212 BUREAU OF AMERICAN ETHNOLOGY [Bull. 197 



Stef 4' The sum of the differences between successive sex ratios is 

 obtained, without regard to signs. 



jStep 5. The sum of the deviations of the male and female age 

 ratios from 100 is obtained, for each sex separately, again without 

 regard to sign. 



Step 6. The sex ratio score is equal to the sum obtained in Step 4, 

 divided by the number of differences calculated. In appendix table 1, 

 this score is equal to 89.8 divided by 13, or 6.9. 



Step 7. Each age ratio score is equal to that sum obtained for each 

 sex in Step 5, divided by the number of age ratios calculated. In ap- 

 pendix table 1, the male age ratio score is 141.8 divided by 13, or 10.9, 

 wliile the female age ratio score is 120.9 divided by 13, or 9.3. 



Step 8. The joint score is equal to the sex ratio score times 3, plus 

 the male and female age ratio scores. In the table, this score equals 

 20.7 plus 10.9 plus 9.3, or 40.9. 



Step 9. In order to compensate for the greater chance variation in 

 age distributions fomid among smaller populations, an adjustment 

 factor is obtained by means of the formula : 



3,500 



Vp 



In the table, the size of the population whose age distribution was 

 utilized was 22,377. The resultant adjustment factor is therefore 

 to 3,500 divided by the square root of 22,377, which comes to 23.4. 



Step 10. The adjusted joint score is equal to the joint score minus 

 the adjustment factor. In the illustration, the adjusted joint score 

 equals 40.9 minus 23.4, or 17.5.- 



A word should be added in regard to the alternative uses or inter- 

 pretations of these measures. The primary use of these measures is 

 to provide an indication of the extent to which the given age and sex 

 ratios depart from a smooth progression. As was mentioned in the 

 text accompanying table 28, such departures may either reflect 

 poor data or actual peculiarities in the given population's age distri- 

 butions. A secondary, and somewhat more hazardous use of these 

 measures is to provide comparisons of age distributions obtained for 

 different populations. Wlien these measures are used only for the 

 internal analysis of a single age distribution, the adjustment for the 

 small size of the population is not necessary. However, this adjust- 

 ment is appropriate in comparing the scores obtained for populations 

 of different size, since a smaller population would be more likely to 

 depart from a smooth progression by chance than a larger population. 



2 The derivation, uses, and limitations of these measures are more fully discussed in 

 United Nations, 1952, pp. 59-79. 



