ISO 19003:2006 ゴムおよびゴム製品—物理的試験への統計の適用に関するガイダンス | ページ 5

3.1

population

totality of data that could (theoretically) be obtained to characterize the property of the rubber, compounding ingredient or process being measured

3.2

sample

data actually available from the population as a result of an experimental test programme having been undertaken

3.3

variability

tendency for tests performed on nominally identical test pieces to produce different test results

3.4

arithmetic mean

sum of the (population or sample) data divided by the number of values used

Note 1 to entry: The “average” is the statistic most frequently used to describe a group of data. There are several kinds of average and they are often used in common parlance without specifying the type, which can be a source of confusion. Averages fall into two categories: computational and positional. The arithmetic mean is the most frequently used computational average. Others are considered in Annex B. Positional averages are the median and mode. The calculation of the arithmetic mean is given in Equations (1) and (2) in 6.2.2.2.

3.5

median

middle value (or average of the two middle values) when the data in a sample are arranged in numerically increasing value

3.6

mode

value of the property being measured which occurs with the greatest frequency

3.7

residual

difference (+ or −) between each value and the mean

Note 1 to entry: The sum of the residuals must be 0.

3.8

variance

arithmetic mean of the squared residuals

3.9

standard deviation

square root of the variance

Note 1 to entry: The calculation of standard deviation is given in Equations (5) and (6) in 6.2.3.2.1.

3.10

coefficient of variation

ratio of the standard deviation to the mean, generally expressed as a percentage

Note 1 to entry: The calculation of coefficient of variation is given in Equation (8) in 6.2.3.4.

3.11

range

maximum value minus the minimum value

3.12

standard error

standard deviation of the estimate of the population mean

Note 1 to entry: The calculation of standard error is given in Equation (7) in 6.2.3.2.3.

3.13

bias

difference between the average statistic and the true value of the parameter it is estimating, arising out of one or more systematic errors

3.14

accuracy

closeness of agreement between a test result and the accepted reference value

3.15

trueness

closeness of agreement between the average value of a large number of test results and the true or accepted reference value

Note 1 to entry: It is usually expressed in terms of bias.

3.16

precision

closeness of agreement between test results

3.17

repeatability

precision obtained under conditions where independent test results are obtained with the same method on identical test material in the same laboratory by the same operator using the same equipment within a short interval of time

3.18

reproducibility

precision obtained under conditions where independent test results are produced with the same method on identical test material in different laboratories with different operators using different equipment

3.19

level of significance

probability of error associated with a significance test

3.20

distribution function

function describing the probability that a random variable will take a value less than or equal to a number x

3.21

density distribution

slope of the distribution function at every value, i.e. the first derivative of the distribution function

3.22

normal distribution

symmetrical “bell-shaped” density distribution which is fully defined by its mean and standard deviation

Note 1 to entry: It is also known as the Laplace Gauss or Gaussian distribution.

3.23

double exponential distribution

asymmetrical distribution, fully defined by a single “shape” parameter, which has been used to characterize the distribution of tensile strengths in rubber compounds

3.24

Weibull distribution

symmetrical distribution fully defined by three parameters and found to be useful in characterizing lifetime tests such as fatigue

3.25

degrees of freedom

number of independent differences between the readings available for an estimate of standard deviation

3.26

confidence interval

range within which a value or parameter can be expected to lie with a given probability

3.27

confidence limits

extreme values of the confidence interval