Six Sigma Test 17 (Analyze Stage
Indiana Council Notes) fully solved
Measurement Correlation - correct answer ✔✔Measurement Correlation
1) Comparison of measurement values from one measurement system with the corresponding
values reported by one or more different measurement systems
2) Both measurement system and standards have variation in their measured value. ---eg.
calibration of instruments
What is Proficiency testing? - correct answer ✔✔Proficiency Testing ---
"Round Robin Testing" ------Measurement system or device is compared against the mean and
St. deviation of other similar devices---->can find components of variation due to artifacts and
instruments , especially if no national reference standards are present for specific range/
condition being measured.
5 Categories of Measurement Error - correct answer ✔✔5 Categories of Measurement Error
1) Bias
2) Repeatability
3) Reproducibility
4) Stability
5) Linearity
,Sensitivity - correct answer ✔✔Sensitivity
1) A gage should be sensitive enough to detect difference in measurement as slight as 1/10 of
the total tolerance specification or process spread, whichever is smaller.
2) Inadequate discrimination will affect both accuaracy and precision of reported Values
Reproducibility - correct answer ✔✔Reproducibility
1) Reproducibility of a singe gage is checked by comparin resuls of different operators taken at
different times --->will affect both accuaracy and precision of reported Values
Accuracy - correct answer ✔✔Accuracy
1) Unbiased true value and is reported as the difference between the average of a number of
measurements and the true value.
Precision - correct answer ✔✔Precision
Same as Repeatability ---can repeat the same number by the same operator every time.
***To improve accuracy and precision, must have define test method and must be statistically
stable.
Bias - correct answer ✔✔Bias
1) Difference between output of measurement system and true value------->Lack of bias=
Accuracy
,2)**Know the formula for Bias
3) Bias is reported as a percent of process variation or percent of tolerance.-----> Negative bias
(underestimated value); Positive bias (over-estimated bias)
4) Statistical significance of bias is tested using the t-distribution-----> increase the number of
measurements increases the discriminatory power of the t-test.
Linearity - correct answer ✔✔Linearity
1) Found by obtaining reference part measurement values throughout the operating range of
the instrument and plotting the bias against the reference values.
2) Default procedure for determining linearity is to measure 10 parts 5 times.
3) % linearity = slope -----> Linearity is equal to the slope multiplied by the process variation
(process spread or tolerance)
Percent Agreement - correct answer ✔✔Percent Agreement
1) Percent agreement between measurement system and either reference value or variable
being measured can be estimated using correlation coiefficient "r"
2) r= 0 then 0% agreement ; r= +/- 1, then 100% agreement
Precision/ Tolerance - correct answer ✔✔Precision/ Tolerance
, 1) Ration between the estimated measurement error (precision) and the tolerance of the
characteristic
2) Best to have a small P/T ratio to be small to reduce the effect of measurement error<-------or
the net effect of all sources of measurement variability that cause an observed value to deviate
from the master value.
6sigma is the standard deviation due to measurement variability and the assumptions are: -
correct answer ✔✔6sigma is the standard deviation due to measurement variability and the
assumptions are:
a) Measurement errors are independent
b) Measurement errors are normally distributed
c) Measurement error is independent of the magnitude of measurement.
Precision/ Total Variation - correct answer ✔✔Precision/ Total Variation (P/TV)
1) Like P/T, it evaluates the acceptability of a measurement system
2) At P/T or P/TV increases, the ability to discriminate a change in process reduces.
3) When the measurement process is inadequate to detect part-to-part variation----> must use
a smaller measurement variation
Control Chart Methods - correct answer ✔✔Control Chart Methods
1) Mathematical Model for variability study using average and range and requires 2-3
replications by 2-3 appraisers on 1- parts.
Indiana Council Notes) fully solved
Measurement Correlation - correct answer ✔✔Measurement Correlation
1) Comparison of measurement values from one measurement system with the corresponding
values reported by one or more different measurement systems
2) Both measurement system and standards have variation in their measured value. ---eg.
calibration of instruments
What is Proficiency testing? - correct answer ✔✔Proficiency Testing ---
"Round Robin Testing" ------Measurement system or device is compared against the mean and
St. deviation of other similar devices---->can find components of variation due to artifacts and
instruments , especially if no national reference standards are present for specific range/
condition being measured.
5 Categories of Measurement Error - correct answer ✔✔5 Categories of Measurement Error
1) Bias
2) Repeatability
3) Reproducibility
4) Stability
5) Linearity
,Sensitivity - correct answer ✔✔Sensitivity
1) A gage should be sensitive enough to detect difference in measurement as slight as 1/10 of
the total tolerance specification or process spread, whichever is smaller.
2) Inadequate discrimination will affect both accuaracy and precision of reported Values
Reproducibility - correct answer ✔✔Reproducibility
1) Reproducibility of a singe gage is checked by comparin resuls of different operators taken at
different times --->will affect both accuaracy and precision of reported Values
Accuracy - correct answer ✔✔Accuracy
1) Unbiased true value and is reported as the difference between the average of a number of
measurements and the true value.
Precision - correct answer ✔✔Precision
Same as Repeatability ---can repeat the same number by the same operator every time.
***To improve accuracy and precision, must have define test method and must be statistically
stable.
Bias - correct answer ✔✔Bias
1) Difference between output of measurement system and true value------->Lack of bias=
Accuracy
,2)**Know the formula for Bias
3) Bias is reported as a percent of process variation or percent of tolerance.-----> Negative bias
(underestimated value); Positive bias (over-estimated bias)
4) Statistical significance of bias is tested using the t-distribution-----> increase the number of
measurements increases the discriminatory power of the t-test.
Linearity - correct answer ✔✔Linearity
1) Found by obtaining reference part measurement values throughout the operating range of
the instrument and plotting the bias against the reference values.
2) Default procedure for determining linearity is to measure 10 parts 5 times.
3) % linearity = slope -----> Linearity is equal to the slope multiplied by the process variation
(process spread or tolerance)
Percent Agreement - correct answer ✔✔Percent Agreement
1) Percent agreement between measurement system and either reference value or variable
being measured can be estimated using correlation coiefficient "r"
2) r= 0 then 0% agreement ; r= +/- 1, then 100% agreement
Precision/ Tolerance - correct answer ✔✔Precision/ Tolerance
, 1) Ration between the estimated measurement error (precision) and the tolerance of the
characteristic
2) Best to have a small P/T ratio to be small to reduce the effect of measurement error<-------or
the net effect of all sources of measurement variability that cause an observed value to deviate
from the master value.
6sigma is the standard deviation due to measurement variability and the assumptions are: -
correct answer ✔✔6sigma is the standard deviation due to measurement variability and the
assumptions are:
a) Measurement errors are independent
b) Measurement errors are normally distributed
c) Measurement error is independent of the magnitude of measurement.
Precision/ Total Variation - correct answer ✔✔Precision/ Total Variation (P/TV)
1) Like P/T, it evaluates the acceptability of a measurement system
2) At P/T or P/TV increases, the ability to discriminate a change in process reduces.
3) When the measurement process is inadequate to detect part-to-part variation----> must use
a smaller measurement variation
Control Chart Methods - correct answer ✔✔Control Chart Methods
1) Mathematical Model for variability study using average and range and requires 2-3
replications by 2-3 appraisers on 1- parts.
