C_k(DATA, ALPHA)
 % Calculate the average rater consistency intraclass correlation coefficient
-%   [ICC,LB,UB] = ICC_C_k(DATA)
+%   [ICC, LB, UB] = ICC_C_k(DATA)
 %
 %   DATA is a numerical matrix of ratings (missing values = NaN).
 %   Each row is a single item and each column is a single rater.
@@ -16,27 +16,27 @@
 %
 %   Reference: McGraw, K. O., & Wong, S. P. (1996).
 %   Forming inferences about some intraclass correlation coefficients. 
-%   Psychological Methods, 1(1), 30–46.
+%   Psychological Methods, 1(1), 30–46.
 
 %% Remove any missing values
-[rowindex,~] = find(~isfinite(DATA));
-DATA(rowindex,:) = [];
+[rowindex, ~] = find(~isfinite(DATA));
+DATA(rowindex, :) = [];
 %% Calculate mean squares from two-way ANOVA
-[~,tbl,~] = anova2(DATA,1,'off');
-MSR = tbl{3,4};
-MSE = tbl{4,4};
+[~, tbl, ~] = anova2(DATA, 1, 'off');
+MSR = max([0, tbl{3, 4}]);
+MSE = max([0, tbl{4, 4}]);
 %% Calculate average rater consistency ICC
-[n,k] = size(DATA);
+[n, k] = size(DATA);
 ICC = (MSR - MSE) / MSR;
 %% Calculate the confidence interval if requested
 if nargout > 1
     if nargin < 2
         ALPHA = 0.05;
     end
-    FL = (MSR/MSE) / finv((1-ALPHA/2),(n - 1),(n - 1)*(k - 1));
-    FU = (MSR/MSE) / finv((1-ALPHA/2),(n - 1)*(k - 1),(n - 1));
-    LB = 1 - 1/FL;
-    UB = 1 - 1/FU;
+    FL = (MSR / MSE) / finv((1 - ALPHA / 2), (n - 1), (n - 1) * (k - 1));
+    FU = (MSR / MSE) / finv((1 - ALPHA / 2), (n - 1) * (k - 1), (n - 1));
+    LB = 1 - 1 / FL;
+    UB = 1 - 1 / FU;
 end
 
-end
+end
diff --git a/fig_launcher.m b/fig_launcher.m
@@ -3,8 +3,8 @@
 % License: https://github.com/jmgirard/DARMA/blob/master/LICENSE.txt
 
     global version year;
-    version = 6.09;
-    year = 2019;
+    version = 6.10;
+    year = 2022;
     % Create and center main window
     defaultBackground = get(0,'defaultUicontrolBackgroundColor');
     handles.figure_launcher = figure( ...