A Type 1 Gage study is used to evaluate the bias and repeatability of a measurement device by repeatedly measuring a known reference sample a number of times. By comparing the measurements recorded to the reference, we can evaluate the bias and repeatability of the measurement system. This can be compared to the tolerance to check if the variation in the measurement device is small enough to be acceptable.
In this study, we will look at measurements on fill volumes. A known volume of 15 ml is measured 20 times. The specifications for the process are 14.25 and 15.75. The worksheet Type 1 Gage.mtw contains 20 measurements on this syringe in the Measure column.
- Open the worksheet
Type 1 Gage.mtwby using Open Worksheet… from the File menu. - Navigate to Stat | Quality Tools | Gage Study and click on Type 1 Gage Study.
- Enter
Measurein the Measurement data field. - Enter
15in the Reference field. - Under Tolerance, enter
1.5in Upper spec - lower Spec. - Click on OK.
Type 1 Gage studies will output a Bias measure and a Capability measure, namely Cg and Cgk. A one-sample T-test is run on the mean of the measurements against the reference value to check for bias. The results of this data should show a significant bias with a mean of 14.916.
The run chart will display the measurements in the order they appear in the worksheet. The data should be presented in the collection order to allow a check for trends or patterns in the results.
The red lines plotted on the chart represent as 10 percent of the tolerance interval. Ideally, a good measurement device should have measurements well within the interval. Results exceeding the interval show a measurement device with too much variation. The measures displayed are as follows:
- Cg: This is calculated by dividing 20 percent of the tolerance width by the study variation: the width of the red lines divided by six times the standard deviation. The percent of the tolerance is given as 20 percent by default but can be changed within the Type 1 Gage study options.
- Cgk: This is calculated from the bias and the Gage variation. In this example, we should see that the results are close to, and cross the lower limit, causing Cgk to be much lower than Cg.
- %Var(Repeatability): This compares the gage repeatability to 20 percent of the tolerance.
- %Var(Repeatability and Bias): This compares both repeatability and bias with 20 percent of the tolerance.
Ideally, Cgk is above 1.33.
The Type 1 study only checks the variation around one sample. As such, it only gives an indication of bias around the reference value. More in-depth checks of bias can be run with a Gage linearity and bias study, which checks the variation of the bias across a range of reference samples.
Usually, Type 1 studies are used as initial checks on a measurement device; uses, for example, might include assessing a new measurement device for suitability or assessing a current device for its suitability for use in a new application. One would then look at Gage R&R studies to examine operator variation, or bias and linearity studies to observe bias across a range.