(Witness: Neill.)

only a willingness but a desire to do it if we could arrange it in any way so that there would be no overlapping. During this past summer, for the first time, we began an investigation we have not yet completed it into the question of the scarcity of farm labor at this time, to see to what extent immigration is meeting the demand for farm labor. That, as far as I know, is the only time the Bureau bas ever had anything to do with farm labor; and that is being done with a very small investigation, and only occupies the time of practically

one man.

The CHAIRMAN. So that, as far as that item is concerned, there would be no duplication of work between your Bureau and that of the Agricultural Department?

Mr. NEILL. No; none whatever.

The CHAIRMAN. Have you had occasion to investigate to any extent the question of statistics and their gathering and utilization by the various Departments of the Government?

Mr. NEILL. Principally in connection with our own Department, the Department of Commerce and Labor, and in order to avoid duplication we have done this in that Department: The Secretary has appointed a committee which is charged with the responsibility of seeing that there is no duplication of statistics, and that committee is to meet from time to time.

The CHAIRMAN. When you say "no duplication of statistics," you use that term with reference to other Departments of the Govern ment?

Mr. NEILL. No; within our own Department.

The CHAIRMAN. Oh, in your own Department?

Mr. NEILL. Yes; you see we have no jurisdiction outside of that. This committee meets from time to time, and each bureau of our own Department which does any statistical work submits a statement of the work which it has undertaken or is about to undertake. The committee goes over all that, and if any of it seems to be duplicated work the matter is taken up with the two bureaus, and, if necessary, will be referred to the Secretary to determine what shall be done about it. We endeavor to avoid, and I think succeed in avoiding, duplication in our own work.

I might say here that there was a great deal of discussion last year, and perhaps more the year before, over the question of duplication in statis ical work; and there has been

The CHAIRMAN. You mean duplication relating to the Departments generally?

Mr. NEILL. Yes; the Departments generally; and there has been, I think, some misunderstanding on account of what appears to be similar work, when in fact it is not similar at all. Take, for example, the investigation which we recently undertook into convict labor. The Census Bureau intended-I do not know whether it was able to carry out its investigation or not-the Census Bureau, at the time the Bureau of Labor planned the investigation, also planned an investigation into convicts. The work that we did they could not do, and the work that they did we could not do, for this reason: Their work in all cases constitutes an enumeration. They have to take every individual. It is work that must be done quickly, rapidly, and requires a large number of men for a short period of

(Witness: Neill.)

time. Our work, on the other hand, requires a much smaller number of men.

The CHAIRMAN. That is for the purpose of getting contemporaneous figures, I suppose.

Mr. NEILL. Yes.

The CHAIRMAN. All relating, so far as may be, to one point of time?

Mr. NEILL. Yes; for example, they simply took up the question of convicts, and their na ionality, and masters of that kind. We took up the question of the general output and its influence on the outside market. For ins ance, one of our men might be in a penitentiary for three or four weeks making a careful, de ailed study of the books, going into the whole question of the cost. We went into the question of what the average output per convict was during the year, the averge cost of maintenance, and those things that bore on the industrial end of the problem.

The Census Bureau is not equipped for and can not do that kind of work. On the other hand, we are not equipped for and can not do and would not undertake the elaborate kind of work that the Census undertakes. You may put it this way: Our work is intensive work, and theirs is extensive; and although we may cover the same subject, we cover it in entirely different ways. The work that we do, I think it is perfectly fair to say, the Census could not do, is not equipped to do, and can not be equipped to do, it is so far from their general line of work. On the other hand, we could not begin to do their kind of work; and if we attempted to expand the Bureau to do work of that kind it would simply unfit us for the more intensive, what you might almost call microscopic, class of work we have to do.

The CHAIRMAN. That is, it is not practicable, in your judgment, for one man to undertake to do both of those things at the same time?

Mr. NEILL. No; it is not so much that as it is that the whole organization of the office is so entirely different.

Let me put it in this way: Every five years, in taking this manufacturing census, the Census Bureau undertakes to get the wages that are being paid in manufacturing establishments. They have to get that information from every establishment in the United States. They prepare schedules. Those schedules are left with the various manufacturers, and, as I understand, they are filled out by the manufacturing establishments and returned to the Census Bureau. The questions they ask are very simple. For example, they will ask the total amount of wages paid in that establishment during the year. That is in the books, and it is a very simple matter to get. There can not be any mistake about giving it. They get that information from every manufacturer in the United States.

The Bureau of Labor endeavors to maintain an annual table, showing the changes in wages from year to year in certain occupations. We can not begin to cover every establishment. We take certain representative establishments, and our agent goes over there, and goes over the pay roll of the establishment. One of our men may be in one place for thirty days getting that material, going over their pay rolls carefully, looking over them, especially where

(Witness: Neill.)

they pay by piecework. He may work there for a period, say, of thirty days; and after going down that list he does not know which of these men's earnings represents full time, and which does not. So he will ask the superintendent of the division, for example, what is the limit below which a man, who could not earn up to that limit, working full time, would not be kept. The superintendent says "So much." All right; he will then go down and strike all those off the roll, as being plainly men who have not worked full time. After he has struck out all those who can be struck out under this general rule, he will get the superintendent of each division to go over that whole roll with him, and take the particular men in charge of smaller groups, and have them indicate to him which people on that roll are not credited with full earnings that is, those who did not work full time. After that is done he will take the remaining men, and take their actual earnings for the period, say, of thirty days or two weeks, as the case may be. Then, by dividing that up, he gets the actual earnings per day and per hour for his records in that given occupation.

It could be said, "The Census Bureau, for example, is taking statistics as to wages and you are taking statistics as to wages.' That would be perfectly true, and yet the work is absolutely different. As I say, it could not be done by the same bureau. A bureau equipped to do this kind of work could not undertake the kind of work the Census does, and the Census could not get down to this kind of work without having an absolute and separate division there to do it, and the head of that division would have to know as much about it as the head of our Bureau would know or its staff of men. What I mean is that when you come down to the detail work in the direction and supervision of this kind of work, if you attempt to combine it, it results in putting one man at the head in the case of whom it is simply impossible that he shall be an expert in all these fields, and the men who, under our method, are real experts would simply be cheap men who could not possibly supervise and direct the work.

The CHAIRMAN. I do not quite see, Mr. Neill, why your man, for instance, who must be a pretty intelligent man to get these results you are speaking of-that is, you intend to have him so, of courseMr. NEILL. Yes.

The CHAIRMAN (continuing). I do not quite see why he, for instance, while he is making an examination for you can not with very little difficulty, inasmuch as he is on the spot, get the larger and more general information that seems to be collected by the Census Bureau. Mr. NEILL. The reason is that we probably do not cover one plant in a hundred.

The CHAIRMAN. You would not cover the whole field, but as to the plants that you do reach, of course you could get the information with very little difficulty?

Mr. NEILL. Yes; that is true in the case of the few plants that we reach, but the question is this: The kind of a man that we have to use is, you can see, an expensive man. The kind of man who leaves a schedule and brings it out is a cheaper kind of a man. You could get almost any kind of man to distribute and collect those schedules under the direction of a competent supervisor for that city or that district. You can see, of course, that to take men of the class that we have to use and send them to every place in the United States

(Witness: Neill.)

would require an almost inconceivable expense, and the kind of men that can only do that kind of work we could not use at all. The CHAIRMAN. Yes; I see.

Mr. NEILL. Take this case now: Last year we were making a study of certain agricultural conditions in the West, or, at least, of the question of farm labor in the West. I took one of the best men we had in the Bureau and sent him out with directions to go through the wheat belt, for example, and make a very careful study there of the difference in the demand for labor, say, in wheat raising between the various periods of the year; that is, I directed him to find out how much increase there was in the demand for a certain short period, to find how long that period was, how long it lasted, what wages they had to pay, where they got their men from, what those men did during the other periods of the year, and, beginning up in Minnesota, to follow that thing right down to the southern end of the wheat belt, studying the problem, not merely as an agricultural but rather as a general industrial problem. It was a problem in casual employment, you see, and the man had not only to go into the question of farm labor, but he had to follow those men through, find the various kinds of work they were employed in, find whether they were local people there or whether they were men that were migratory and moved back and forth, and it required a good deal of knowledge and experience outside of agricultural work. The agricultural end of it was really the less important end; it was an incident to the larger study.

The CHAIRMAN. Yes; but the wage factor was the dominant feature of the whole thing, was it not?

Mr. NEILL. No. The question there was to determine what were the difficulties in the way of farmers getting labor during the harvest period-to find why they could not get it at that period and to find what would be necessary in any given locality to get a permanent supply of labor there for the farmers. Of course, that involved going into the general question of mixed industries in such a way that at the time the farmer needed the man he would not be in the other industry, because that industry would be idle, you see; to find where those men were being drawn from. It was not simply a question of what wages they were being paid; it was a wider investigation than that, although, of course, at bottom the fact was that the farmer could get them if he would pay enough wages. But, for example, Mr. Littlefield, it was not merely a question of offering $3 a day for a short season for a man that was only getting $1.50 somewhere else, because a man would not come for $3 a day for three weeks if he felt that he could not get his other position back."

The CHAIRMAN. It was not so much a labor problem as it was an industrial problem?

Mr. NEILL. I use the term "labor problem" in the sense of that aspect of an industrial problem.

The CHAIRMAN. And, of course, labor is involved as one of the features?

Mr. NFILL. Certainly.

The CHAIRMAN. But you were treating this from a larger sense?

Mr. NEILL. Certainly.

The CHAIRMAN. And a broader point of view, and labor was only one of the minor factors?

(Witness: Neill.)

Mr. NEILL. Yes; and yet it might be said that we were making an investigation of farm labor, for in one sense the question of farm wages came into it, you see.

I myself believe thoroughly in concentration and centralization of this work wherever it can be done. But there is this danger, Mr. Littlefield: That a concentration that seems to effect good results may, when you come to study it much closer, be perfectly demoralizing to the actual efficiency of the various kinds of work.

The CHAIRMAN. Of course, no one would knowingly submit a centralization proposition that destroyed efficiency.

Mr. NEILL. I mean to put this down as a general proposition: A man will say to me, "Are you a statistician?" "Yes." Will you take this thing up?" I will say: "I do not know anything more about it than the man in the moon." Statistics are simply one way of expressing results attained in various lines of study. A man has to have some knowledge of the principles of statistics and the methods of handling them; but more than that, he has to be somewhat of a trained man in the particular field that he is investigating.

The CHAIRMAN. In other words, your proposition is that a man can not efficiently conduct a statistical investigation or examination unless he has information and knowledge as the result of study and experience of that subject?

Mr. NEILL. Certainly. In the first place, he has to know the subject thoroughly in order to know how to properly investigate it. His statistics are merely the method of expressing the results of that study. It is one of the methods; it is not the only method; it is only a partial method. And, as I say, the difficulty is, I find, in my own Bureau, that it takes all the time and all the intelligence I am capable of to wrestle with these problems, and I am not at all sure that I am meeting them successfully or wisely. If to that work were added a more diverse field of work, I should simply give up in despair.

So that it seems to me that if the Government work is going to be efficiently done, it is going to take a high-grade, well-trained man at the head of every department of statistical work which deals with a different topic from any other one. As I say, in census work there is a wide field, because, if it is a question merely of enumeration, the methods of enumeration are practically the same in all lines of work; and if it is a question of census, which is a kind of enumeration, one man can properly direct an enormous number of census undertakings. But if it is going to come down to getting statistics showing the vital questions not connected with industrial problems, but vital problems, or social problems

The CHAIRMAN. That is, the relation one factor in the equation has to another?

Mr. NEILL. Exactly; the man has to be a trained man in that particular field.

The CHAIRMAN. Right there; have you, in your Bureau, made from time to time analytical and concrete summaries of the work that you do in all these various branches?

Mr. NEILL. In what way do you mean?

The CHAIRMAN. Do you summarize, for instance, the results of your investigations so that in connection with the statistics of your