Numeric response vector (or matrix, for the one-way column form) used to fit the ANOVA model. This property is read-only.

Table whose variables are the factors used to fit the ANOVA model, with one column per factor named after FactorNames and one row per observation. This property is read-only.

Character vector describing the fitted ANOVA model formula. (MATLAB returns a formula object; Octave returns the equivalent character vector.) This property is read-only.

Cell array of character vectors used as factor names in the fitted ANOVA table. This property is read-only.

Cell array of character vectors naming the model coefficients when the selected backend exposes them, otherwise an empty cell array. This property is read-only.

Character vector selecting 'one', 'two', 'three', or 'hierarchical' sums of squares. This property is read-only.

Positive integer indices of random factors. This property is read-only.

Character vector 'all' or positive integer indices of factors treated as categorical. This property is read-only.

Character vector used as the response name in formula display. This property is read-only.

Scalar number of response observations used by the model. This property is read-only.

Numeric coefficient table returned by the fitted anovan backend, or derived from a LinearModel object. This property is read-only.

Table with variables Raw (observed minus fitted values) and Pearson (raw residuals scaled by the root mean squared error) when the selected backend exposes residuals, otherwise empty. This property is read-only.

Table with variables MSE, RMSE, SSE, SSR, SST, RSquared, and AdjustedRSquared summarising the fitted model. This property is read-only.

anova: obj = anova (Y)

anova: obj = anova (factors, Y)

anova: obj = anova (…, name, value)

anova: obj = anova (mdl)

Y is a non-empty numeric response vector or matrix. factors contains grouping variables for vector responses and may be a grouping vector, a matrix of grouping variables, or a cell array of grouping vectors. If factors is omitted and Y is a matrix, columns of Y are treated as groups following anova1 matrix syntax.

Supported name-value arguments include 'ModelSpecification', 'SumOfSquaresType', 'FactorNames', 'CategoricalFactors', 'RandomFactors', 'ResponseName', 'Alpha', and 'Display'. Passing 'Reps' selects the balanced two-way anova2 backend when Y is a non-vector matrix.

mdl may be a LinearModel object, in which case the ANOVA object is populated from the fitted linear model’s public properties.

anova: disp (obj)

The display includes the selected backend, fit state, number of factors, sum-of-squares type, and significance level.

anova: s = stats (obj)

anova: s = stats (obj, type)

The model is fitted first if needed. The optional type argument is accepted for MATLAB syntax compatibility; currently both 'component' and 'summary' return the backend ANOVA table.

anova: means = groupmeans (obj)

anova: means = groupmeans (obj, factors)

The returned value is a table with one row per factor-level combination and columns for the level, mean, standard error, and confidence bounds.

anova: boxchart (obj)

anova: h = boxchart (obj, …)

This method uses boxplot as the graphics backend in Octave.

anova: plotComparisons (obj)

anova: h = plotComparisons (obj, …)

The method delegates interval computation to multcompare.

anova: v = varianceComponent (obj)

anova: v = varianceComponent (obj, …)

Random-factor variance component estimates are not yet implemented.

anova: C = multcompare (obj)

anova: [C, M, H, GNAMES] = multcompare (obj, …)

The method fits the object if needed and delegates to the package function multcompare using the backend Stats structure. Additional arguments are passed through unchanged.

 y = [1; 2; 3; 4; 5; 6; 10; 11; 12];
 g = [1; 1; 1; 2; 2; 2; 3; 3; 3];
 a = anova (g, y, 'SumOfSquaresType', 'two');
 summary (a);

ANOVA TABLE (Type II sums-of-squares, backend = anovan):

Source      Sum Sq.     d.f.        Mean Sq.    Eta Sq.     F           Prob>F      
------------------------------------------------------------------------------------
X1          126         2           63          0.95455     63          9.3914e-05  
Error       6           6           1                                               
Total       132         8                                                           

MSE: 1    DFE: 6    Alpha: 0.05

 y = [1; 2; 3; 4; 5; 6; 10; 11; 12];
 g = [1; 1; 1; 2; 2; 2; 3; 3; 3];
 a = anova (g, y, 'SumOfSquaresType', 'two');
 C = multcompare (a, 'display', 'off')

C =

   1.0000e+00   2.0000e+00  -4.9979e+00  -3.0000e+00  -1.0021e+00   1.0402e-02  -3.6742e+00   6.0000e+00
   1.0000e+00   3.0000e+00  -1.0998e+01  -9.0000e+00  -7.0021e+00   9.9474e-05  -1.1023e+01   6.0000e+00
   2.0000e+00   3.0000e+00  -7.9979e+00  -6.0000e+00  -4.0021e+00   6.4995e-04  -7.3485e+00   6.0000e+00

 y = [10; 12; 11; 14; 16; 15; 9; 8; 10];
 g = [1; 1; 1; 2; 2; 2; 3; 3; 3];
 a = anova (g, y, 'SumOfSquaresType', 'two');
 plotDiagnostics (a);