The results of the experiment were inconclusive. Due to the small sample size there was a large variance in the results; this is evident in the large values of standard deviation.
The access time for the backspace key was lowest for Keyboard B and highest for keyboard C. Time to hit the backspace key was about three times higher than the time to hit other keys. This result was consistent across the three keyboards.
The typing speed is determined by dividing the number of words in a given sample by the total test time, which includes correcting time also. Subjects were not directed as to correcting as they typed or to do all the corrections at the end. Typing speed for keyboards A and B are similar, showing no evidence of an advantage for the split spacebar. The typing speed on keyboard C, is notably slower. This is not surprising since the hand position for this keyboard is very different from the other two keyboards, and the user would need some time to acclimate to this style.
A source of variance in the test results are subject correction methods. Most subjects (who were not touch typists) spent time reading the screen or test paper during the test. This results in an inflated time for a particular key combination. To improve the experiment, keystroke times of greater than one second should be disregarded.
The error rate is determined by dividing the number of time that the backspace key was used over total test time. This method doesnt consider the user using alternate methods, such as the arrow keys, to correct an error. An alternative to this method would be to manually record stroke errors, though this may be difficult to execute. Using the method mentioned first, keyboard C has the largest number of errors.
The subjective satisfaction was determined by a user survey. Keyboard B scores the lowest in this category with keyboards A and C having similar scores. Keyboard C has the most fluctuation in responses, indicating users were either very pleased or very frustrated with this new style of keyboard.
Subject Comments
Keyboard A - Mostly neutral reaction. They have used this keyboard (or one similiar to it for a number of years).
Keyboard B - Most users did not use the modified space bar during the test, instead favoring the normal backspace bar on the upper right. Further testing might include a disabling of the traditional back space key, either through hardware or a simply constructed barrier.
Keyboard C - Positive reaction by experienced and inexperienced typists. Most users felt they would need more time to become comfortable with this keyboard. Two subjects strongly disliked this keyboard.