SAS, single-factor model. Iris sepal, petal measurements.

Investigate: species effect on sepal, petal widths and length.

Datasets.sas

data iris;

infile '/home/u60665181/_Storage/iris.dat';

input sl sw pl pw species;

run;

Analysis_IrisMeasures.sas

%include "/home/u60665181/SAS Revision/GLM/Datasets.sas";

/*sl sw pl pw species. s: sepal. p: petal. l:length. w: width.*/

proc discrim data=iris pool=test short;

class species;

var sl sw pl pw;

run;

Homogeneity of Within Covariance Matrices test failed. MANOVA on species - result not guaranteed.

proc glm data=iris;

class species;

model sl sw pl pw = species /nouni;

manova H=species M=sl, sw, pl, pw;

means species;

means species/ cldiff bon;

output out=residuals residual=r1-r4;

run;

Significant species effect on the responses.

Parameter estimates for 1-factor model can be obtained from proc reg.

For sepal length, Specie 3> Specie 2; Specie 3> Specie 1; Specie 2> Specie 1; i.e. Specie 3 > Specie 2 > Specie 1 significantly.

%include "/home/u60665181/_Storage/Chisq_plot.sas";

%chisq_plot(residuals, r1 r2 r3 r4);

Multivariate normality assumption not hold.