Processing kodak motion picture films, module 3 analytical procedures

Reproducibility
Customer Standard Deviation, 1sc & 95 Percent
Confidence Estimate (not including bias)
Reproducibility or customer standard deviation (1sc ) is an
estimate of the variability a customer could expect when
submitting a sample to any Photoprocessing Quality
Services laboratory, where any trained analyst could test the
sample using any instrument on any day.
The 95 percent confidence estimate (calculated using the
customer standard deviation) around a single test result will
include the mean value 95 percent of the time.
Three ferricyanide bleach samples were analyzed by four
analysts, on two different days. Duplicate analyses were
performed on each sample, on each of the two days. These
samples were:
1. A “fresh” tank solution prepared at 40.000 g/L
K3Fe(CN) 6 .
2. A “seasoned” ferricyanide bleach sample analyzed at
46.185 g/L K3Fe(CN) 6 .
3. The same “seasoned” solution, as in number 2, above,
analyzed in the same manner, after making a standard
addition of 13.901 g/L K3Fe(CN) 6 .
Sample
“Fresh” at
“Aim”
(40.000 g/L
K 3Fe(CN) 6)
“Seasoned”,
As Received
“Seasoned”
with
Standard
Addition
K 3 Fe(CN) 6 (Potentiometrically)
Mean
(g/L
K 3Fe(CN) 6)
N
Reproducibility
Standard
Deviation, 1S c
(g/L K 3 Fe(CN) 6 )
95 Percent
Confidence
Estimate
(g/L
K 3Fe(CN) 6)
40.400 16 0.090 ± 0.19
46.185 16 0.131 ± 0.28
59.795 16 0.188 ± 0.40
Bias
Bias is a statistically significant deviation of the mean from
the known mix level at a 95 percent confidence level. It is
determined for “fresh” samples only. Bias is not determined
for “seasoned” samples, since the component concentration
level was not determined independently of the test method.
A statistically significant high bias of (+0.400 g/L) for
K 3 Fe(CN) 6 was found for a Process ECN-2 “fresh” tank
developer sample. However, the bias was judged not to be
practically significant.
Recovery
Recovery is used instead of bias for seasoned samples, since
the component concentration level was not determined
independently of the test method. It is defined as the
calculated mean for the seasoned sample with a standard
addition of the component minus the mean for the seasoned
sample, divided by the actual amount of the standard
addition. It is expressed as a percentage. Statistically, the
recovery of 97.9 percent for Process ECN-2 was
significantly different from 100 percent, but was judged not
to be practically significant.
Visual Titration
Repeatability Standard Deviation (1sr) ) & 95 Percent
Confidence Estimate (not including bias)
Repeatability standard deviation is an estimate of the
variability one trained analyst should be able to obtain under
favorable conditions (analyzing a sample, with one
instrument, within one day).
The 95 percent confidence estimate (calculated using the
repeatability standard deviation) around a single test will
include the mean value 95 percent of the time.
To obtain the repeatability data, a single skilled analyst
performed five (5) replicates on each of the following
solutions during methods development (this procedure was
performed by both potentiometric and visual end point
detection):
1. A “fresh” ferricyanide bleach prepared with all
components at their respective “working tank” aim
concentrations.
2. A “seasoned” ferricyanide bleach analyzed as received
at 46.350 g/L.
3. The same “seasoned” solution as in number 2, above,
reanalyzed after making an analytically weighed,
standard addition of 13.901 g/L K3Fe(CN) 6 .
Sample
(ECN-2)
“Fresh” at
“Aim”
(40.002 g/L
K 3 Fe(CN) 6 )
“Seasoned”,
As Received
“Seasoned”
with
Standard
Addition
Mean
(g/L
K 3Fe(CN) 6)
K 3 Fe(CN) 6 (Visually)
Bias
Bias is a statistically significant deviation of the mean from
the known mix level at a 95 percent confidence level. It is
determined for “fresh” samples only. Bias is not determined
for “seasoned” samples, since the component concentration
level was not determined independently of the test method.
A statistically significant low bias of (-0.067 g/L) for
K 3Fe(CN) 6 was found for a Process ECN-2 “fresh” tank
developer sample. However, the bias was judged not to be
practically significant.
2 Processing KODAK Motion Picture Films, Module 3, Analytical Procedures H24.03
N
Repeatability
Standard
Deviation, 1S r
(g/L K 3 Fe(CN) 6 )
95 Percent
Confidence
Estimate
(g/L
K 3 Fe(CN) 6 )
39.935 6 0.034 ± 0.09
46.350 5 0.069 ± 0.19
59.778 5 0.046 ± 0.13