Quite often, there are several measured observations for each of the units of an analysis. For instance, it is a good practice to digitize each specimen twice to reduce the possible effects of measurement error. But the result is that there are two landmark configurations for each specimen, and the unit of analysis is the specimen. It is therefore necessary to average over the replicates.
In MorphoJ, observations can be averaged if a classifier has been set up that specifies which observations belong to one unit. Averaging is then done within each of the groups designated by this classifier.
To start, select the dataset in the Project Tree window and then choose Average Observations By from the Preliminaries menu. A dialog box like the following will appear:
In this screen shot, the user has selected a dataset named 'Keratohyale', as can be seen from the line of text at the top of the box. The text field below it is for the name of the new dataset that will contain the averaged data. By default, this is the same as the name of the original dataset, with ', averaged' appended.
The classifier to be used as a grouping criterion for averaging can be selected in the drop-down menu labeled Average by ... to the left (a classifier named 'individual' in the screen shot).
The list Data types: to the right of it shows the data types found in the dataset. By default, all data types are selected in this list, but it is possible to change that selection if the user wants to exclude some data types. (For instance, it may or may not be sensible to average the covariates.)
Click the Execute button to start, or the Cancel button to abort the procedure.
As a result, a new dataset with the averaged data is appended to the original dataset in the Project Tree. For each observation in the new dataset, the identifier has one of the values of the classifier selected in the Average by ... menu. In the example of the screen shot above, the identifiers in the new dataset 'Keratohyale, averaged' will therefore have the values of the classifier 'individual' in the dataset 'Keratohyale'.