Home / Statistical Tools / MSA / Create MSA Template / Create MSA Template How-To

Create MSA Template How-To¶

This walkthrough builds a blank data collection sheet for a study of three operators, ten parts and two trials, then runs the analysis straight from it.

There is no data to paste. The point of a template is that it is empty: you print it or fill it in on screen as the measurements are taken, and the analysis reads it directly afterwards.

Steps¶

Open the dialog

With any workbook open, choose QXL Stat Tools > MSA / Gage R&R > Create MSA Template (Optional).

Unlike the analysis tools, this one does not read a selection, so there is nothing to select first. The template goes onto a new worksheet of its own.
Choose the data format

Under Data Format, leave Quantitative (Variable) chosen. That is the format the Crossed, Nested and Extended analyses read.

Qualitative (Attribute) builds a sheet for ratings instead, read by Attribute MSA: Crosstabulations rather than by the three analyses this walkthrough is for.
Say whether you have reference values

Under Reference Values, leave I do not have reference values chosen.

I have reference values for the parts adds a Ref column to the sheet for the known true value of each part. When that column holds values, the analysis report adds the bias and linearity tables.
Set the size

Under Template Size:
- Number of operators (appraisers): 3
- Number of trials (reps): 2
- Number of parts: 10
The sheet you get holds 3 times 2, or six, measurement columns and ten part rows, so sixty measurement cells. The smallest study the dialog accepts is 1 operator, 1 trial and 2 parts.
Choose Finish

Quantum XL adds a worksheet named MSA Template after the sheet you were on and writes the template there.

What the template looks like¶

The block is laid out for a person with a gage in their hand, not for a spreadsheet formula.

A title cell reading MSA Template 1: Gage R&R: XbarR, EMS, and REML.

A two-row header. The first row holds Part and then one heading per operator, Operator (Appraiser) 1, Operator (Appraiser) 2 and Operator (Appraiser) 3, each merged across that operator's trial columns. The second row holds Trial 1 and Trial 2 under each operator. Then one row per part, numbered down the Part column.

You can rename an operator by typing over its heading, and the analysis reads the name from there.

So the grid is wide: one column per operator and trial combination, which for this study is six columns of measurements beside the part numbers.

A Gage Information area, with labelled cells to fill in: Gage name: (filled in as My gage), Gage No.:, Gage type:, Part name:, Part No.:, Characteristics: and Specifications:.

A limits and signoff area to the side, holding USL:, LSL:, Date: and Performed by:. The date is filled in with today's date. Enter the limits here if you know them: the analysis reads them back off the sheet, so they do not have to be typed again.

Printed instructions under the heading How to use this template, in these words:

Enter the gage and part information.

Enter the USL and LSL if known (needed for %Tolerance).

Have each operator (appraiser) measure each part in random order; type each result under that operator's trial column.

When the table is full, run the MSA analysis on this template.

Filling it in¶

Work down each operator's pair of columns, one measurement per cell. Leave a cell empty if a measurement was not taken; an empty cell is read as a missing measurement, never as a zero.

You can put more than one template on the same sheet. Add Template adds another tab to the dialog before anything is written, starting from a copy of the selected tab's settings, and each block is written below the previous one.

Running the analysis from the template¶

With the filled-in template on screen, choose QXL Stat Tools > MSA / Gage R&R > Crossed.
Quantum XL finds the template and shows the small MSA Analysis Method dialog rather than asking you to assign columns. Estimation Method offers:
- Automatic (balanced: EMS; unbalanced: REML)
- Restricted Maximum Likelihood (REML)
- Expected Mean Squares (EMS) - Equivalent to ANOVA when balanced
- Xbar and Range (XbarR)
Beneath it, Specification Limits shows one row for this template, already carrying whatever the template's LSL: and USL: cells hold and blank on either side that was left empty. Three checkboxes follow: Display Confidence Intervals ticked, Display gage performance curve clear, and Part variation isn't representative clear.

Leave all of it alone. Choose Automatic and press OK.
The report appears on its own worksheet, exactly as it would from data you had arranged yourself.

The Data tab is skipped because the reshaping is done for you

The analyses read one row per measurement, with the part and the operator in their own columns. The template's grid is the other shape: one column per operator and trial.

Running the analysis from a template reshapes the grid into that row-per-measurement form automatically, which is why there are no columns to assign. Each operator is named from its own heading cell in the first header row, so typing a name over Operator (Appraiser) 1 uses that name on the report. The template as written gives Operator (Appraiser) 1, Operator (Appraiser) 2 and Operator (Appraiser) 3, and a heading cleared to blank falls back to the same wording.

The specification limits and the gage information are read back off the sheet at the same time.

A template and your own layout, compared¶

Both run the same analysis on the same measurements. What differs is the route, and one default:

	A template	Your own layout
shape	wide: one column per operator and trial	one row per measurement
starting Crossed or Nested	the MSA Analysis Method dialog, no columns to assign	the Data tab, assigning each role column
starting Extended	the full dialog, prefilled from the template	the full dialog
limits and gage info	read back off the sheet	typed into the dialog
Misclassification sweep charts	on	off unless ticked on the Options tab

The Crossed How-To walks through the row-per-measurement route.