b. Matters of employment.

Lack of employment, 12-37.67; insufficient employment, 0-14.47; poorly paid employment, 0-10.5; unhealthy and dangerous employment, 0-.6.

c. Matters of personal capacity.

Ignorance of English, 0-1; accident, 1.2-5; sickness or death in family, 13.75-26.50; physical defects, 1.28-7; insanity, .25-1; old age, 0-7.

C. Not classified:

Large family, 0-4.5; nature of abode, 0-1; other or unknown, .35-10.

What can be said of the scientific value of these studies? The case-counting method, dealing as it does with actual cases, with the "facts," seems at first sight to offer the most direct line of approach to the problem. Its statistical appearance arouses the hope that the effect of various causes may be measured and the importance of causes disclosed by quantitative comparison. Upon the important causes ameliorative efforts could then be concentrated, and the practical value of this method of study be demonstrated. Closer examination, however, will show that suggestive as the study of concrete cases of destitution may be, nothing is gained by the application of statistical methods. In other words, the relative weight of causes of poverty cannot be established by counting

cases.

In the first place the data used are cases of dependency, not of destitution. Agencies of relief are not yet so discriminating as to exclude from assistance all who are not destitute and none who are. The number of dependents therefore does not coincide with the number of the destitute. Assuming, however, that the correspondence between these two classes is sufficient to permit inferences drawn from the study of one to be applied, without qualification, to the other, we are confronted with a more serious difficulty. Mere number of cases attributable to a given cause is not an accurate index of its effect. A hundred slight cases obviously cannot be regarded as the equivalent of a hundred serious ones. It will be neces

sary, therefore, to ascertain the degree or amount of destitution in each case due to a given cause, if the total effect of this cause is to be summed up in such manner as to make possible comparison with other causes. Comparison of the effect of different factors would be a simple matter if (1) each case were due to only one cause and (2) there were some method by which degree of destitution could be measured. Usually there are many causes for each case. The method suggested by Professor Warner and employed by Mr. Simons of assigning ten units to each case and indicating the relative force of each of the causes at work by its proportionate share of these ten units would be only a very rough index of the proportion of the destitution of a particular case due to a given cause, and fails to allow for the differences in amount of destitution of the different cases. The degree of want in each instance, however, cannot even be approximated. If the only ideally correct method of indicating the relative weight of causes be insisted on, the case-counting method becomes impossible.

If it be granted that it is legitimate to follow Warner, Simons, and others in treating all cases as of equal value, the results are yet inevitably so incomplete and subject to such large allowance for possible errors as to be worthless for scientific purposes. First, be it observed that not all the destitute of any community or country are taken, but only certain groups, as, for instance, charity organization society cases and almshouse inmates. Out of these groups can be selected only those individuals or families concerning whom the recorded information has an appearance of completeness. Transient cases cannot be considered. The group of cases investigated constitutes a small fraction of the whole number of the destitute, and may be not at all representative. Moreover, the different studies of this kind cannot be combined. Between the two groups hitherto studied, almshouse inmates and charity organization society cases, there are important differences. The families or individuals that the societies have under treatment for some length of time and concerning whom, therefore,

the information gathered is sufficient to permit an entry in the "statistical blank," are ordinarily of a distinctly higher grade than the "sodden driftwood" of the average almshouse. If the numerical data concerning these two groups are to be combined, each group must be "weighted" according to the percentage it constitutes of the whole number of the destitute. Owing to insufficiency of data and differences of classification and terminology, such combination of results is, and for a long time will be, unattainable. The study of charity organization society cases only is pursued, in this country, in such manner as to promise results, and to this group the following discussion of the case-counting method will be confined. That it is difficult to secure adequate information even concerning these cases no one who has had experience in investigation and relief work can deny. The subjects of investigation, the human beings designated as cases, may have reasons for deception and concealment, and at all events their testimony concerning themselves, one of the main sources of information, is biassed and incomplete. It should further be observed that "investigations" or "visits" are made primarily not for a scientific, but for a practical purpose, by overworked, hurried, and sometimes inexperienced agents. A very small number of facts concerning an individual or family may be decisive as to the practical measures to be taken. These entered on the "record card" do not present to the statistician looking for data to tabulate a sufficiently complete view of the condition and history of the family or individual to warrant inferences in regard to the cause of distress. Many a practical worker, seeing how decisive facts may remain undiscovered till almost the very close of the investigation or course of treatment, must have been haunted by the thought that perhaps one hour more devoted to some case might have revealed facts overthrowing the conclusions reached, and that serious mistakes may have been committed. Are we not all convinced by our own experiences that the decisive factors in the lives of many are so elusive that years of acquaintance may fail to disclose them? Somewhere, however, because of the merciless flight of time,

an investigation must be closed and the incomplete record allowed to stand.

Some factors, because not easily discovered, are credited with too small a proportion of cases, and, unless these cases are assigned to the caption "Unknown," the percentages left to other causes are thereby unduly increased. In Warner's table "crime and dishonesty" are credited with 3 per cent. from Cincinnati in the year 1891-92 (but only .7 per cent. for the previous year), and from other cities are not held to account for more than 1.5* per cent. in any year, and for most years no figure appears under this head. These numbers certainly give no adequate idea of the many who fail to keep employment because of petty acts of dishonesty or because they are felt to be unreliable. Concerning the numerous suspected but not proven acts no testimony whatever is likely to reach the charity organization society visitor. Even flagrant offences are not willingly disclosed by those who have detected them. The inquirer is frequently impressed by the unwillingness of employers to disparage the character of persons they have discharged and would not be willing to employ again. Another striking example of a forgotten cause is immorality. No percentage whatever is attributed to it in the charity organization society cases, and the heading "Immorality" in Warner's table serves no purpose except to include Mr. Booth's data concerning English paupers. Health and character are so undermined by this cause that unemployment and destitution are highly probable results. That it is wide-spread there are grounds for suspecting. Certain evidence, however, is not to be obtained. In spite of widely prevailing cynicism the charity visitor seeking information finds a remarkable reticence. Men and women are only suspected, not convicted, of sexual license. What charity worker would venture to set down such a suspicion in black and white? How many there are that would hesitate to record even certain knowledge! The record cards of charity organization societies, therefore, contain no evidence of this wide-spread and active cause of poverty.

*The highest percentage, it will be observed, in Professor Lindsay's table.

And now we come to the fundamental difficulty of the casecounting method. It has employed data for only a small proportion of the world's destitution, and incomplete information at best in all cases. Yet, if all the facts open to observation and record were given, their interpretation would meet insuperable difficulties. How is one to pick out a "principal" cause from a tangle of interacting forces? Take this entirely possible instance suggested by A. M. Simons:*

"The husband, a not very competent workman, and an occasional drinker, is thrown out of employment by the stopping of the factory where he has been working. A child falls sick; owing to defective drainage, and this unusual expense causes him to allow his trades-union dues to lapse just before a period of general financial depression. Discouraged and tired of 'looking for work' and his resources exhausted, he applies for charity. Is the 'cause of distress' lack of employment, incompetency, intemperance, sickness, bad sanitation, tradeunionism, or 'general social conditions' beyond the control of the individual?"

How far back of the immediate cause is it permissible to go? Warner holds that the method should deal only with immediate causes, and therefore objects to the inclusion in the list of causes of "pauper association and heredity" and of "nature and location of abode." "Both of these," he states, "are by their nature predisposing causes rather than immediate or exciting causes; and it is confusing to mix the two." Dr. Ayers,† however, in discussing the new blank prepared by the committee of the National Conference, advises that:"In cases where it is doubtful which of several causes should be indicated, some within and some outside the family, the emphasis should be thrown upon the primary cause, st it is surely known and can be stated. A suggestion from the Baltimore Charity Organization Society may be safely followed: 'Give the cause that is farthest back provided you really know it!'" Evidently there is no agreement as to the object of the inquiry.

*American Journal of Sociology, vol. iii. † Charities Review, December, 1898.