Well...it depends on
what data results from the use of these gages. Are you generating pass/fail data or 'estimates' of a variables result in size? It can be either attributes or variables data...
Technically 'attribute' data is 'count' data that either has no natural rank or has a natural rank but violates other 'mathematical requirements' for adding, subtracting dividing and multiplying. These math requirements are:
- True zero
- Equal cell widths
- Equal distance between cell widths
So colors and gender have no natural rank, but categories of size, speed or other likely scale data (ordinal data) has rank but cannot be mathematically manipulated other than to describe the values in the categories as simple proportions of all of the categories. Pass/fail data falls into this category; it is nominal data. A simple count of values that fall into specified categories that are not equal in their bin size.
Attributes: Gage blocks and 'feeler gages' can be used to generate nominal data (pass/fail). Some dimension is either larger than or smaller than the gage block(s).
Variables: Gage blocks and feeler gages can also be used to determine a variables result. (Ratio data that meets all of the math requirements). Think about using a ruler tha only has 1/8" increments but your dimension is specified to 1/16". HOWEVER, this would be very chunky data; data with very little resolution. This type of data 'behaves' like ordinal data when we use statistical measures such as everage and standard deviation.