Requirements for a valid diagnostic of attention: 10 postulates
(provisional translation, all rights reserved, © JOA 1997)

definition   ease   coordination   distraction   probability  
information   recording   sequence   latent power   allocation integration  

1) Definition: Like physical power is defined as work per time, in order to measure concentration power, the testee has to perform discriminative task on certain stimuli in a limited time.

2) Ease: The stimuli shall be easy and known by everybody, to avoid unneccessary confounding with training effects.

3) Coordination: The stimuli shall vary on at least two relevant dimensions, thus requiring mental coordination (not just one-dimensional yes/no decisions).

4) Distraction: Furthermore, the stimuli shall vary on at least one irrelevant stimulus dimension, in order to involve a distraction task.

5) Probability: The distribution of different kind of items on the test forms has to obey strict stochastic rules, thus leading to well defined error probabilities. This is neccessary for estimating the number of items processed with concentration, and to seperate those from "correct" responses, which result from processing without concentration, e.g. guessing.

6) Information: It must be granted, that each item requires full information processing, i.e. coordination of all stimulus dimensions. It has to be avoided, that items can be judged errorfree solely on the basis of partial information processing. Many tests fail to fullfill this criterion, because they allow rejection of non-targets after checking only one of two relevant stimuls dimensions. The FAIR avoids this problem by introducing a special (copy right protected) target definition

7) Recording The behaviour of the testee has to be fully recorded. Recording only part of his or her responses gives ambiguous information.

8) Sequence: It must be granted (or at least controlled) that the testee processes the items in the pre-determined order, processing the relevant (at least two) stimulus dimensions at a time. Otherwise, the testee could - by partial exclusion - process the stimulus dimensions seperately. This could result in an error free protocol which does not correspond to concentration power at all. Namely paper and pencil tests fail to grant or even check the sequence of item processing. The FAIR solves this problem through his (copy right protected) mode of full recording

9) Latent Power: The evaluation of the test protocol has to be an unbiased estimate of "how many items" a testee can process with concentration and thus errorfree, given the time limit. Corrections for randomly "correct" responses (e.g. guessing) have to be performed. Penalizing for errors shall be announced to the testees.

10) Allocation Integration: Since the attention recources can be allocated to the task itself as well as to monitoring the (quality of the) task, both aspects should be measured and integrated.

Only if a concentration power test complies with all 10 postulates, it will have a reliable basis for a valid psychological measurement of attention.