Preprint
Article

This version is not peer-reviewed.

A Robust Method for Calculating Precision for Interlaboratory Studies with a Staggered-Nested Design

Submitted:

29 April 2025

Posted:

19 May 2025

You are already at the latest version

Abstract
Outlier testing and elimination can be avoided via application of robust estimators. Amongst robust estimators, the Q/Hampel method displays the best performance (in terms of breakdown point and efficiency). While the formulas and correction factors for Q/Hampel in the case of the design with two variance components (e.g. within- and between-laboratory variance) have already been made available, corresponding formulas for other designs have not. A case in point is the staggered-nested design, which is a highly efficient design for e.g. the estimation of intermediate precision in method validation studies. Accordingly, the formulas and correction factors for the use of Q/Hampel in the staggered-nested design are provided here.
Keywords: 
;  ;  ;  ;  

1. Introduction

In the case of quantitative methods, the aim of interlaboratory validation studies is to characterize method performance in terms of trueness and precision. ISO 5725-2/-3/-4 ([1], [2], [3]) provide a number of different designs allowing the evaluation of these two performance characteristics.
A particularly powerful design is the staggered-nested design described in ISO 5725-3 [2]. The simplest such design is the two-factor staggered-nested design, where each laboratory obtains three test results. For laboratory i , test results y i 1 and y i 2 are obtained under repeatability conditions, and y i 3 under intermediate conditions, e.g. on a different day.
The standard calculation method for precision is analysis of variance (ANOVA), preceded by outlier testing. While ANOVA, under the normal distribution assumption, is very efficient in the case of balanced designs, in the case of unbalanced designs, such as the staggered-nested design, the usual ANOVA does not use all the information available in the data. This is related to the fact that the reproducibility standard deviation is not calculated directly, but rather from the laboratory and intermediate standard deviation values; while the latter, in turn, is calculated from the standard deviation under intermediate conditions and the repeatability standard deviation. Another disadvantage of ANOVA and other conventional methods is that they very much depend on the data following a normal distribution and are thus highly sensitive to outliers. This can only be partially offset via outlier tests, since even if conspicuous values are not yet statistically significant outliers, considerable deviations can result in the determined precision data.
The advantage of robust estimators is that no outlier testing is required. For the estimation of means and standard deviation values, different robust estimators exist. The performance of a given robust estimators can be characterized via breakdown point (proportion of data that can be outliers without the estimate being affected) and efficiency (ratio of the statistical uncertainty of the estimate to that of the classical estimator under the normal distribution assumption). An overview of different breakdown point and efficiency values is provided in ISO 13528 [4].
The Q/Hampel and Qn methods have both the highest breakdown point and the best efficiency.
The Q/Hampel method uses the Q method for the calculation of the robust reproducibility standard deviation s R and repeatability standard deviation s r together with the Hampel estimator for the calculation of the location parameter x * as described in ISO 13528 [4]. The theoretical basis for the Q method, including asymptotic performance and finite sample breakdown, is described in Müller et al. [5] and Uhlig [6].
The Q method is not only robust against outlying results, but also against a situation where many test results are identical, e.g. due to quantitative data on a discontinuous scale or due to rounding distortions. In such a situation other Q-like methods (e.g. the Qn method originally introduced by Rousseeuw et al. [8] for univariate data) can fail because many pairwise differences are zero.
The Q method was introduced for the one-way random effect model and later modified – for ISO 13528 [4] – to deal with the situation that that there are many equal data, e.g., for the case where the number of significant digits is too small.
The Q/Hampel method is typically used for conventional designs, but it can also be applied for the staggered-nested design with two factors according to ISO 5725-3 [2] – in particular for estimating the intermediate standard deviation in addition to the reproducibility and repeatability standard deviation.

2. Robust Statistical Analysis of Results by Means of the Q/Hampel Method in a Staggered-Nested Design with Two Factors

For each level, the data obtained in the experiment are denoted y i k l (with i representing factor 0, i.e. laboratory, i = 1 , , p ; k representing factor 1, k = 1,2 ; and l representing the replicate, with l ranging from 1 to n k , with n 1 = 2 and n 2 = 1 ), i.e. for laboratory i there are three measurement results y i 11 , y i 12 , y i 21 . In summary, test results, grouped by laboratory, are denoted as follows:
y 111 , y 112 , y 121 L a b   1 ,   y 211 , y 212 , y 221 L a b   2 , , y p 11 , y p 12 , y p 21 L a b   p

2.1. Determination of the Robust Reproducibility Standard Deviation s R

Using the Q Method
The calculation relies on the use of pairwise differences within the data set and, thus, does not depend on the estimate of the mean or median.
The algorithm can be described as follows.
Based on the measurement results as structured in equation (1), the cumulative distribution function of all absolute between-laboratory differences is calculated as follows:
H 1 x = 2 9 p ( p 1 ) 1 i < j p I y i 11 y j 11 x + I y i 11 y j 12 x + I y i 11 y j 21 x + I y i 12 y j 11 x + I y i 12 y j 12 x + I y i 12 y j 21 x + I y i 21 y j 11 x + I y i 21 y j 12 x + I y i 21 y j 21 x
where I y i k l y j k l x = 1 if y i k l y j k l x 0 otherwise denotes the indicator function.
Discontinuity points of H 1 x are denoted
x 1 , , x m ,   where   x 1 < x 2 < < x m
For each positive discontinuity points x 1 , , x m , define
G 1 x i = 0,5 H 1 x i + H 1 x i 1 if i 2 0,5   H 1 x 1 if i = 1 ; x 1 > 0  
and let
G 1 0 = 0
For each x within the interval 0 , x m   , G 1 x is obtained by linear interpolation between discontinuity points 0 < x 1 < x 2 < < x m .
Finally, the robust reproducibility standard deviation s R is obtained as
s R = G 1 1 0,25 + 0,75 H 1 0 2 Φ 1 0,625 + 0,375 H 1 0 b p
where H 1 0 is calculated as in equation (2) and is set equal to zero unless there are identical values in the data set.
In Equation (4), Φ 1 q denotes the q t h quantile of the standard normal distribution and b p denotes the correction factor corresponding to the number of laboratories p .
The correction factors b p were obtained via a simulation study, which will now be briefly described. In each simulation step, normally ( N 0,1 ) distributed data corresponding to p laboratories were generated and the robust reproducibility standard deviation s R was calculated in accordance with the Q method from the formulas given above. Taking the mean value across 10 6 simulation steps – separately for each p – it was possible to calculate the expected value for the reproducibility standard deviation s R . For a given p , the correction factor b p was then obtained by taking the reciprocate of the expected value. For each value p between 4 and 100, Table 1 provides the expected value for the reproducibility standard deviation s R along with the corresponding relative standard error and correction factor b p .
For p > 12 , it was possible to derive a functional relationship between p and the correction factor b p via nonlinear optimization. This functional relationship is given in equation (5) and presented in Figure 1 for 12 < p 100 .
b p = 0,2680 1 p 2,3363 + 0,5810 1 p + 0,9998 1

2.2. Determination of the Robust Intermediate Standard Deviation s I ( 1 ) Using the Q Method

Take the case that factor 1 is day. As explained above, in the two-factor staggered-nested design, there are two results for day 1, and one for day 2. For a given laboratory, there are thus two within-laboratory differences corresponding to factor 1. Accordingly, the cumulative distribution function corresponding to factor 1 is calculated as follows:
H 2 , I ( 1 ) x = 1 2 p i = 1 p I y i 11 y i 21 x + I y i 12 y i 21 x
where I y i 1 l y i 21 x = 1 if y i 1 l y i 21 x 0 otherwise ( l = 1,2 ) denotes the indicator function.
As above, the discontinuity points of H 2 , I ( 1 ) x are denoted
x 1 , , x m ,   where   x 1 < x 2 < < x m
For each positive discontinuity points x 1 , , x m , define
G 2 , I ( 1 ) x i = 0,5 H 2 , I ( 1 ) x i + H 2 , I ( 1 x i 1 if i 2 0,5   H 2 , I ( 1 x 1 if i = 1 ; x 1 > 0  
and let
G 2 , I ( 1 ) 0 = 0
For each x within the interval 0 , x m   , G 2 , I ( 1 ) x is obtained via linear interpolation between discontinuity points 0 < x 1 < x 2 < < x m .
Finally, the robust intermediate standard deviation s I ( 1 ) is obtained as
s I ( 1 ) = G 2 , I ( 1 ) 1 0,5 + 0,5 H 2 , I ( 1 ) 0 2 Φ 1 0,75 + 0,25 H 2 , I ( 1 ) 0 c p
where H 2 , I ( 1 ) 0 is calculated as in equation (6) and is set equal to zero unless there are identical values in the data set.
As above, Φ 1 q denotes the q t h quantile of the standard normal distribution.
Similarly to b p in the previous section, for a given number of laboratories p , the correction factor c p was calculated via a simulation study.
For each value p between 4 and 100, Table 2 provides the expected value for the intermediate standard deviation s I ( 1 ) along with the corresponding relative standard error and correction factor c p .
For p > 12 , it was possible to derive a functional relationship between the number p of laboratories and the correction factor c p via nonlinear optimization. The functional relationship for 12 < p 100 is given in equation (9) and shown in Figure 2.
c p = 2,1251 1 p 11,3592 + 0,3051 1 p + 0,9999 1 p   odd 2,9723 1 p 4,6860 + 0,3199 1 p + 0,9998 1 p   even
If s I ( 1 ) results in a value greater than s R , then s I ( 1 ) is set to s R .

2.3. Determination of the Robust Repeatability Standard Deviation s r

Using the Q Method
As far as repeatability is concerned, there is only one difference which can be calculated per laboratory in the two-factor staggered-nested design: namely, that corresponding to the this first level of factor 1 (e.g. the two results obtained on day 1, if factor 1 is day). Accordingly, the cumulative distribution function corresponding to repeatability precision is calculated as follows::
H 2 , r x = 1 p i = 1 p I y i 11 y i 12 x
where I y i 11 y i 12 x = 1 if y i 11 y i 12 x 0 otherwise denotes the indicator function.
As above, the discontinuity points, the discontinuity points of H 2 , r x are denoted
x 1 , , x m ,   where   x 1 < x 2 < < x m
For each positive discontinuity points x 1 , , x m , define
G 2 , r x i = 0,5 H 2 , r x i + H 2 , r x i 1 if i 2 0,5   H 2 , r x 1 if i = 1 ; x 1 > 0  
and let
G 2 , r 0 = 0
For each x within the interval 0 , x m   , the function G 2 , r x is calculated via linear interpolation between discontinuity points 0 < x 1 < x 2 < < x m .
Finally, the robust repeatability standard deviation s r is obtained as
s r = G 2 , r 1 0,5 + 0,5 H 2 , r 0 2 Φ 1 0,75 + 0,25 H 2 , r 0 c p
where H 2 , r 0 is calculated as in equation (10) and is equal to zero unless there are identical values in the data set.
As above, Φ 1 q denotes the q t h quantile of the standard normal distribution.
The correction factor c p is the same as in the previous section, see Table 2 and equation (9).
If s r is greater than s I ( 1 ) , then s r is set equal to s I ( 1 ) .

2.4. Determination of the Robust Mean x *

Using the Hampel Estimator
Calculate the weighted means for each laboratory, denoted y 1 ,   , y p , i.e.
y i = 1 4 y i 11 + y i 12 + 2 y i 21
Calculate the robust mean, x * , by solving the equation
i = 1 p Ψ y i x *   s * = 0
where
Ψ q = 0 ,     q 4,5 4,5 q ,   4,5 < q 3 1,5 ,   3 < q 1,5 q ,   1,5 < q 1,5 1,5 ,     1,5 < q 3 4,5 q ,     3 < q 4,5 0 ,     q > 4,5
and
s * = s R 2 1 2 s I 1 2 1 8 s r 2
where s R , s I ( 1 ) and s r denote the robust reproducibility, intermediate and repeatability standard deviations obtained in accordance with the Q method (as described in 2.1, 2.2 and 2.3), respectively.
The exact solution may be obtained in a finite number of steps (not iteratively) using the property that Ψ is partially linear in x * and by means of the interpolation nodes of the left side of equation (14) (interpreted here as a function of x * ).
The interpolation nodes are obtained as follows
- for the first value y 1 :
d 1 = y 1 4,5 s * , d 2 = y 1 3 s * , d 3 = y 1 1,5 s * , d 4 = y 1 + 1,5 s * , d 5 = y 1 + 3 s * ,
d 6 = y 1 + 4,5 s * - for the first value y 2 :
d 1 = y 2 4,5 s * , d 2 = y 2 3 s * , d 3 = y 2 1,5 s * , d 4 = y 2 + 1,5 s * , d 5 = y 2 + 3 s * ,
d 6 = y 2 + 4,5 s * - and so on for all values y 3 ,   , y p .
The notes d 1 , d 2 , d 3 , , d 6 p are sorted in ascending order: d 1 , d 2 , d 3 , , d 6 p .
For each m = 1 , , ( 6 p 1 ) , the following quantity is then calculated:
p m = i = 1 p Ψ y i d m   s *
It is then checked whether
p m = 0 . If so, d m is a solution of equation (14).
p m + 1 = 0 . If so, d m + 1 is a solution of equation (14).
p m p m + 1 < 0 . If so, x m = d m p m p m + 1 p m d m + 1 d m is a solution of equation (14).
Let S denote the set of all solutions of equation (14).
The solution x * S nearest to the median is taken as the location parameter x * , i.e
x * median ( y 1 , y 2 ,   , y p ) = min x median ( y 1 , y 2 ,   , y p ) ; x S
Several solutions may exist. If there are two solutions nearest the median, or if there is no solution at all, the median itself is taken as the location parameter x * .
This implementation of Hampel’s estimator has approximately 96 % efficiency for normally distributed data.
If this estimation method is used, laboratory results differing from the mean by more than 4,5 times the reproducibility standard deviation no longer have any effect on the calculation result, i.e. they are treated as outliers.

3. Conclusion

The Q/Hampel procedure described in this paper extends the range of robust statistical methods to staggered-nested designs, providing for the first time an adequate approach for handling outliers in such complex experimental formats. In conventional ISO 5725 approaches, especially under unbalanced conditions, the absence of a straightforward outlier identification process at the intermediate stage often necessitates the exclusion of all results from a laboratory, resulting in significant information loss. By contrast, the robust approach presented here allows the retention of valuable data from all laboratories, maximizing the information available for the estimation of precision parameters and means. Thus, the introduction of robust estimators in staggered-nested designs ensures both the integrity of statistical analysis and the efficient use of all available data in method validation studies.

Conflicts of Interest

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

  1. ISO 5725-2, Accuracy (trueness and precision) of measurement methods and results — Part 2: Basic method for the determination of repeatability and reproducibility of a standard measurement method.
  2. ISO 5725-3, Accuracy (trueness and precision) of measurement methods and results — Part 3: Intermediate measures of the precision of a standard measurement method. -.
  3. ISO 5725-5, Accuracy (trueness and precision) of measurement methods and results — Part 5: Alternative methods for the determination of the precision of a standard measurement method.
  4. ISO 13528, Statistical methods for use in proficiency testing by interlaboratory comparisons.
  5. Müller, C.H. Müller C.H. and Uhlig S., Estimation of variance components with high breakdown point and high efficiency; Biometrika; 88: Vol. 2, pp. 353-366, 2001.
  6. Uhlig S., Robust estimation of variance components with high breakdown point in the 1-way random effect model. In: Kitsos, C.P. and Edler, L.; Industrial Statistics; Physica, S. 65-73, 1997.
  7. Uhlig S., Robust estimation of between and within laboratory standard deviation measurement results below the detection limit, Journal of Consumer Protection and Food Safety, 2015.
  8. Rousseeuw, P.J. and Croux, C., Alternatives to the Median Absolute Deviation. Journal of the American Statistical Association, 88, 1273-1283, 1993.
Figure 1. Functional relationship between the number p of laboratories and the correction factor bp for sR.
Figure 1. Functional relationship between the number p of laboratories and the correction factor bp for sR.
Preprints 157734 g001
Figure 2. Functional relationship between the number p of laboratories and the correction factor c p for s I ( 1 ) .
Figure 2. Functional relationship between the number p of laboratories and the correction factor c p for s I ( 1 ) .
Preprints 157734 g002
Table 1. Simulation results for each value of p : Expected value and relative standard error for s R as well as the correction factor b p . .
Table 1. Simulation results for each value of p : Expected value and relative standard error for s R as well as the correction factor b p . .
p s R Rel .   se ( s R ) b p p s R Rel .   se ( s R ) b p p s R Rel .   se ( s R ) b p
4 1.3212 0.058% 0.7569 37 1.0163 0.014% 0.9839 70 1.0084 0.010% 0.9917
5 1.1864 0.054% 0.8429 38 1.0158 0.014% 0.9845 71 1.0082 0.010% 0.9919
6 1.1490 0.044% 0.8703 39 1.0154 0.014% 0.9848 72 1.0080 0.010% 0.9921
7 1.1173 0.041% 0.8950 40 1.0149 0.013% 0.9853 73 1.0079 0.010% 0.9922
8 1.1001 0.036% 0.9090 41 1.0147 0.013% 0.9855 74 1.0078 0.009% 0.9922
9 1.0857 0.034% 0.9211 42 1.0140 0.013% 0.9861 75 1.0077 0.009% 0.9924
10 1.0737 0.032% 0.9313 43 1.0139 0.013% 0.9863 76 1.0075 0.009% 0.9925
11 1.0657 0.030% 0.9384 44 1.0138 0.013% 0.9864 77 1.0077 0.009% 0.9924
12 1.0586 0.028% 0.9446 45 1.0133 0.013% 0.9869 78 1.0075 0.009% 0.9925
13 1.0538 0.027% 0.9490 46 1.0130 0.012% 0.9872 79 1.0073 0.009% 0.9928
14 1.0494 0.026% 0.9529 47 1.0125 0.012% 0.9876 80 1.0071 0.009% 0.9930
15 1.0452 0.024% 0.9568 48 1.0124 0.012% 0.9877 81 1.0072 0.009% 0.9928
16 1.0417 0.023% 0.9600 49 1.0119 0.012% 0.9882 82 1.0071 0.009% 0.9929
17 1.0391 0.022% 0.9624 50 1.0119 0.012% 0.9883 83 1.0069 0.009% 0.9931
18 1.0365 0.022% 0.9648 51 1.0117 0.012% 0.9885 84 1.0069 0.009% 0.9931
19 1.0342 0.021% 0.9669 52 1.0115 0.012% 0.9886 85 1.0068 0.009% 0.9932
20 1.0322 0.020% 0.9688 53 1.0112 0.011% 0.9889 86 1.0067 0.009% 0.9933
21 1.0304 0.020% 0.9705 54 1.0109 0.011% 0.9892 87 1.0065 0.009% 0.9936
22 1.0292 0.019% 0.9716 55 1.0107 0.011% 0.9894 88 1.0066 0.009% 0.9935
23 1.0278 0.019% 0.9730 56 1.0105 0.011% 0.9896 89 1.0067 0.009% 0.9933
24 1.0261 0.018% 0.9746 57 1.0104 0.011% 0.9897 90 1.0065 0.009% 0.9935
25 1.0252 0.018% 0.9754 58 1.0102 0.011% 0.9899 91 1.0063 0.008% 0.9938
26 1.0238 0.017% 0.9768 59 1.0099 0.011% 0.9902 92 1.0062 0.008% 0.9938
27 1.0231 0.017% 0.9774 60 1.0096 0.011% 0.9905 93 1.0061 0.008% 0.9939
28 1.0221 0.017% 0.9784 61 1.0096 0.011% 0.9905 94 1.0061 0.008% 0.9939
29 1.0214 0.016% 0.9791 62 1.0095 0.010% 0.9905 95 1.0061 0.008% 0.9939
30 1.0203 0.016% 0.9801 63 1.0096 0.010% 0.9905 96 1.0059 0.008% 0.9941
31 1.0200 0.016% 0.9804 64 1.0092 0.010% 0.9909 97 1.0058 0.008% 0.9942
32 1.0192 0.015% 0.9812 65 1.0090 0.010% 0.9911 98 1.0058 0.008% 0.9942
33 1.0185 0.015% 0.9818 66 1.0088 0.010% 0.9913 99 1.0058 0.008% 0.9943
34 1.0180 0.015% 0.9823 67 1.0087 0.010% 0.9914 100 1.0058 0.008% 0.9942
35 1.0172 0.014% 0.9830 68 1.0086 0.010% 0.9915
36 1.0168 0.014% 0.9835 69 1.0084 0.010% 0.9917
Table 2. Simulation results for each value of p : Expected value and relative standard error for s I ( 1 ) as well as the correction factor c p . .
Table 2. Simulation results for each value of p : Expected value and relative standard error for s I ( 1 ) as well as the correction factor c p . .
p s R Rel .   se ( s I ( 1 ) ) c p p s R Rel .   se ( s I ( 1 ) ) c p p s R Rel .   se ( s I ( 1 ) ) c p
4 1.0855 0.046% 0.9212 37 1.0081 0.019% 0.9920 70 1.0043 0.014% 0.9957
5 1.0561 0.046% 0.9469 38 1.0080 0.018% 0.9920 71 1.0041 0.014% 0.9959
6 1.0550 0.040% 0.9479 39 1.0077 0.018% 0.9924 72 1.0043 0.014% 0.9957
7 1.0409 0.040% 0.9607 40 1.0077 0.018% 0.9923 73 1.0040 0.014% 0.9960
8 1.0410 0.036% 0.9606 41 1.0074 0.018% 0.9927 74 1.0041 0.013% 0.9959
9 1.0324 0.036% 0.9686 42 1.0073 0.018% 0.9928 75 1.0039 0.013% 0.9961
10 1.0321 0.033% 0.9689 43 1.0072 0.017% 0.9929 76 1.0040 0.013% 0.9960
11 1.0272 0.033% 0.9735 44 1.0068 0.017% 0.9932 77 1.0038 0.013% 0.9963
12 1.0270 0.031% 0.9737 45 1.0064 0.017% 0.9936 78 1.0040 0.013% 0.9960
13 1.0233 0.031% 0.9772 46 1.0068 0.017% 0.9933 79 1.0039 0.013% 0.9961
14 1.0231 0.029% 0.9774 47 1.0066 0.017% 0.9935 80 1.0038 0.013% 0.9962
15 1.0206 0.029% 0.9798 48 1.0064 0.016% 0.9937 81 1.0039 0.013% 0.9962
16 1.0200 0.027% 0.9804 49 1.0064 0.016% 0.9937 82 1.0034 0.013% 0.9966
17 1.0178 0.027% 0.9825 50 1.0063 0.016% 0.9937 83 1.0035 0.013% 0.9965
18 1.0173 0.026% 0.9830 51 1.0057 0.016% 0.9943 84 1.0037 0.013% 0.9963
19 1.0157 0.026% 0.9846 52 1.0059 0.016% 0.9941 85 1.0035 0.013% 0.9965
20 1.0157 0.025% 0.9845 53 1.0058 0.016% 0.9942 86 1.0036 0.012% 0.9964
21 1.0147 0.025% 0.9855 54 1.0055 0.016% 0.9946 87 1.0034 0.012% 0.9966
22 1.0140 0.024% 0.9862 55 1.0053 0.016% 0.9947 88 1.0036 0.012% 0.9964
23 1.0131 0.024% 0.9870 56 1.0055 0.015% 0.9946 89 1.0035 0.012% 0.9965
24 1.0134 0.023% 0.9867 57 1.0052 0.015% 0.9948 90 1.0036 0.012% 0.9964
25 1.0122 0.023% 0.9880 58 1.0054 0.015% 0.9946 91 1.0033 0.012% 0.9967
26 1.0121 0.022% 0.9880 59 1.0050 0.015% 0.9950 92 1.0034 0.012% 0.9966
27 1.0109 0.022% 0.9893 60 1.0051 0.015% 0.9949 93 1.0031 0.012% 0.9969
28 1.0112 0.021% 0.9889 61 1.0052 0.015% 0.9948 94 1.0032 0.012% 0.9968
29 1.0102 0.021% 0.9899 62 1.0050 0.015% 0.9950 95 1.0031 0.012% 0.9969
30 1.0102 0.021% 0.9899 63 1.0048 0.015% 0.9952 96 1.0031 0.012% 0.9969
31 1.0099 0.020% 0.9902 64 1.0051 0.014% 0.9949 97 1.0031 0.012% 0.9969
32 1.0095 0.020% 0.9906 65 1.0046 0.014% 0.9954 98 1.0031 0.012% 0.9969
33 1.0091 0.020% 0.9909 66 1.0048 0.014% 0.9952 99 1.0029 0.012% 0.9971
34 1.0092 0.019% 0.9909 67 1.0046 0.014% 0.9954 100 1.0032 0.012% 0.9968
35 1.0084 0.019% 0.9917 68 1.0044 0.014% 0.9956
36 1.0088 0.019% 0.9913 69 1.0043 0.014% 0.9958
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.
Prerpints.org logo

Preprints.org is a free preprint server supported by MDPI in Basel, Switzerland.

Subscribe

Disclaimer

Terms of Use

Privacy Policy

Privacy Settings

© 2025 MDPI (Basel, Switzerland) unless otherwise stated