Indiscreet advocates of the numerical method have sought to apply the general results of collections of cases expressed in the language of figures to the treatment of individual cases of the same disease, without making allowance for those differences between case and case which confessedly existed in the collections themselves. They have used a general prin ciple, as if it had been a rigid and unbending rule of action ; forgetting that though, as the experience of assurance offices abundantly tes tifies, the general results obtained from a large number of individual facts may be safely re applied to an aggregate of facts of the same nature, they cannot be brought to bear on a single case, or on a small number of cases, without the greatest danger. Each case must be viewed in practice, first as a generality governed by some large law of prognosis, diagnosis, and treatment ; and secondly, as a specialty demanding a careful consideration of all its peculiarities.
Among the rules which ought to govern the observer in the collection of facts destined to form the elements of averages, there are some of so simple and obvious a nature as to re quire no discussion. Such are, the previous preparation of some simple and available form of register, by means of which the several facts may be committed to paper at the very time of observation, so that nothing may be trusted to the memory ; the careful selection of the facts themselves ; the shaping of the inquiries which may be necessary to elucidate those facts as nearly as may be in the same terms ; the avoidance as much as possible of such leading questions as would be likely to bias the respondent, when the facts in question, like most of the particulars which make up the history of diseases, are dependent upon testi mony; and especially the purging of the ob server's own mind of prejudices and precon ceptions in respect of the subject of inquiry.
The careful selection of the facts which are to form the materials of our averages is by far the most important of these rules, and one which demands a little further consider ation. If we consider the facts we are ob serving in the light of phenomena, or events brought about by a multitude of concurrent causes, it will be obvious that care will re quire to be exercised, not so much in verify ing each phenomenon or event as that of which we are in search, as in ascertaining that all the concurrent causes or conditions arc, or have been, in operation to bring about that phenomenon or event. The absence of a single cause or condition will vitiate the in dividual fact, and impair or destroy the value of our average results. A few illustrations will suffice to show what is here intended. We are anxious to determine the true average frequency of the pulse in adult males, in a state of rest, and as free as possible from the influence of all disturbing causes ; but, either from ignorance or oversight, we count it in differently in every position of the body, and at all times of the day. In this case, our facts
cease to be comparable facts ; for it is well known, that both posture and time of day have a remarkable influence on the number of the pulse. Or, to take another case, we wish to ascertain the influence of some employment upon. health (say that of the letter-press printer); but we overlook the important fact, that in every printing office, two or three very distinct occupations are carried on, of which the most important are those of the com positor and pressman. Not being fully aware of this fact, and of the wide difference exist ing between the two employments, we pro ceed to extract from some mortuary register the ages at death of printers as a class, calcu late the average age at death, and then proceed to group the whole class of printers with that large class of occupations carried on in-doors, with little bodily exertion, to which the com positor alone properly belongs, but from which the pressman is, by the nature of his employment, excluded. In this case, we should have been misled by the common name borne by men following two really dis tinct occupations, and our facts would again cease to be comparable facts. A third and apt illustration is afforded by the Asiatic cholera. We wish to compare two different remedies or plans of treatment ; but we administer the one remedy, or adopt the one plan, at the onset of the epidemic, and the other during its decline. Here, again, our facts are not comparable facts ; for it is one of the well known characteristics of this disease, that it is more severe on its first occurrence than during the period of its decline. The same sort of error would be committed, if one remedy were administered in an early, and the other in an advanced, stage of the attacks themselves. The principle which these illus trations are intended to enforce, is the ne cessity of selecting, as the elements of the same average, facts strictly comparable, or, in other words, brought about by the same com bination of causes. Over the intensity with which each cause acts in individual instances, the observer can exercise no control. His province is to ascertain that the same com bination of causes is at work to bring about each phenomenon or event. If from ignor ance or oversight he fails in this duty, he im pairs the value of his facts, and vitiates his Inferences in proportion to the number and force of the conditions that he has overlooked or omitted.