Are there task specific performance effects for differently configured numeric keypads?

Previous studies of numeric keypad user preference and performance have indicated that the telephone layout (TEL) was superior to the layout seen on computer keyboards and adding machines (ADD). A recent study (Straub and Granaas, 1993) suggested that the TEL preference was subject to task specific effects. To investigate the possibility of task specific performance in using keypads, 24 subjects were tested on four different keypad layouts (TEL, zero at top; TEL, zero at bottom; ADD, zero at top; ADD, zero at bottom) using three different tasks (four digit strings, seven digit strings, and seven digit strings depicted like standard North American telephone numbers). Results indicated that differences in rate of performance across the four keypad layouts were the result of zero placement, with the zero in the bottom position yielding the fastest keypad use. No significant differences were found for error rate across the different keypads. No task specific performance effects were found. These findings suggest that either the ADD or TEL layouts could be adopted universally for numeric keypads, with the stipulation that the zero key be placed below the other keys.