| 概要 |
Critique on the former method - In order to revise the former method of sample survey let us criticize the former one from the stand point of aim, items, scope, calculation, range of error and expense...s and clarify the reasons why we are planning positively its revision. Aim - In the former method as in the revised one, the principal aims are compute the average wages and to form the wage index. But what do the average wages mean ? If it means the wage level, the survey is to be undertaken necessarily not only on the wage distribution, but also on the structure of wages by industry, occupation, age, number of dependents and sex. Scope - In the former sample survey, all establishments with 75% or more of total employees of each all industries-exclusive of establishments with 10 or less employees and of agriculture, forestry and fishery-are covered. But by this method there occurs, such inconvenient cases where in one industry only the establishments with 1,000 or more employees are selected and in other industry only the establishments with 50 or more employees are sampled. This means that the object of the survey is different by industry and the scope of the survey lacks precision. The idea to undertake a cut-off survey by the percentage of employees betrays the desire to survey, if possible, all establishments and the necessity to exclude the establishments of small size in view of the difficulty of the survey. If so a more precise and less biassed sample survey is to be planned with the same expenses and time. Which of the former method and the revised method is better will be shown on following tables by analysing concrete figures. Error - The scope of survey-that is, the population-determined, the error of estimated values brought from the method of survey should be called in question. It's most desirable that the error of value should be within a certain range, selecting for the population the employees of establishments with 30 or more employees in each of industries (shown on the attached Industrial Classification Table)-in the undertakings of extremely small size, employees of establishments with a certain number of employees or more. Expenses and Time - If expenses and time permit, the areal or more-Expenses and Time - If expenses and time permit, the areal sampling method is the best . The cut-off survey is not desirable, because the employees of the establishments with a certain number of employees or less are omitted and therefore the survey does not reflect general wage conditions. As is shown on following table formed from the analysis of concrete figures, the wage level of employees of small size establishments is remarkably low. But this survey on wage conditions is not undertaken now because of erorr brought from the method of current survey. Even if the scope of the survey is limitted in cut-off survey, the samples selected from the list of establishments do not necessarily represent the population. Only the establishments registered at the time of Establishment Census are shown on the list, so later changes in the number on size of these establishments is not to be seen on the list. From this point of view, the areal sampling method, showing everchanging aspect of industrial world is to be adopted. In this paper the auther intended to design a reasonable sample survey, using modern sampling theory.続きを見る
|
Mendeley出力
