A stratified cross-over study was conducted to evaluate the effects of familiarity with computer on the performance of computerized neurobehavioral tests. Effects of keyboard type on the relationship between familiarity with computer and performance of computerized neurobehavioral tests were also examined.

We tested 70 persons classified into three groups according to degree of familiarity with computer: Group 1 was people with no computer competency, Group 2 had slight working experience on the computer, and Group 3 were highly competent and efficient on the computer. For each group, we executed 4 tests of the Korean version of the Swedish Performance Evaluation System (Simple Reaction Time, Addition, Symbol Digit, and Finger Tapping Speed). Three types of keyboard were applied for all three groups. Type 1 keyboard was a conventional layout with 106 keys. We removed the unused keys except for the 17 which were used in the test and remodeled the normal keyboard into Type 2 keyboard with 17 keys (10 numeric keys, 4 arrow keys, space bar, and both Ctrl keys). Type 3 keyboard with 11 keys consisted of 10 numeric keys and the space bar was newly developed for the computerized neurobehavioral performance test.

It was found that Simple Reaction Time, Addition, and Finger Tapping Speed were not affected by familiarity with computer and type of keyboard. The mean reaction time of Symbol Digit adjusted by age and educational level with Type 1 keyboard was found to have significant difference among the three groups (F=3.347, p=0.041). The performance of Symbol Digit in Group 1 was found to be significantly lower than that in Group 2. The performances of Symbol Digit measured with Types 2 and 3 keyboards were not found to have significant differences among the three groups. In Groups 1 and 3, the performances of Symbol Digit measured with Type 1 keyboard were found to be significantly lower than those of other keyboards.

According to the above results, Simple Reaction Time, Addition, and Finger Tapping Speed have high priority over other neurobehavioral tests for the workers having varying degrees of computer familiarity. It is also suggested that familiarity with computer and keyboard type needs to be considered in interpretation of the performance of Symbol Digit. We recommended executing the computerized neurobehavioral test with keyboards 2 or 3 to reduce the effect of computer familiarity.