得苗調査における経済的標本抽出単位の型について

その他

The type of unit affects usually both the accuracy and the cost of a sampling survey therefore, it may be veryinportant to find the optimum type of unitin the sampling survey of seed-bed.
To find the economical unit type, the authors made some reserch works in the nursery of Simane Forest Experiment Station by the survey of Kuromatsu and Sugi seedlings.
For a comparative study, we prepared eight type of unit, as shown in Table 1 and Fig. 1. Following the accurate population total number of seedlings was as certained in the areas by a census, we could get the variance of population per unit due to each types. Next were estimated the cost of sampling works by a time study (last row of Table 3 and 4).
In studes of this type the item to be estimated is population total number of seedlings. If finite population correction is ignored, the variance of the estimated population total is
(N^2_u S^2_u)/(n_u)
Since this variance is to be the same for all units, the smallest unit (Type 1) was chosen as a standard. Then the values of the other n that give the same precision as the type I are obtained from above equation, as shown in the first lines of Table 7 and 8.
Next step is to find the comparable sample size in term of type I. (second line of same Tables).
The cost ot taking n of the type I may be expressed as (Kuromatsu)
c_1 = (n_1)/(47.2)
the values of the other c that compared with the type I are shown on the third lines of same Table.
The type IV unit appers the best to be used for Kuromatsu and Sugi seedlings. The last lines of Table 7 and 8 shows the reciprocals of the costs compared with the Type I (0.1×0.1m) taken as 100. In the tables these figures have been called relative net precision.