Page:Wealth and Income of People of United States.djvu/12

viii The reader of this book will do well to bear in mind that there are two broad varieties of concrete statistical studies: first, those depending on the correct counting of many individual items; second, those consisting of estimates based on counts made by others. This study is distinctly in the latter class. The estimates have been made, in most instances, from fragmentary material gathered by different persons at different times and for different purposes. In some cases, the original counts (principally by government officials) were doubtless faulty, but only when the errors were evident has the author attempted to go behind the returns and criticize the validity of government reports. Frequently, estimates have been made on the basis of assumptions that are possibly decidedly erroneous. In some cases, details were filled in by the use of careful guesses based on general information only.

The critics will immediately assert that such methods are useless and that the results are not worth the paper upon which they are printed. To this view the author takes exception. The primary value of statistics is usually due to relative rather than to absolute accuracy. It is believed that the figures cited are, in most instances, sufficiently accurate to justify fully the conclusions made concerning relative sizes, amounts, or changes. An effort has been made to state fairly the probabilities of error