612 APPENDIX. 



It is not, however, always as easy as it might seem to know what statistics 

 really teach when we have them. For example, statistical tables and diagrams 

 which follow show that the purchasing power of gold has enormously increased 

 during the past thirty years, and that the farmer who for that period has kept 

 alive an old mortgage, hy renewals, must now dispose of far more produce 

 to pay the deht, than would have paid it when originally contracted. On the 

 other hand, tables of current rates of interest, for the same period, equally 

 authentic, and indeed with less liability to error, show that in 1896 it required 

 twice as much money invested in gilt-edged bonds, to produce an income of 

 $1,000, as would have produced that income in 1870. To arrive at a correct 

 conclusion these two tables, possibly, upon analysis, not so contradictory as 

 they seem, must somehow be reconciled. A skilful controversialist, by using 

 only one set of these tables, can make an argument, unanswerable by a popular 

 audience, which will show that debtors are suffering very severely by the appre- 

 ciation of money, and immediately thereafter take the other table before another 

 audience, and with equal conclusiveness, prove that the poor creditors have 

 practically lost half their debts. One can prove almost anything by statistics, 

 so long as he is permitted to select his tables. Imperfect as many of them are, 

 statistics are invaluable for the study of economic problems, as long as they are 

 intelligently and honestly employed, and all the facts, instead of only part of 

 them, are considered. This is sometimes very difficult. Not only is our 

 information limited, but our mental powers, which largely accounts for the 

 zeal with which good men difler on economic problems. It is not everyone who 

 has the intellectual grasp to compass great subjects. 



Statistical tables make a very dreary looking page for the ordinary reader, 

 and are, of course, not intended to be read by any one, but to be referred to 

 when it is desired to make use of the fiicts which they show. For the purpose 

 of quickly conveying the facts of statistical tables to the reader, it is a common 

 practice to employ diagrams, drawn to a scale, upon which the fluctuations of 

 the tables are shown by lines. This is a very useful practice, as it enables the 

 eye to take in at a glance the lesson of a long column of figures. Like many 

 other good things, however, the method may be easily abused, and, in the 

 hands of designing persons, be made to convey impressions not at all warranted 

 by the facts. The trick is very simple, and consists merely in making the 

 vertical spaces of the scale large or small, in relation to the horizontal spaces, 

 according as the draughtsmen desire to convey the impression of great or small 

 fluctuations. For example, the ratio of silver to gold previous to 1873 was 

 constantly fluctuating, and yet the variations were so small, that popular 

 discussions of the relations of silver and gold usually take no account of them, 

 and silver and gold are often, and properly enough, represented on diagrams 

 by an identical line, for some years previous to 1873. It would bo very easy, 

 however, for a series of those years, by constructing a diagram in which vertical 

 spaces of one inch should represent one per cent of variation, and horizontal 

 spaces of one-half inch represent years, to convey the impression that the 

 ratios of the two metals were fluctuating within very wide limits. The increas- 

 ing use of these diagrams in popular discussion, sometimes very disingenuously, 

 makes it proper to give this caution. Economists and students, and political 



