2015.14 Vitamin B (July 2018)

2015.14 (JULY 2018) BVit‐02 Total Bs ERP Evaluation Form

FOR ERP USE ONLY DO NOT DISTRIBUTE

Accuracy ‐ Analytical bias compared to reference material Certified Reference Material

Table 2

Measured Results

T stat Bias

Confidence interval 

CRM product

Analyte

Certified value Uncertainty Certified range Concentration Difference from  certified value

p ‐value

Nominal

990 987

90.1 92.8 38.8 76.0

1041 1004 1074 1067 1079 1117

6.3

13.3

SRM 1849a

Niacin (ug/100g)

1080

100

980 ‐ 1180

‐50.3

3.792

0.003

1000.5 ‐ 1058.9

1.4

‐37.3 100.9 103.1

979 977

1030 1012

49.8 68.5 50.3 46.0

Mean   

1029.7

SD   

46.0

d

T stat =

/ sd

n

where:       = mean of the differences between measured results and reference value sd  = standard deviation of the differences between results and reference value n   =  number of replicate analyses of SRM d

H 0 :       = 0,   The null hypothesis is that there is no difference between the measured results and the certified value:

d

p ‐value: is the probability (at the 95% levels of confidence) of obtaining the test statistic assuming that the null hypothesis is true p ‐value < 0.05 indicate a difference of statistical signifance between measured results and SRM.

95% confidence interval: The true population of the measured value expected to fall within the confidence interval estimate 95% of the time

sd

sd

t crit

t crit

lower limit  = mean ‐

upper limit = mean +

n

n

crit t

where:           =  critical values of the t‐distribution

sd  = standard deviation of the differences between results an reference value n   =  number of replicate analyses of SRM