INSTRUMENTS FOR PRESENTING QUANTITATIVE DATA 



87 



pendent on the linear separation of gradu- 

 ation marks or on unit scale distance and, 

 for circular scales, as independent of dial 

 diameter. The data of Williams and 

 Grether which are presented in Fig. 6 repre- 

 sent readings on dials of four different diam- 

 eters, from 1 to 4 inches. The data of 



desired ranges of values. Should the read- 

 ing tolerances for any particular task be such 

 that there is no need to employ the most 

 effective unit scale distance or scheme of 

 graduation. Figs. 6 and 7 may be used to 

 suggest to what limits the graduation mark 

 spacing may be reduced. 



so 



A. DIALS OKADc/ArcD ay ''^'^^ 



J I I L 



0X107 dO/<t- 0.02Z 



OM^ 



0.086 



0J)07 OiJIU- OJJZZ 



OXiUU- 



0J)88 



AKC L£nGTH £OUAL 7D ONE SCAIS UMT - //\l l/iCHtS 



Fig. 7. Frequency of reading errors for dials graduated by tens and by fives. Data from Kappauf 

 and Smith (41) for six subjects, each of whom made 30 readings for each type and size of dial under 

 instructions to be as accurate as possible. Brightness level: six foot lamberts. 



Kappauf and Smith in Fig. 7 are for two 

 sizes of dial, 1.4 and 2.8 inches in diameter. 

 The two figures suggest that the foregoing 

 assumptions are not far from the truth. 

 The subject is currently under further in- 

 vestigation, but for the present it seems 

 reasonable to use the unit scale distance data 

 cited above in designing scales to cover any 



Kappauf and Smith (41) report the read- 

 ing time data shown in Fig. 8. In their 

 test situation, requiring the serial reading of 

 12 dials of identical design and for the full 

 360 degree scales which they used, reading 

 time proved to be related primarily to the 

 number of units represented on the scale and 

 was relatively uninfluenced by dial size 



