92 



DESIGN AND USE OF INSTRUMENTS 



also investigated pointer form very briefly, 

 obtaining the result that full-length pointers 

 starting at the center of a circular dial are 

 superior to pointers on which a short length 

 of the tip only is visible. Instruments with 

 twin pointers, as used especially in aircraft 

 to indicate the functioning of two different 

 engines, are difficult and slow to read. The 

 trend at the moment is to replace such dual 

 instruments with pairs of smaller, single- 

 pointer instruments (59) . 



Scale Form. It is the goal of instrument 

 designers to make instruments in such a way 

 that the direction of the indicated change -^all 

 be easily perceived. The fact that vertical 

 scales seem to be the natural way of ex- 

 pressing changes in magnitude, and of alti- 

 tude or depth, in particular, has caused this 

 type of scale to receive repeated consider- 

 ation. Its principal disadvantage is that it 

 is often difficult to mechanize, but it is con- 

 ceivable that this could be more than out- 

 weighed by advantages from the operator 

 point of view. Studies of thermometer type 

 and other vertical scales are currently under 

 way in Great Britain (68) and two were 

 recently reported in this country. One of 

 the latter, by Sleight (70), unfortunately 

 offers little worthy evidence, since the results 

 were distorted by the fact that the subjects 

 had no opportunity to refixate during the 

 0.12 second exposure periods which were 

 used. They were of necessity less accurate 

 in reading any scale which was spread out 

 over an extended area. More satisfactory 

 are the data of Grether (31) which have 

 already been referred to in Fig. 11, Grether 

 used pencil-and-paper test materials and had 

 his subjects work as rapidly as they could. 

 In this situation, as will be seen from the 

 figure, the speed and accuracy of reading the 

 vertical scales were about the same as for 

 reading the combined circular scale and 

 counter. 



If the chosen scale form for any particular 

 apphcation is a circular dial, Christensen's 

 results (17) on the errors in reading clock- 

 wise and counter-clockwise scales are worthy 



of note. Christensen used a serial dial-read- 

 ing situation in which the subject was shifted 

 randomly between clockwise and counter- 

 clockwise dials. Under these conditions, re- 

 versal errors, reading the scale in the wrong 

 direction from a numbered graduation, were 

 very frequent. They occurred in about 4% 

 of all the readings made. It is not surprising 

 that a significantly greater number of these 

 reversal errors occurred on the counter-clock- 

 wise dial. But by showing that these errors 

 were two and one-half times as frequent as 

 on clockwise scales, the study adds a quanti- 

 tative emphasis to the warning against using 

 counter-clockwise scales wherever possible. 



Some Summary Data on the Accuracy of 

 Making Quantitative Readings 



The paragraphs above have dealt \\ith 

 variables which influence, or have been 

 thought to influence, operator performance 

 in taking quantitative data from instruments 

 or indicators. For the most part, the data 

 presented have been stated in relative terms, 

 i.e., design A is better than design B. Some 

 absolute data on reading accuracy are offered 

 now in Table II. This table serves at least 

 two functions. It provides a partial sum- 

 mation of the problems discussed to this 

 point, including representative data from 

 those experiments which have dealt the most 

 systematically with quantitative reading 

 problems. It also provides performance 

 records which should be especially useful to 

 persons who are planning future experiments 

 in this area and who wish to know what levels 

 of reading accuracy they might expect under 

 particular test conditions. 



Instruments for Check Reading 



Check-reading operations have attracted 

 special attention because of their particular 

 importance for operators working with banks 

 of instruments. Checking such indicators 

 has often been an instrument-by-instrument 

 affair. As a result of the growing number of 

 instruments used on some panels, various 

 schemes have been suggested for simplifying 



