If a dataset contains observations belonging to multiple groups, it is often necessary to divide the dataset for more detailed analyses within groups. For instance, a dataset containing specimens from different species, we may want to create separate datasets for each species.
In MorphoJ, it is possible to subdivide the observations in a dataset by the values of a classifier variable.
To do this, click on the dataset in the Project Tree window and then select Subdivide Dataset By from the Preliminaries menu. A dialog box like the following will appear:
The list at the top contains the names of the classifiers in this dataset. Select one as a criterion for subdividing the dataset into groups.
The two buttons below that switch between two possibilities: either a new dataset is created for every value of the classifier or only for the values selected in the list below the buttons. In the screen shot above, the classifier 'Species' is used as the criterion for forming the groups. Because the user does not want a new dataset for all nine species, the second option was chosen and four of the species ('Mauri', 'Melan', 'Seich' and 'Simul') were selected in the list. Therefore, in this case, only four new datasets would be generated. In the Project Tree window, the new datasets are attached to the original one.
The check box Separate Procrustes fits is for selecting whether each new dataset has a separate Procrustes fit and therefore will have its own coordinate system, or whether the new datasets keep the Procrustes fit from the dataset that is being subdivided and will therefore share the same coordinate system. The latter option is useful if a user wants to compare the results of analyses in several groups, all in exactly the same coordinate system.
Click Execute to generate the new datasets. The name of each new dataset is the value of the classifier (in the example, these would be the species labels 'Mauri', 'Melan', 'Seich' and 'Simul').
To stop the procedure, click Cancel.