"Reproducibility" refers to the percentage of accuracy with which the scores on the individual items making up a scale may be predicted knowing only the total score (ratings of selected occupations in this case) of all occupations in that scale. A reproducibility of 80-85 may indicate adequate scalability if satisfactory levels of rigorousness on the other criteria are met but generally would reveal a quasi-scale. An index of 85 or, perhaps better, 90 indicates satisfactory scalability. Reproducibility indexes, must, however, be interpreted in the light of the minimum marginal reproducibility which represents the per cent of ratings falling at the modal value. Thus, the lower the minimum marginal figure and the higher the reproducibility, the greater is the improvement in predictability of responses to the items over that which would result from chance alone.
Herman P. Miller
Table 1 shows the occupational distribution of employed persons by sex, as well as the mean and the median income received by persons in each occupation group. . . .
There is a very large gap between the incomes of the highest paid workers like doctors and dentists and the lowest paid workers such as farm and nonfarm laborers. However, for most occupation groups, the income differentials are quite narrow. Nonfarm laborers, for example, had a median income of $2300 in 1951 or about $44 per week. The corresponding median for service workers was only slightly higher ($2500 or $48 per week), and the median for operatives was $3100 or $60 a week. The craftsmen, clerical workers, and sales workers each had median incomes ranging between $3400 and $3700 for the year or between $65 and $70 a week. The use of over-all averages based on