TX contains the sum of the size variable for the stratum to be sampled. 



TS,TP contain components of the sample variance estimate defined on page 4, 



NRS contains the number of units drawn from the stratum to be sampled 

 (NRS = n) . 



N contains the total number of elements in the sampled stratum. 

 KR 



LR contain odd random integers used by the random number generator. 

 MR 



CALCULATING ESTIMATES 



Population Total 



The population total is estimated from a pps sample by 



. , ^ ■ ^ z=n+l 

 %=l ^=l 



in which 



n is the number of units measured in the sample. 



I is the number of units assigned to the stratum of too-large 

 units to be measured with absolute certainty. 



Y. is the variable of interest for the ith measured unit. 

 ^ 



N 



P. = n X ./ l X., the probability with which the ith unit was selected for 



measurement where x. equals the measure of size of the ith unit in 

 the sampled stratum. P. equals unity for units in the too-large 

 stratum . 



The last summation would not contribute to the variance of Y. 



Note that depends on n. Accordingly, inadvertent change in n after the P. 

 is calculated should be avoided. ^If u sample units are not measured for reasons 

 independent of their sizes, then Y could be estimated by 



n y I 



n-u up. L-' 1 

 . . ^ . , 

 ^=l ^=n+l 



in which = for the omitted sample units. 



3 



