Abstract:

The analysis of variance (ANOVA) is a statistical technique by which a hypothesis test is performed. If the accuracy of this contrast could be relied upon, it could be ensured that after the statistics have been calculated there is no more pertinent information left in the original data and the waste and original data could be dispensed with and all the attention in the interpretation of statistics. In practice, it would be unwise to do so without further checks, since the data can be removed from the assumptions of the applied model and therefore obtain invalid results. In these cases it is necessary to work the "raw data" to adapt them to the model. The validity of the results obtained in the analysis of variance is subject to the assumptions of the model being met.

These assumptions are normality, homoscedasticity of variance and independence. The observed residuals are the basis of the diagnostic tools that are used to verify such assumptions of the model because if the assumptions are met, the residuals can be seen as a random sample of a normal distribution with zero mean, constant variance and without any pattern or trend.

To improve the teaching of this topic, an applet was selected for a nominal variable and 6 levels and for each one of these you can select the mean, the standard deviation and the sample size. With the use of the applet the student constructs the knowledge about how much the data of the assumptions can be diverted and which of these is more important to fulfill.

With the help of new technologies, (such as applets, in this case), the teaching of these subjects is facilitated, in order that the students have a clearer and more concise idea of the aforementioned concepts. It is important that the student understands that the analysis of residuals should be considered as mandatory to detect serious deviations from the assumptions of the model, which would result in erroneous results.

Keywords:

ANOVA, ICTs, Residual.