The calculation of relevance is initiated by the analyst as soon as the analyst is convinced to have recorded all variants with acceptable success. By clicking "Process" the program starts its quest for relevance. At the rightmost corner of the status line a status bar shows the progress of the calculation.
Initially, the program checks whether all active examples are described completely. All factors have to be filled with values. If not, the program terminates execution. If there were incompletely described examples the program could not judge on identity and difference of values and therefore examples. Moreover, the example base must consist of at least two active factors and at least two active examples. At most, the program processes 15 factors simultaneously. This upper limit is set because with an increasing number of active factors the number of required examples increases as well. Actually, it is not expected to process analytical problems with more than 15 active factors.
The user always can interrupt processing by clicking "exit". After regular execution of the procedure the program displays the relevance values for each factor. The relevance values are shown on top of each factor button (as long as the actually selected filter does not restrict the view only to hidden factors).
The more similar examples the analyst has recorded the more the relevance of the factors decreases. Adding more examples that are similar to the known ones decreases the relevance values even further.
Single variants possibly receive especially low intermediate results. This indicates ideal examples. By high intermediate results the analyst is warned that this example differs from most of the other examples. This information is given by a new window that is opened after clicking "Session" and "Tree view".
The FACTORFINDER-Software shows the effects of entering extremely different examples on the result of the analysis. Such examples differ from ideal examples even in important factors. After adding such a variant to the example base the relevance values decrease. This can be expected as with each new acceptable example the possibility for plausible comparisons is increases. The tolerability of non-ideal values seems to be larger. A larger tolerability equals a smaller relevance of these factors.
Despite a decreased relevance of the factors the variance in relation to supposedly ideal examples is larger than critical. Thereby, the FACTORFINDER-Software identifies those different examples. This either disconfirmes the validity of the success hypothesis associated with this example or it disconfirmes the current relevance hypothesis.