Sample Size Determination and Probability Level Estimation

樣品取樣數的決定及機率水凖的估計－圖表法

楊恭志(Kung-Chi Yang) ； Reino E. Pulkki

《台灣林業科學》 17卷2期 (2002/06) Pp. 135-141

https://doi.org/10.7075/TJFS.200206.0135

摘要

在生物界的計量中，樣品取樣數(sample size)的大小取決於機率水凖(probability level)、允許誤差率(allowable error)及樣品變方(variance of specimen)。樣品取樣數的決定可使用一個簡單的圖表法來估計，此圖表法包括了三個機率水凖(α=0.01, 0.05, 0.1)和三個允許誤差率（均值的5、10、20•），使用易捷，且僅需計算樣品變異係數(coefficient of variation)即可。本文以兩個實例來說明使用傳統的反覆計算方法(iteration calculation method)和圖表法(charting method)的差異，並以一個實例說明圖表法用於估計機率水凖的方法。使用圖表法有下列的優點：(1)簡單及容易使用，(2)不需要用t-table，(3)減少計算錯誤的機會，(4)減少計算時間，(5)增加樣品取樣數估算的精確度。

關鍵字

樣品取樣數 ； 允許誤差率 ； 機率水凖 ； 變異係數 ； 樣品變方

並列摘要

In biological measurements, the sample size required to achieve a given confidence level and a given allowable error depends upon the variance of the specimen used. A simple chart to determine the sample size required for 3 probability levels (α＝0.01, 0.05, and 0.1) and 3 levels of allowable error of the mean (E＝5, 10, and 20•) was developed. The chart is easy to use, and only the calculation of the coefficient of variation is required. Two examples are demonstrated with a traditional iteration calculation and the proposed charting method. One sample is used to illustrate the estimation of probability level by the charting method. This charting method has the following advantages over the traditional iteration calculation method: 1) it is simple and easy to use; 2) no t-table is needed; 3) there is less chance of error due to calculation; 4) less time is required for calculation; and 5) it produces a more-accurate sample size estimation.

並列關鍵字

sample size ； allowable error ； probability level ； coefficient of variation