An Analysis of First Impression in the Reading of Electrical Schematic Diagrams

The initial impression of electronic troubleshooters of schematic diagrams was studied in a series of two experiments. In one experiment the time available to the subjects to view the information was restricted to 12 seconds and in the second experiment the subjects could look at the material as often and as long as they wished. The characteristics of the first element redrawn were: (1) the same elements were consistently selected, (2) the element was chunked with other elements, (3) the element was in a branch instead of a loop, (4) the element was along the exterior of the element and (5) the element was active more often than would be expected from the number of active elements in the circuit.

preferences are summarized in Table 1. The numbers used to identify different circuit components were assigned after the experiment was completed, and they were not available to the technicians during any phase of the memory or perception tasks.
The first point to be noted from Table 1 is that there was perfect agreement between the memory and perception tasks with regard to the element most often redrawn first. Indeed, on the second most preferred initial element, there was agreement on eight of the ten schematics, with the only differences being on schematic seven and schematic nine. The remaining choices are, for the most part, one time selections. Using a simple binomial distribution calculation, it can be rejected with 0.9999 confidence in favor of the alternative nypothesis of a patterned selection.

Analysis o f First Element Chunking
?'he information on first element chunking is summarized below for the memory and perception tasks. The data reflect a clear disposition on the part of the technicians to initially absorb information in chunks, rather than element by element separately. Hypothesizing a 4 to 1 chunking ratio for first elements, one can perform a goodness o f fit test. The 4 to 1 chunking ratio implies that 4 out of every 5 first elements are chunked. Under the null hypothesis, the chi square value for the perception task is 7.04 and for the memory task, the chi square value is 0.81. dence, f o r the perception task, there is no question whether the hypothesized chunking ratio applies, while for the PERCEPTION TASK

First Elements
T o t a l -Number (Number in parenthesis indicates the of number of times a particular element Circuit was picked first) Elements 18 20 11 24 31 21 17 2 1 16 16 memory task, the null hypothesis cannot be rejected with any significant degree of confidence. Of the two tasks, there is more inclination on the part of the technician to chunk information in the memory task than in the perception task, as discussed earlier. With perception,,he had free access to the schematic, and therefore he could encode as little as one element per glance. This difference between the two tasks very likely accounts for the discrepancy between the observed and the hypothesized chunking ratio in the perception task.

Analysis of Branch Versus Loop First Elements
In courses on circuit theory, there are two principal methods of circuit analysis which are taught. One method applies to closed loops and is called loop analysis, while the other method applies primarily to branches and is called node analysis. Therefore, the circuit geometry applicable to a particular element determines how that element would be viewed from a circuit analysis standpoint.
The ten schematics used in the study contained a total of 195 elements. Of these, 91 were branch elements and 104 were loop elements. Since there were only a total of 150 first elements in each of the tasks, the actual numbers of both types of elements will be changed from a base of 195 to a base of 150 f o r purposes of analysis. These new expected values, along with the observed values, are snown in Table 3 below.
Employing a chi square goodness of fit test under the null hypothesis that chance factors alone dictate whether a loop element or a branch element will be chosen first, the memory chi square value is computed as 1a.35. These correspond to P values which are both l e s s than 0.005 (chi square: 7.88, 1). The observed frequencies suggest that f o r the circuits employed in this study, branch elements are preferred or selected first on a 2 to 1 ratio over loop elements. Employing a goodness of fit test under this hypothesis results in a memory chi square value of 0.27 and a perception chi square value of 2.43. Hence the null hypothesis cannot be rejected with any significant degree of confidence.

Analysis of Interior Versus Exterior First Elements
The impression analysis along this dimension is similar to that just undertaken with regard to branch and loop elements. Exterior elements are those which are located on the perimeter of a circuit, while interior elements are those which are not exterior elements. The observed and expected frequencies are shown in Table 4. As in the pre-vious section, the expected frequencies have been based on a total of 150 element s .
Under the null hypothesis that chance factors alone determine whether an interior or an exterior element is selected first, the memory chi value is 14.62, as is the perception chi square value. For one degree of freedom, the null hypothesis may be rejected with greater than 0.995 confidence. Active elements contribute energy to a circuit, while passive elements either store o r dissipate circuit energy. The observed and expected frequencies for these two categories are shown in Table 5. As in the previous two sections, the expected frequencies have been based on a total of 150 elements. Table 5 Active versus passive first elements, indicating the extent to which active or passive initial elements were selected by technicians during the memory and perception tasks.

MEMORY PERCEPT ION
Observed/Expected Passive Elements... 103/242 104/142 Under the null hypothesis that chance factors alone determine whether an active or a passive element is selected first, the memory chi square value is 200.84 and the perception chi square value is 190.67. For one degree of freedom, the n u l l hypothesis may be rejected with greater than 0.995 confidence.

PROCEEDINGS of the HUMAN FACTORS SOCIETY-24th ANNUAL MEETING-1980
The Hence t h e n u l l h y p o t h e s i s c a n n o t be r ej e c t e d w i t h any s i g n i f i c a n t d e g r e e of c o n f i d e n c e .
A n a l y s i s o f F i r s t Element S p a t i a l Locat i o n s