As regards i., we have the wage sheets for some 470 separate weeks, in addition to the complete lists of two very small firms for one and four years respectively; ii., the complete earnings of about 130 hands for periods varying from one to fourteen years.
Owing to the fact that it was impossible to get the complete lists through 1899 for many firms, and that the periods of slackness and full work were not the same in different places, it proved very difficult to handle the wage lists. At last the plan was adopted of getting complete lists of one busy week, one typical week, and one slack week in 1899, leaving the employers to choose the weeks, unless our investigators could make a complete record. In the following analysis we have endeavoured to bring out the salient features of the statistics of each firm separately, and we have then grouped together all the typical weeks, either chosen by the employer or selected by us from the series; and it is believed that this grouping gives an adequate idea of the wages at a time which the trade regards as ordinary.
The earnings of the 130 individual hands is a very valuable and, it may be, almost unique record. Many interesting facts are brought out by their study, and the records should have a place in sociological literature apart from their interest in the present connection.
It has been necessary to make a technical use of averages in collating and tabulating the material, and we offer the following explanations. Where the word "average" is used without qualification, it is the ordinary arithmetic average, obtained by dividing the total by the number of payees. This is the best for general quantitative measurements.
In most cases the median and quartiles and sometimes the dispersions have been calculated. They may be explained as follows. Suppose the wages of, say, sixty persons to be arranged in ascending order, e.g., 5s., 5s. 3d., 6s., 6s. 1d. ... 11s. 9d., 12s., 12s. 6d., then the wage halfway up the list is the median wage; thus, there are as many individuals above the median as below it. The wages halfway from the ends to the median (i.e., fifteenth and forty-fifth from bottom), are the quartiles, so that between the quartiles half the wages are grouped. Thus, if the median and quartiles in the above list were 7s. 6d., 10s. 6d., 12s. 6d., there would be fifteen earning less than 7s. 6d., fifteen more than 12s. 6d., thirty between 7s. 6d. and 12s. 6d., thirty below and thirty above 10s. 6d. For a single measurement of the grouping of the wages about their median, the distance between it and the quartiles is significant: in this example 3s. and 2s. are these distances. The more convenient way of stating this is to express half the distance between the quartiles (2s. 6d.) as a fraction of their average 10s., which is generally very nearly the median. This fraction (1/4 or ·25) we call the dispersion, and it enables us to study the changing character of a group in a very simple and efficient manner.
I.—STATISTICAL VIEW OF THE VARIOUS FIRMS.
FIRM A.—Information obtained.—Wages of thirty-six hands tabulated week by week through 1899.
Total amount paid in wages and total number employed each week, 1885-1899.
The whole wages sheet for one week in July and one week in November for each of these fifteen years.
A. is a firm employing from fifty to one hundred and ten women and girls as folders, stitchers and sewers. The number employed has changed gradually; in 1885-7 there were about a hundred: from 1888 to 1894 the number continually diminished to sixty, and after a brief spurt in the autumn of 1894 to one hundred and fifteen and a rapid fall, has from 1895 to 1899 gradually risen from fifty to ninety.