Capability of detection—Part 3:Methodology for determination of the critical value for the response variable when no calibration data are used
Some standard content:
ICS03.120.30
National Standard of the People's Republic of China
GB/T33260.3—2018
Capability of detection-Part 3: Methodology for determination of the critical value for the response variable when no calibration data is available calibration data are used (ISO 11843-3:2003, MOD)
Released on June 7, 2018
State Administration for Market Regulation
Standardization Administration of the People's Republic of China
Implementation on January 1, 2019
Normative reference documents
Terms and definitions
Experimental design
Selection of the reference state whose net state variable value is 0 4.3
5 Critical value y of the response variable. Calculation of.
Basic method
Actual calculation
Reporting and use of critical values
Appendix A (Normative Appendix)
Appendix B (Informative Appendix)
Symbols used in this part
GB/T33260.3—2018
GB/T33260% Detection Capability" is currently divided into the following parts: Part 1: Terms and definitions;
Part 2: Method for determining the detection limit in the case of linear calibration; Part 3: Method for determining the critical value of the response variable in the case of no calibration data; Part 1: Method for comparing the minimum detectable value with the given value; Part 5: Method for determining the detection limit in the case of non-linear calibration. This part is Part 3 of GB/T33260. This part was drafted in accordance with the rules given in GB/T1.1—2009. GB/T33260.3—2018
This part adopts the redrafting method to modify and adopt ISO11843-3:2003 "Detection capability Part 3: Method for determining the critical value of the response variable in the absence of calibration data". Compared with ISO11843-3:2003, the main technical changes are as follows: Regarding normative reference documents, this part has made technically different adjustments to apply to my country's technical conditions. The adjustments are concentrated in Chapter 2 "Normative Reference Documents". The specific adjustments are as follows: ISO3534-1 is replaced by GB/T3358.1-2009, which is equivalent to the international document; ISO3534-2 is replaced by GB/T3358.2-2009, which is equivalent to the international document; GB/T3358.3-2009, which is equivalent to the international document, is replaced by GB/T3358.4-2009, which is equivalent to the international document. Replaced ISO3534-3; ·
replaced ISO11095:1996 with GB/T225542010 which was modified and adopted international documents; replaced ISO11843-1:1997 with GB/T33260.1-2016 which was modified and adopted international documents; replaced ISO11843-2:2000 with GB/T33260.2-2018 which was modified and adopted international documents; replaced ISOGuide30 with GB/T15000.2-1994 which was not equivalent to the international documents. This part was proposed and managed by the National Technical Committee for Standardization of Statistical Methods (SAC/TC21). Drafting organizations of this part: China National Institute of Standardization, Beijing University of Technology, Tianjin University, Xiamen Optimization Technology Co., Ltd., Hefei Normal University, CITIC Dicastal Co., Ltd. Drafters of this part: Zhang Fan, Xie Tianfa, Zhao Jing, Zhao Chao, Wu Gang, He Zhen, Zeng Jie, Chen Shiju, Ding Wenxing, Yu Zhenfan, Huang LiangI
GB/T33260.3--2018
The ideal requirement for the detection capability of a selected state variable is that the actual state of the observed system can be accurately distinguished as a basic state or a non-basic state. However, due to the influence of systematic and random variations, this ideal requirement cannot be met because: in fact, the values of the state variables of all reference states, including the basic state, are unknown. Therefore, all states can only be accurately described by the difference from the basic state, that is, the net state variable. Note: In GB/T15000.2-1994 and GB/T22554-2010, there is no distinction between state variables and net state variables. Therefore, both standards unequivocally assume that the state variables of the reference state are known. In addition, the calibration and sampling and sample preparation processes will introduce random errors to the measurement results. In this part, α represents the probability of (erroneously) detecting that the system is not in the basic state when the system is in the basic state. 1 Scope
Detection capability Part 3: Method for determining the critical value of the response variable in the case of no calibration data
GB/T33260.3-2018
This part of GB/T33260 specifies the method for estimating the critical value of the response variable using the mean and standard deviation of repeated observations under the reference state in the specific case where the net state variable value is zero (see 5.1), thereby determining whether the value of the response variable in the actual state (or the sample to be tested) exceeds the critical value provided by the reference state: GB/T33260.2-2018 specifies the general calculation method for determining the critical value of the response variable, the critical value of the net state variable and the minimum detectable value of the net state variable. These methods are applicable to the case where there is a linear calibration and the residual standard deviation is a constant or a linear function of the net state variable. This part specifies the method for determining the critical value of the response variable when no calibration data are available. It is assumed that the data follow a normal distribution or an approximate normal distribution.
The method given in this part is used in situations where a large number of basic conditions are available but a large number of actual conditions are difficult to obtain. 2 Normative references
The following documents are indispensable for the application of this document. For all dated references, only the dated version applies to this document. For any undated referenced documents, the latest version (including all amendments) applies to this document GB/T3358.1—2009
9 Statistical Vocabulary and Symbols Part 1: General Statistical Terms and Terms Used in Probability (ISO3534-1:2006, IDT)
GB/T3358.22009
GB/T3358.32009
Statistical Vocabulary and Symbols Part 2: Applied Statistics (ISO3534-2:2006, IDT) Statistical Vocabulary and Symbols Part 3 Division: Experimental design (ISO3534-3: 1999, IDT) GB/T4882—2001
Statistical processing and interpretation of data Normality test (ISO5479: 1997, IDT) GB/T6379.2—2004
Accuracy (trueness and precision) of measurement methods and results Part 2: Basic methods for determining the repeatability and reproducibility of standard measurement methods (ISO5725-2: 1994, IDT) GB/T15000.2—1994 Guidelines for standard sample work (2) Commonly used standard samples Terms and definitions (ISO Guide 30: 1992, NEQ)
GB/T 22554—2010
Linear calibration based on standard samples (ISO 11095: 1996, MOD) GB/T 33260.1 2016
GB/T 33260.2—2018
3 Terms and definitions
Detection capability Part 1: Terms and definitions (ISO 11843-1: 1997, MOD)Detection capability
Part 2: Detection limit in the case of linear calibration Determination method (ISO11843-2: GB/T3358.1-2009, GB/T3358.2-2009, GB/T3358.3-2009, GB/T4882-2001, GB/T15000.2-1994, GB/T6379.2-2004, GB/T22554-2010 and GB/T33260.1-2016 The terms and definitions defined apply to this document.
GB/T33260.3-2018
4 Experimental design
4.1 Overview
It is assumed that the measurement method is standardized and has been calibrated for similar types of measured quantities, but calibration has not been performed or is impossible under the conditions of the specific conditions of the study and the low level of the net state variable currently under consideration. All repeated measurements of the reference state (the state variable is 0) and the actual state (the sample to be measured) use exactly the same measurement method. The order of all measurements of the base state and the actual state should be randomized. If the response variable has negative values, it should not be discarded or modified (such as replaced by 0). 2 Selection of a reference state with a net state variable value of 0 4.2
One of the assumptions of the method given in this part is that the net state variable value of the reference state is 0. This assumption has been discussed in 4.1 of GB/T 33260.2-2018. In fact, the actual value of the state variable of the reference state is unknown, but its difference from the base state is known. wwW.bzxz.Net
The base state selected in this part only needs to have a state variable value that is sufficiently lower than the actual level of the measured quantity of the measurement method. When a prepared reference substance is used to represent the basic state, its composition should be as close as possible to that of the substance being measured. For example, in analytical chemistry, when a blank matrix is selected, its composition should be as similar as possible to that of the sample being measured in the measurement sequence, even if it is not completely the same. The influence caused by the possible presence of other substances or components or by the difference in the physical state of the sample can be very significant. In particular, when the research object is a solution, a solvent extract commonly encountered in the measurement should be used instead of a pure solvent. 4.3 Repetitions
Number of Repetitions J
In order to obtain a good estimate of the mean and standard deviation, it is necessary to use this method to perform a certain number of repeated measurements (number of repetitions J) under the basic state to obtain enough observations of the response variable, so that there is enough data to test whether the response variable obeys a normal distribution or an approximately normal distribution. This is very important.
The difference between the difference and the standard deviation does not exceed 30%
Usually, about 30 measurements per day can ensure that the estimate of the standard deviation is consistent with the true standard with a probability of . Note: In some cases, it is impossible to complete the above number of repeated measurements due to the limitation of the amount of material available or other reasons. In this case, the uncertainty of the estimate of the standard deviation obtained is large. When an estimate s of the true standard deviation is given (see 5% in 5.2), the flow can be based on: . The interval estimate of . The probability that the interval contains . is 1-α (α is given in advance). This is a statistical problem, (if the normality assumption is valid and , is taken as the sample standard deviation) is usually solved using the chi-square distribution related to the number of measurements: The interval estimate of is: xi-喜(v)
where -J-1, the quantile of the chi-square distribution can be obtained by looking up the table, α has been defined in the introduction. The symbols used here and elsewhere in this part are defined in Appendix A. Its
actual state (sample to be tested) The number of repeated measurements K will reduce the critical value of the response variable to some extent [see formula (4), but cost constraints should also be carefully considered.
4.3.2 Consistency of Repetitions
In order to measure the response variable, when sampling the base state, it is important to keep everything consistent throughout the sampling process. Standard reference materials should be used whenever possible, as their homogeneity can be guaranteed. It is worth noting that due to surface phenomena, electrostatic effects, precipitation, etc., the samples obtained may not be completely consistent. 4.3.3 Possible Interference Factors
The variability of possible interference factors should be minimized, which is listed in 4.1 of GB/T33260.2—2018. 2
5 Critical value y of the response variable. Calculation of
5.1 Basic method
GB/T33260.3-2018
The critical value y of the response variable defined in GB/T33260.1-2016. It means: if the value of the response variable y exceeds this value, the conclusion is drawn that the system is not in the basic state; the determination of the critical value of the response variable should satisfy that when the system is in the basic state, the above conclusion is drawn only with a small probability α. In other words, the critical value is the minimum significant value of the measurement or signal, which can be used as a discriminator for the background (noise). By comparing the arithmetic mean determined by the actual state with the critical value of its distribution. , the conclusion of "detection" or "non-detection" is drawn. The probability that the arithmetic mean of the measured value exceeds the critical value y of the distribution in the basic state (=0) should be less than or equal to a suitable predetermined probability α.
The critical value y of the response variable. It can usually be expressed as follows P(y.>ye/z=0)≤α
Note: P(y.>y.la
) refers to the probability of >y. under the condition of =0. .（1）
The inequality sign is used instead of the equality sign in formula (1) to consider the case where the random variable is discrete (such as Poisson cloth, whose α value does not always satisfy the equality case).
If:
a) y follows a normal distribution with a standard deviation of. b) The sample of the actual state is as close as possible to
e) The measurement is unbiased.
From formula (1), we can get the response variable
....(2)
In the formula:
Standard normal distribution
True value of the original hypothesis
Quantile
Number of repeated measurements of the basic state;
Arithmetic mean of the repeated measurements;
Number of repeated measurements of the actual state
Note: When the response variable increases with the increase of the level of the net state variable, the symbol "+" is used. When the response variable decreases with the increase of the level of the net state variable, the symbol "-" is used.
If α is estimated by s., the quantile of the distribution with a degree of freedom of > is used instead of -., that is, ye=y±ti--(v)son
Note: The use of the symbols "+" and "-" here is the same as in formula (2). (3)
When there is reasonable reason to believe that the value of the state variable under the actual basic state is close to 0, then. . =6b. The situation studied in this section is estimated by the standard deviation sb of repeated measurements of the response variable under the actual basic state. This is one of several methods for estimating. 5.2 Actual calculation
Use the method in GB/T4882-2001 or some other methods to test the normality of repeated measurements of the response variable under the basic state.
In this section, in a measurement sequence, the average of the three repeated measurements of the response variable under the basic state is: 3
GB/T33260.3—2018
y is the estimate of the expected ye of y; and the sample standard deviation Sh of y
is the estimate of ah.
So a good estimate of the critical value of the response variable is: y
ye=yhti-.()
（4）
Where, white degree=J-1. The statistical test is one-sided, and α is usually taken as 0.05 as recommended in GB/T33260.1-2016, and the corresponding quantile of the t distribution can be obtained through the quantile table. Note: The symbols "+\" and "+" are used in the same way as in formula (2). Formula (5) is directly applied to the single measurement of the sample to be tested: y. = yhti-()sh
Note: The symbols "+\\_\ are used in the same way as in formula (2). For examples of calculating the critical value of the response variable, see Appendix B. 5.3 Reporting and use of critical values
The number of measurements J of the response variable under the basic state should be given together with the standard deviation S of the measurement sequence. And the number of repetitions K of the response variable under the actual state should be reported. The selected α value (usually 0.05) and the critical value calculated when the number of repetitions of the response variable under the basic state and the actual state are given should be reported. The above content can be conveniently given in Table 1. Table 1 Critical values of response variables and corresponding experimental parameters Experimental parameters
Number of repetitions of response variables in the basic state Number of repetitions of response variables in the actual state Selected value (default value is 0.05)
The average value of the response variable in the basic state The average value of the response variable in the actual state Sample standard deviation of the response variable in the basic state When no calibration data is available, the critical value of the response variable derived by the simplified method of this part Symbol
If the average value of K repeated measurements in the actual state is not greater than the critical value, it indicates that there is no significant difference between the actual state and the basic state. However, the average value in the actual state should also be reported, and should not be reported as 0.62
j-1,2,.,J
Kurtosis test statistic
Appendix A
(Normative Appendix)
Symbols used in this part
GB/T33260.3—2018
Under the basic state, that is, the number of repeated measurements of the response variable when the net state variable is zero (blank matrix), the identification variable of the sample preparation under the basic state, that is, the net state variable is zero (blank matrix), the actual state (sample), the number of repeated measurements of the response variable, the probability
the sample standard deviation of the response variable
the basic state, that is, the sample standard deviation of the response variable when the net state variable is zero (blank matrix), the actual
the sample standard deviation of the response variable under the basic state, the t distribution or statistics
Shapiro
wiks test
net state variable value
response variable signal
the actual basic state of the response variable (arithmetic mean), the actual state (to be tested), Sample)
Critical value of response variable
Magic child under Shengtejiajin||tt| ... (Recorded)
The mass fraction of cadmium in the BCR soil samples was determined by atomic luminescence spectrometry after dissolution in aqua regia. The cadmium content was analyzed using a 0.5 certified standard sample No. 142 light sandy soil sample, and the cadmium content level was known from other data to be below the critical limit of the measurement method reported here (approximately one-tenth of the critical limit). The samples were dissolved and filtered in aqua regia in a batch process to prepare 25 mL samples for spectral measurement. Cadmium was measured and calibrated using a 24-band inductively coupled plasma emission spectrometer (spectral line wavelength 226nm), with a total of J=30 readings, see Table B.1. Measurement results of cadmium elements in certified standard sample No. 142 soil Table B.1
Response variable
Response variable|| tt||response variable
Some normality tests (skewness test, kurtosis test and Shapiro-Wilks test) were used for the response variable, and the results showed that there was no significant deviation from normality. An outlier test (Grubbs test) was further performed. α was taken as 0.05. From the quantile table, ti- was obtained. (v) = to.95 (29) -1.699. The mean value of the response variable y = 2.1898mV, and the standard deviation 5b = 0.0186mV. Three measurements were made simultaneously in a similar soil sample, and the response variable values were 2.177mV, 2.183mV and 2.161mV. Using formula (4), the critical value of the response variable for the three measurements of the actual sample (retaining 3 decimal places) was calculated as: 11
y. =2.1898+1.699×0.0186×
=2.1898+0.0191
-2.209(mV)
The results are reported in Table B.2.
GB/T33260.3—2018
Table B.2 Critical values of response variables for elements measured by atomic emission method in certified standard sample No. 142 soil
Number of replicates of response variables in basic state; α value selected in actual state;
Average value of response variables in basic state; Average value of response variables in actual state; Sample standard deviation of response variables in basic state; When no calibration data is available, the critical value of response variables derived by the simplified method of this section does not exceed the critical value of response variables, indicating that there is no difference between the basic state and the sample to be tested, and the value
Note 1: The above critical value is higher than the critical value (0.815) when only reagents are used in the whole process mV), which is much higher than the "pure solution ion" claimed by the instrument manufacturer (about 0.027mV), which shows that the sample matrix has a great influence on the critical value. Note 2: The above data are from Soil Science Department of the IACR, Rothamsted, Harpenden, Hertfordshire, UK Example 2
Determine the chemical oxygen demand in water by titration. It should be noted that the calibration curve of the whole process of determining the chemical oxygen demand in water is monotonically decreasing: as the oxygen demand increases, the available oxygen decreases, so that the volume of ammonium sulfate iron (III) solution in the back titration decreases. For the 0.06mol/L ammonium sulfate iron (III) solution used in the titration, the measurement Thirty blanks were used to determine the oxygen demand (COD) of water. See Table B.3. Table B.3
Volume of solution used in titration
Determination of oxygen demand in water by titration
Volume of solution used in titration
Volume of solution used in titration
Volume of solution used in titration
-Some normality tests (skewness test, kurtosis test and Shapiro-Wilks test) and outlier tests (Grubbs test) show slight deviations from normality: the kurtosis test at α=0.01 failed (b2=1.737, while the critical values are 1.79 and 5.12 respectively), and =0.057
GB/T33260.3-—2018|| The Shapiro-Wilks test for tt|| failed (W = 0.9045, while the critical values are 0.900 for α = 0.01 and 0.927 for α = 0.05). As indicated by the two normality tests, the distribution of the original data can be considered to be approximately normal. However, even a simple frequency distribution plot indicates that the results may be a mixture of two distributions. Therefore, it is wise to ask the laboratory providing the data to determine whether there are any anomalies in the data records of the response variable. If the data records are confirmed to be accurate, the steps to determine the critical value of the response variable based on a single measurement of an actual (to be tested) sample are as follows: The mean value of the response variable is calculated as: = 19.829 mL, and the standard deviation 5b = 0.0774 mL. From the quantile table, it can be found that ti-. (v) to.95 (29) = 1.699. When using equation (5), the regressiveness of the calibration requires that the average response variable in the baseline state be subtracted from the variance term (rather than added to it), so the critical value of the response variable for a single measurement in the actual sample (to two decimal places) is: y. =19.829-1.699×0.0774X
-19.829-0.1337
=19.70(mL）
The results are shown in Table B.4g
4 Measurement data of critical value of response variable for chemical oxygen demand in water by titration
Number of repetitions of response variable in basic state; α value selected for number of repetitions of response variable in actual state
Average value of response variable in basic state; Sample standard deviation of response variable in basic state; Critical value data of response variable derived by simplified method of this part when no calibration data is available
Because the minimum titration of 0.06mol/L ammonium sulfate iron (II) in actual (to be tested) sample is not less than 19.70mL, there is no difference between basic state and the samples to be tested completed at the same time
Note: The above data are quoted from the document of ISO/TC147 (water quality). 85 The Certified Standard Sample 142 light sandy soil sample was analyzed for cadmium content, and other data showed that the cadmium content level was below the critical limit of the measurement method reported here (approximately one tenth of the critical limit). The sample was dissolved and filtered in aqua regia in batch mode to prepare 25 mL samples for spectral measurement. Cadmium was measured using a 24-band inductively coupled plasma emission spectrometer (spectral line wavelength 226nm) and calibrated, with a total of J=30 readings, see Table B.1. Results of the measurement of cadmium elements in the certified standard sample 142 soil Table B.1
Response variable
Response variable
Response variable
Some normality tests (skewness test, kurtosis test and Shapiro-Wilks test) showed no significant deviation from normality. A further outlier test (Grubbs test) was performed. α was taken as 0.05. ti- was obtained from the quantile table. (v) = to.95 (29) - 1.699. The mean value of the response variable y = 2.1898 mV, standard deviation 5b = 0.0186 mV. Three measurements were made simultaneously in a similar soil sample, and the response variable values were 2.177 mV, 2.183 mV and 2.161 mV. Using formula (4), the critical value of the response variable for the three measurements of the actual sample (retain 3 decimal places) is calculated as: 11
y. = 2.1898 + 1.699 × 0.0186 ×
= 2.1898 + 0.0191
- 2.209 (mV)
The results are reported in Table B.2.
GB/T33260.3—2018
Table B.2 Critical values of response variables for elements measured by atomic emission method in certified standard sample No. 142 soil Measurement data
Number of repetitions of response variables in basic state Number of repetitions of response variables in actual state α value for selection
Average value of response variables in basic state Average value of response variables in actual state Sample standard deviation of response variables in basic state When no calibration data is available, the critical value of response variables derived by the simplified method of this part does not exceed the critical value of response variables, indicating that there is no difference between basic state and sample to be tested, value
Note 1: The above critical value is higher than the critical value (0.815) when only reagents are used in the whole process mV), which is much higher than the "pure solution ion" claimed by the instrument manufacturer (about 0.027mV), which shows that the sample matrix has a great influence on the critical value. Note 2: The above data are from Soil Science Department of the IACR, Rothamsted, Harpenden, Hertfordshire, UK Example 2
Determine the chemical oxygen demand in water by titration. It should be noted that the calibration curve of the whole process of determining the chemical oxygen demand in water is monotonically decreasing: as the oxygen demand increases, the available oxygen decreases, so that the volume of ammonium sulfate iron (III) solution in the back titration decreases. For the 0.06mol/L ammonium sulfate iron (III) solution used in the titration, the measurement Thirty blanks were used to determine the oxygen demand (COD) of water. See Table B.3. Table B.3
Volume of solution used in titration
Determination of oxygen demand in water by titration
Volume of solution used in titration
Volume of solution used in titration
Volume of solution used in titration
-Some normality tests (skewness test, kurtosis test and Shapiro-Wilks test) and outlier tests (Grubbs test) show slight deviations from normality: the kurtosis test at α=0.01 failed (b2=1.737, while the critical values are 1.79 and 5.12 respectively), and =0.057
GB/T33260.3-—2018|| The Shapiro-Wilks test for tt|| failed (W = 0.9045, while the critical values are 0.900 for α = 0.01 and 0.927 for α = 0.05). As indicated by the two normality tests, the distribution of the original data can be considered to be approximately normal. However, even a simple frequency distribution plot indicates that the results may be a mixture of two distributions. Therefore, it is wise to ask the laboratory providing the data to determine whether there are any anomalies in the data records of the response variable. If the data records are confirmed to be accurate, the steps to determine the critical value of the response variable based on a single measurement of an actual (to be tested) sample are as follows: The mean value of the response variable is calculated as: = 19.829 mL, and the standard deviation 5b = 0.0774 mL. From the quantile table, it can be found that ti-. (v) to.95 (29) = 1.699. When using equation (5), the regressiveness of the calibration requires that the average response variable in the baseline state be subtracted from the variance term (rather than added to it), so the critical value of the response variable for a single measurement in the actual sample (to two decimal places) is: y. =19.829-1.699×0.0774X
-19.829-0.1337
=19.70(mL）
The results are shown in Table B.4g
4 Measurement data of critical value of response variable for chemical oxygen demand in water by titration
Number of repetitions of response variable in basic state; α value selected for number of repetitions of response variable in actual state
Average value of response variable in basic state; Sample standard deviation of response variable in basic state; Critical value data of response variable derived by simplified method of this part when no calibration data is available
Because the minimum titration of 0.06mol/L ammonium sulfate iron (II) in actual (to be tested) sample is not less than 19.70mL, there is no difference between basic state and the samples to be tested completed at the same time
Note: The above data are quoted from the document of ISO/TC147 (water quality). 85 The Certified Standard Sample 142 light sandy soil sample was analyzed for cadmium content, and other data showed that the cadmium content level was below the critical limit of the measurement method reported here (approximately one tenth of the critical limit). The sample was dissolved and filtered in aqua regia in batch mode to prepare 25 mL samples for spectral measurement. Cadmium was measured using a 24-band inductively coupled plasma emission spectrometer (spectral line wavelength 226nm) and calibrated, with a total of J=30 readings, see Table B.1. Results of the measurement of cadmium elements in the certified standard sample 142 soil Table B.1
Response variable
Response variable
Response variable
Some normality tests (skewness test, kurtosis test and Shapiro-Wilks test) showed no significant deviation from normality. A further outlier test (Grubbs test) was performed. α was taken as 0.05. ti- was obtained from the quantile table. (v) = to.95 (29) - 1.699. The mean value of the response variable y = 2.1898 mV, standard deviation 5b = 0.0186 mV. Three measurements were made simultaneously in a similar soil sample, and the response variable values were 2.177 mV, 2.183 mV and 2.161 mV. Using formula (4), the critical value of the response variable for the three measurements of the actual sample (retain 3 decimal places) is calculated as: 11
y. = 2.1898 + 1.699 × 0.0186 ×
= 2.1898 + 0.0191
- 2.209 (mV)
The results are reported in Table B.2.
GB/T33260.3—2018
Table B.2 Critical values of response variables for elements measured by atomic emission method in certified standard sample No. 142 soil Measurement data
Number of repetitions of response variables in basic state Number of repetitions of response variables in actual state α value for selection
Average value of response variables in basic state Average value of response variables in actual state Sample standard deviation of response variables in basic state When no calibration data is available, the critical value of response variables derived by the simplified method of this part does not exceed the critical value of response variables, indicating that there is no difference between basic state and sample to be tested, value
Note 1: The above critical value is higher than the critical value (0.815) when only reagents are used in the whole process mV), which is much higher than the "pure solution ion" claimed by the instrument manufacturer (about 0.027mV), which shows that the sample matrix has a great influence on the critical value. Note 2: The above data are from Soil Science Department of the IACR, Rothamsted, Harpenden, Hertfordshire, UK Example 2
Determine the chemical oxygen demand in water by titration. It should be noted that the calibration curve of the whole process of determining the chemical oxygen demand in water is monotonically decreasing: as the oxygen demand increases, the available oxygen decreases, so that the volume of ammonium sulfate iron (III) solution in the back titration decreases. For the 0.06mol/L ammonium sulfate iron (III) solution used in the titration, the measurement Thirty blanks were used to determine the oxygen demand (COD) of water. See Table B.3. Table B.3
Volume of solution used in titration
Determination of oxygen demand in water by titration
Volume of solution used in titration
Volume of solution used in titration
Volume of solution used in titration
-Some normality tests (skewness test, kurtosis test and Shapiro-Wilks test) and outlier tests (Grubbs test) show slight deviations from normality: the kurtosis test at α=0.01 failed (b2=1.737, while the critical values are 1.79 and 5.12 respectively), and =0.057
GB/T33260.3-—2018|| The Shapiro-Wilks test for tt|| failed (W = 0.9045, while the critical values are 0.900 for α = 0.01 and 0.927 for α = 0.05). As indicated by the two normality tests, the distribution of the original data can be considered to be approximately normal. However, even a simple frequency distribution plot indicates that the results may be a mixture of two distributions. Therefore, it is wise to ask the laboratory providing the data to determine whether there are any anomalies in the data records of the response variable. If the data records are confirmed to be accurate, the steps to determine the critical value of the response variable based on a single measurement of an actual (to be tested) sample are as follows: The mean value of the response variable is calculated as: = 19.829 mL, and the standard deviation 5b = 0.0774 mL. From the quantile table, it can be found that ti-. (v) to.95 (29) = 1.699. When using equation (5), the regressiveness of the calibration requires that the average response variable in the baseline state be subtracted from the variance term (rather than added to it), so the critical value of the response variable for a single measurement in the actual sample (to two decimal places) is: y. =19.829-1.699×0.0774X
-19.829-0.1337
=19.70(mL）
The results are shown in Table B.4g
4 Measurement data of critical value of response variable for chemical oxygen demand in water by titration
Number of repetitions of response variable in basic state; α value selected for number of repetitions of response variable in actual state
Average value of response variable in basic state; Sample standard deviation of response variable in basic state; Critical value data of response variable derived by simplified method of this part when no calibration data is available
Because the minimum titration of 0.06mol/L ammonium sulfate iron (II) in actual (to be tested) sample is not less than 19.70mL, there is no difference between basic state and the samples to be tested completed at the same time
Note: The above data are quoted from the document of ISO/TC147 (water quality). 8For the 0.06 mol/L ammonium sulfate iron (III) solution, 30 blank samples were measured to determine the oxygen demand (COD) of the water. See Table B.3. Table B.3
Volumes of solution used in titration
Determination of oxygen demand in water by titration
Volumes of solution used in titration
Volumes of solution used in titration
Volumes of solution used in titration
Some normality tests (skewness, kurtosis and Shapiro-Wilks tests) and outlier tests (Grubbs test) indicate slight departures from normality: the kurtosis test for α = 0.01 fails (b2 = 1.737, while the critical values are 1.79 and 5.12, respectively), and the Shapiro-Wilks test for W = 0.057
GB/T 33260.3-—2018
fails (W = 0.9045, while the critical values are 0.900 for α = 0.01 and 0.927 for α = 0.05). As indicated by the two normality tests, the distribution of the raw data can be considered to be approximately normal. However, even a simple frequency distribution plot indicates that the results may be a mixture of two distributions. Therefore, it is wise to ask the laboratory that provided the data to determine whether there are any anomalies in the response variable data record. If the data record is confirmed to be accurate, the steps to determine the critical value of the response variable based on a single measurement of an actual (test) sample are as follows: The mean value of the response variable is calculated as: =19.829mL, and the standard deviation 5b = 0.0774mL. From the quantile table, it can be found that ti-. (v) to.95 (29) = 1.699. When using equation (5), the regressivity of the calibration requires that the average response variable in the baseline state is subtracted from the variation term (rather than added to it), so the critical value of the response variable for a single measurement of the actual sample (retained to two decimal places) is: y. =19.829-1.699×0.0774X
-19.829-0.1337
=19.70(mL）
The results are shown in Table B.4g
4 Measurement data of critical value of response variable for chemical oxygen demand in water by titration
Number of repetitions of response variable in basic state; α value selected for number of repetitions of response variable in actual state
Average value of response variable in basic state; Sample standard deviation of response variable in basic state; Critical value data of response variable derived by simplified method of this part when no calibration data is available
Because the minimum titration of 0.06mol/L ammonium sulfate iron (II) in actual (to be tested) sample is not less than 19.70mL, there is no difference between basic state and the samples to be tested completed at the same time
Note: The above data are quoted from the document of ISO/TC147 (water quality). 8For the 0.06 mol/L ammonium sulfate iron (III) solution, 30 blank samples were measured to determine the oxygen demand (COD) of the water. See Table B.3. Table B.3
Volumes of solution used in titration
Determination of oxygen demand in water by titration
Volumes of solution used in titration
Volumes of solution used in titration
Volumes of solution used in titration
Some normality tests (skewness, kurtosis and Shapiro-Wilks tests) and outlier tests (Grubbs test) indicate slight departures from normality: the kurtosis test for α = 0.01 fails (b2 = 1.737, while the critical values are 1.79 and 5.12, respectively), and the Shapiro-Wilks test for W = 0.057
GB/T 33260.3-—2018
fails (W = 0.9045, while the critical values are 0.900 for α = 0.01 and 0.927 for α = 0.05). As indicated by the two normality tests, the distribution of the raw data can be considered to be approximately normal. However, even a simple frequency distribution plot indicates that the results may be a mixture of two distributions. Therefore, it is wise to ask the laboratory that provided the data to determine whether there are any anomalies in the response variable data record. If the data record is confirmed to be accurate, the steps to determine the critical value of the response variable based on a single measurement of an actual (test) sample are as follows: The mean value of the response variable is calculated as: =19.829mL, and the standard deviation 5b = 0.0774mL. From the quantile table, it can be found that ti-. (v) to.95 (29) = 1.699. When using equation (5), the regressivity of the calibration requires that the average response variable in the baseline state is subtracted from the variation term (rather than added to it), so the critical value of the response variable for a single measurement of the actual sample (retained to two decimal places) is: y. =19.829-1.699×0.0774X
-19.829-0.1337
=19.70(mL）
The results are shown in Table B.4g
4 Measurement data of critical value of response variable for chemical oxygen demand in water by titration
Number of repetitions of response variable in basic state; α value selected for number of repetitions of response variable in actual state
Average value of response variable in basic state; Sample standard deviation of response variable in basic state; Critical value data of response variable derived by simplified method of this part when no calibration data is available
Because the minimum titration of 0.06mol/L ammonium sulfate iron (II) in actual (to be tested) sample is not less than 19.70mL, there is no difference between basic state and the samples to be tested completed at the same time
Note: The above data are quoted from the document of ISO/TC147 (water quality). 8
Tip: This standard content only shows part of the intercepted content of the complete standard. If you need the complete standard, please go to the top to download the complete standard document for free.