抄録
To investigate into the errors caused in setting an index by an interpolation at a certain position between 1mm graduations with an accuracy of 1/10mm, results of the calculations performed by 40 persons with slide rules were examined. In these tests, each person was requested to repeat five times a round of calculations-1/(55.0)10, 1/(55.1)10, 2/(55.2)10, ……and 1/(56.0)10. The errors caused in setting an index at each position could be estimated from the difference between the values obtained by the above method and true values. As a result of statistical treatment by means of one-way layout in analysis of the variance, it was found that the error can satisfactorily be represented in such 3 items as mean value and variance of personal biases and variance of accidental error. Obtained results are as follows:
1) Biases set at the middle position of the graduation interval and on the graduation are much smaller than those at other positions.
2) Mean values of personal biases set at various positions are symmetrically distributed to the middle position of the graduation interval and there is a tendency to set the cursor line near to the middle position.
3) Variances of personal biases set at various positions are symmetrically distributed to the middle position and are particularly large at 1/10, 2/10, 3/10, 7/10, 8/10 and 9/10 positions.
4) Accuracy in setting biases at the middle position is almost same as that on the graduation.
5) Due to the effect of training, time required for calculation gradually decreases until about 200-time operations (e.g. 1/55.1).
