Persons with no training in engineering are generally unaware of the nature of engineering analysis, and so tend to assume that testing, as a means of determining the causation of accidents, is a dominant tool of the engineer. In the following examples, we shall undertake to explain the nature of engineering analysis, and to show that it is more basic than testing because testing without analysis is meaningless. Further, while analysis is always necessary in accident reconstruction, testing is only sometimes necessary.
Consider, for example, a flight of steps in which the tread of each step (the horizontal surface) is only 6 Inches deep (in the direction from front to back). Since the shoes of most persons are considerably greater than 6 Inches in length, the toes of a descending adult will tend to overhang the tread by a substantial amount, especially since It Is not to be expected that the heel will always be placed as far back as possible, thereby increasing the overhang all the more. The result will be that the footing will not be as secure as if the tread were 10 inches deep. Thus, if a person has fallen while descending the 6-Inch-deep steps, the fall may be ascribed to the inadequate depth with reasonable probability (providing of course that there is no other contributing reason for the fall).
Straightforward as the above reasoning is, it nevertheless constitutes a valid (though simple) example of engineering analysis. Now let us consider what it would take to demonstrate the defect of the steps by testing rather than by analysis. To do this, there must first be devised a suitable test procedure, and this can be arrived at only by further analysis-which is another word for organized and systematic thinking with relevant technical considerations taken into account. From this analysis, there emerged the following requirements:
1. The tests must be done with different subjects, who must not know they are being tested and must not observe each other performing the descent of the stairs-else their performance will be affected, and so will not represent “normal” descent of the stairs by an unwarned person.
2. The descent is dangerous because of the shallowness of the treads, and there seems to be no practical way of eliminating the danger without creating special conditions-such as the use of safety nets-which will alert the subjects to the fact that something unusual is going on, and so prejudice the tests.
3. If stumbling actually occurs during the test, the effects will be quite variable. In many cases, the stumbling individual will regain his balance before injury occurs. To be indicative, the tests must be repeated with enough subjects so that injury, or near-injury, occurs in some cases-and this is of course unacceptable.
Thus, in the absence of reliable injury-data based on relevant accident records, the subject danger can be evaluated only by analysis and not by testing. Similar reasoning can be applied to many different cases, in spite of differences in detail.
Engineering analysis is not always as easy as in the above case. As an extreme example, consider the structural design of a skyscraper, which requires involved and sophisticated calculations (whether done by computer or otherwise). Here again testing is impractical, and prior experience is of little value unless gained from similar structures, which have been in use over an extended period of time. Thus, again the role of analysis is predominant.
Having said this, we must mention certain instances in which testing is indeed indispensable. One case involves the coefficient of friction (as discussed earlier under Slip and Fall Accidents). This can be determined only by testing, and not by analysis. Another case involves the strength of different materials, which again must be determined by actual test. However, once such determinations have been carried out, it remains for analysis to incorporate the results into a comprehensive argument relating to, for example, a structural failure or a slip-and-fall accident.