Johnston] NAVAHO POPULATION 175 



The death rates of Navaho children under 5 also suggest that a 

 significant decline has occurred, at least in the period 1950 to 1955. 

 The lower of the two death rates calculated for this age group in 

 1950 is 32.4 per 1,000. The range of chance variation associated with 

 this rate gives a lower limit (at 95 percent confidence limits) of 30.9 

 per 1,000. The upper limit of the corresponding range in 1955 is 25.9 

 per 1,000, thus indicating a real decline. 



The mortality of the Navaho population aged 45 and over, as ex- 

 pressed in the death rates shown in table 35, displays a reversal of this 

 general downward trend, rising from 11.9 per 1,000 in 1945 to 13.7 

 per 1,000 in 1955. However, the overlap in the ranges of chance varia- 

 tion that are associated with these two rates indicates that the signifi- 

 cance of the observed trend is highly questionable.^* 



In summary, it can be concluded that Navaho mortality, as reflected 

 in the death rates shown in table 35, has declined in the period from 

 1945 to 1955. This decline is especially noteworthy since 1950, and 

 is stronger in the female segment of the population than in the male. 

 Comparing the lower limits of the ranges of chance variation in 1945 

 with the corresponding upper limits in 1955 suggests that the ap- 

 parent declines are significant. It must be emphasized, however, that 

 the continuing possibility that Navaho deaths are underreported, to- 

 gether with the persistent uncertainty with regard to the size of the 

 base population, combine to throw considerable doubt upon these 

 conclusions. 



In concluding this analysis of the reliability of the available data 

 on Navaho vital phenomena, a brief summary of certain measures of 

 the underregistration of births occurring in the Navaho area is in 

 order. In connection with the decennial censuses of 1940 and 1950, 

 nationwide surveys were conducted in order to ascertain the com- 

 pleteness of birth registration. The procedure in both tests was 

 essentially the same. In 1950, the census enumerators recorded the 

 names of all infants born during the 3 months preceding the census 

 date on a special "infant card." These names were then matched with 

 birth certificates on file with the respective State registrars of vital 

 statistics. The 1940 test differed from the 1950 survey in that it 

 covered births occurring during a 4-month period prior to the census 

 date, and included infants born during this interval but dying prior 

 to the census date (Hadley, 1952 a) . 



Since all pertinent records included information on the race and 

 birthplace of the infant, separate statistics for Indian births in the 

 area of the Navajo Eeservation could be compiled. The basic measure 

 used to summarize the results of these tests is the "percentage of 



1* One might speculate, in this connection, that the apparent rise in the mortality of 

 Navahos aged 45 and over is merely a reflection of improved reporting of deaths occurring 

 among older Navahos, without a concomitant improvement in the corresponding base 

 population figures. 



780-568—66 13 



