TYPES OF ERROR IN THE D2 TEST OF ATTENTION

The validity of different error types in the d2 Test (Brickenkamp, 199 4) was examined in three studies with 225 healthy and 153 brain-damage d subjects. The total error score (sum of all errors divided by the nu mber of items worked on) largely consisted of deletion errors. Items w hich differed from the targets with respect to two features (letter, n umber of dashes) were cancelled extremely seldom (''double errors''). Two kinds of confusion errors (item with wrong number of dashes or wit h wrong letter cancelled) and deletion errors were slightly correlated (r(s) =.14 to .42). Correlations with errors in two concentration tes ts (new test, ''Revisionstest''), as well as with missings and false a larms in two attention tests, were low (r(s) =.10 to .38). Error score s correlated with construct divergent variables (tests on intelligence , short-term and long-term memory, attention) as high as with construc t convergent variables (r(s) less than or equal to \.51\). Deletion an d dash errors differed somewhat with respect to their correlations wit h other variables. Proneness to errors as assessed by the d2 Test seem s to be a very heterogenous construct. Double errors are discussed as the optimal indicators of working against instructions or simulation.