The chart, however, is likely to be very misleading, as it has been plotted by methods which are not in accordance with usual practice. The variables have been reversed, and the independent variable has incorrectly been made the vertical scale. Besides that, the vertical scale reads downward instead of upward. In all kinds of curve plotting it is common to have the two scales begin with zero at the lower left-hand corner of the chart. Here the two scales begin the zeros at the upper left-hand corner of the chart. Unless the reader will turn Fig. 154 on its side so as to make the two zeros at the lower left-hand corner, he may find great difRculty in interpreting the chart. Fig. 155 shows a replot of the data of Fig. 154. Here the curves are plotted on a "more than" basis, but it would have been better if the

'hit

a JO to

1 £34 56 7 e 9 10 II le 13 14 15 16 n le 19 £0

Time of Answer in Seconds

Courtesy of Data, Chicago

Fig. 152. Time Required for Operators to Answer Telephone Calls in Towns of Different Size in Wisconsin

These curves start at the lower left-hand portion of the field and trend upward, showing that they are plotted on a "less than" basis. Curve A shows a smaller time required to answer calls than Curves B or C, yet the actual position of Curve A on the chart is higher than either curves B or C. If cumulative frequency curves are plotted on a "more than" basis the position of several curves on a chart is relatively such that the reader is not confused so much as when curves are plotted on a "less than" basis