146 BUREAU OF AMERICAN ETHNOLOGY [Bull. 197 



lation that is under 15 years of age (for both sexes combined). This 

 convenient measure is a useful indicator of the general level both of 

 the fertility and the mortality of a population. A percentage of 40 

 or above may be said to typify the age distribution of a population 

 experiencing high fertility. Like the median age, this measure is 

 naturally insensitive to minor differences in the age composition of a 

 population. Hence the wide fluctuations that are evident in the per- 

 centage under 15 as shown for the 25 Navaho age distributions in table 

 28 are a further indication of the serious biases which exist among 

 them. 



Setting aside the extreme values for this percentage among the 

 Navaho age distributions, we find that they tend to vary between 44 

 and 49. This central range of values would, of course, imply a median 

 age in the neighborhood of 16 or 17 years, which is in fact observed 

 among the same distributions. 



The remaining summary characteristics of the 31 distributions 

 presented are shown in table 28 under the heading "measures of ac- 

 curacy." These measures were developed by staff members of the 

 Population Division, United Nations, in order to facilitate comparisons 

 among different age distributions in terms of the internal inconsisten- 

 cies of each (United Nations, 1952) . These measures provide a numeri- 

 cal score which expresses the degree to which the given age distribution 

 departs from a smooth progression of values with respect to the char- 

 acteristics measured. The "joint score" is simply a weighted sum of 

 the separate scores, giving a weight of three to the sex ratio score and 

 a weight of one to each of the age ratio scores. The joint score thus 

 provides a single summary measure which describes the "goodness of 

 fit" of the component parts of a given age distribution. 



In order to improve the comparability of the summary measure, tak- 

 ing into account the greater chance variation in age-specific values that 

 is associated with populations of smaller size, a further modification 

 was introduced. This modification has the effect of reducing the joint 

 score by an amount which is inversely proportional to the size of the 

 population. The resultant score, which is also shown in table 28, is 

 termed the "adjusted joint score." ^^ The approximate magnitudes of 

 scores that can be associated with data of different quality are as fol- 

 lows: Sex ratio scores below 2 and age ratio scores below 3 (implying 

 joint scores below 12) reflect data of excellent quality, provided the 

 age distribution of the actual population is not marked by unusual 

 abnormalities. Data of fair to good quality would be indicated by 

 sex ratio scores of 2 to 4 and age ratio scores of 3 to 6 (implying joint 



^ The value of the adjustment, which is subtracted from the joint score to obtain the 

 adjusted joint score, is given by the formula : 



S600 where P is the size of the total population. 

 -fP' 



