once that $2,000,000 of capital cannot produce $4,000,000 worth of goods. The conclusive answer to his statement is that such is the return he has made. He then tells you that census statistics are vicious, and he is correct in this particular instance. If you attempt to correct this error by asking him to account for his whole capital, that paid in and that borrowed, in fact, all the capital which he uses to secure a given product, he antagonizes the work of the state by say. ing that you are prying into his private affairs. He does not perceive that he has deceived not only himself and his employés but the public generally, and that every argument, conclusion, or deduction, based upon the statistics reported as giving the capital invested in manufacturing, is false in all its elements and consequently exerts a vicious influence in every direction."

It has been the plan of social economists and labor reformers to add the value of stock used and wages paid together, subtract this sum from the value of goods made, and consider the balance as the gross profit. This gross profit has then been compared with the capital invested and a supposed ratio of profit obtained. But inasmuch as the statistics of capital invested have shown too small an aggregate, so this ratio has been correspondingly too large. Double your capital invested and you at once cut down your ratio of profit one-half. Until the actual capital, both cash and credit, is ascertained the ratios of profit must be set aside, and when properly ascertained these important ratios must be recast upon the new lines. Thus statistics will furnish the basis for the consideration of the most important question of our times -the question of profits, and its no less important corollary-the distribution of wealth.

"Another error which has been made relates to average earnings, as ascertained by dividing the aggregate wages paid in any industry by the number of employés involved. In all censuses where industrial statistics have been a feature in this country, until the Tenth Census, the question relating to employés has simply been the 'number of hands employed.' What is the number of hands employed? Can you ascertain the true quotient which shall represent average wages by dividing the total amount of wages paid during a year in any given industry by the total number of hands employed? I assert that you cannot, for the number of hands employed returned under the old form of inquiry simply meant the number of employés borne on the rolls at the close of the census year, while the total wages paid represented the aggregate amount disbursed in the form of wages to all who had been employed at any time during the same year; the number of people returned might represent a much larger number at the close of the year than were employed at any other period, or the reverse, a much smaller number. In any event, and with the utmost care on the part of the manufacturer, it would be impossible in the answer to such a form of inquiry to state anything more than the number employed at the end of the year, or at some stated period.”

Instead of these unreliable wage statistics the classified wage, as explained in Chapter V., and the actual wage for each branch of occupation must take its place. There should be no margin of uncertainty in wage statistics. They should be actual, and absolute, and as worthy of credence as the reconstructed statistics of profits. Then capital and labor can be placed side by side and judged by their fruits. That this work is under way is shown by the following:

"The division of labor which marks this age as distinct in all its industrial features from previous periods constitutes a very interesting study. The attempt has been made in the Census of 1885, for the Commonwealth of Massachusetts, the reports of which are now rapidly approaching completion, to ascertain and report to the finest degree the various but distinctly designated divisions of labor, so far as the occupations of the people are involved. Incomprehensible as it may seem, this classification shows over 20,000 distinct designations. This classification, in connection with nationality as shown by place of birth, age periods, and illiteracy, constitutes a fine contribution to sociological knowledge. This has been done in Massachusetts. It may be done for the United States in the future, now that the basis has been laid."

When wage statistics based upon this scheme are secured, we shall hear no more of average wages in any industry, but instead the actual wages paid in each branch of occupation in that industry. Then the high-priced and low-priced workers will be classed each by themselves and not merged in an aggregate in which high-priced or low-priced workers may preponderate, giving, naturally, a fictitious, unreliable, untruthful average.

"Perhaps the greatest difficulty which confronts the federal census taker, and which offers the least possibility of solving problems, is to be found in the agricultural department of the work. The products of various localities differ so widely in their character and kind, that any single schedule framed to comprehend them all becomes in itself impracticable. Yet certain facts, like total value and quantity of the great staple products, ought to be easily ascertained. The question of renting or ownership, of the acreage of the farms, of alien ownership, are entirely within the possibilities of census taking without enlarging the field of operations. Such problems ought to be readily solved. They have been solved for the time being in past censuses. As the conditions vary the problems vary, and the intelligence of the government should see to it that the facts sought should be those necessary to indicate the true relations of things in the agricultural field of knowledge.

From what has been said it is readily seen that there is quite a range of problems, the elements for the solution of which can only be supplied through comprehensive censuses. As yet, however, no co-ordinate effort has been made

between the state and federal governments for systematic work. This question is difficult and yet important. Its importance lies in this reflection, that no member of a state or the national government can intelligently discuss the land question, the solution of the immigration problem, the question of how far educational efforts shall be extended, the tariff problem, the questions of currency, of railroad transportation, in fact, any question of any great state or national importance, without full and constant reference to the tables of the census reports. This being the case, and census work becoming more and more complicated, as all these great questions come nearer and nearer to the people, how important it is that the whole range of census taking should be reduced to systematic forms and methods. The United States cannot do it all, and do it well. The country is too large to admit of extended inquiries through the methods of enumeration. The States should enlist in the work, and they could do it generally without much expense through their established machinery. They should be called upon by the federal government, but aided to the extent of the results furnished, to do certain things at short intervals relative to the enumeration of the people, the ascertaining of products, the ownership of land, and may be half-a-score of leading features, the federal census itself being confined to those things which reach over the whole country and which may be called purely national topics. Such a system, involving the co-operation of the states, would not only simplify labor, but would reduce the burden in many respects and would certainly remove much of the irritation which exists whenever the census is taken; and further simplification should be reached, as I have indicated, through special investigations. All census work, both state and national, should leave out of sight every question which can be brought under the methods of special investigation, in which aggregates are not essential but in which representative facts are just as important as the total facts."

The quotations from Mr. Wright's address cover its main points, and bring out forcibly the great statistical questions of the time. To these, in time, many as valuable points will be added. Industry pictures will be formed showing the composition of each industry-partners, stockholders, salaried workers, and wage laborersmarshalled, as it were, in battle array. Then with statistics of profits and wages we shall get clear views at both sides of the statue of labor. Statistics of trade will be demanded in time, and the dealer will be called upon to disclose for the general good the facts concerning his business. The only way in which the middle-man can be honestly judged is to ascertain the facts concerning him first. If he is a public benefactor, he has nothing to fear. If he is a public evil, he will prefer to be judged on facts rather than opinions-so the facts will have to be secured. At the next National Census a request has already been made that the number of survivors of the Rebellion be ascertained, with cer

tain facts as to their condition. This enumeration should include Statistics is a catholic science and should

all soldiers, in all wars.

always give the facts for the people, and not for a class.

This Chapter could be extended indefinitely, but enough has been written to show that statistics is interwoven with the life of the nation. Its results, like the physician's diagnosis, must precede salutary treatment. The best physicians are specialists. Special statistics reach facts while general statistics only reach numbers.

12

CHAPTER X.

THE TEACHING OF PRACTICAL STATISTICS.

Statistics being a science and destined to have in the future a much greater influence than in the past as regards the progress of mankind, it follows that as the statisticians of to-day retire from active duty or die, others must take the places that they vacate. In other words, the statistician of the present time must teach his science to others.

This teaching may be done in two ways: theoretically in the institution of learning, or practically in the statistical office. We shall endeavor to show hereinafter how statistics may be studied practically in the college or academy.

The subject of the study of statistics in colleges has been considered only in the last few years. Mr. Wright, in a paper read before the joint session of the American Economic Association and the American Historical Association, at Sanders Theatre, Harvard University, May 24, 1887, brought into twenty pages all that could then be said upon the matter. We shall quote liberally from the paper in question, for it puts the case strongly as regards the absolute necessity of making educational provisions for the instruction of statisticians. Referring to the European statisticians, Mr. Wright says:

"America has no counterpart to the continental school of statisticians, whose members have entered their particular field of science after special training by a systematic course of instruction. We have our statisticians, to be sure, but they have taken up their work accidentally, and not as a profession. Men engaged in the practice of law or of medicine, or in the other learned professions, enter them only after careful preparation. Our government trains its soldiers and sailors; our colleges and higher institutions of learning fit men for various special scientific and professional labors, but we have not yet reached the advanced stage of educational work in this country which comprehends administration in its broadest terms. The European has an advantage over