The normal probability plot shows the percent of acceptances for each reference value. The attribute gage study regresses the probability of acceptance on the reference values to estimate bias and repeatability.

Minitab provides the fitted line equation in the gage information table and displays the fitted line on the normal probability plot.

The R-sq value indicates the percentage of the variation in the observed responses that is explained by the model.

For the length data, the reference values increase as the percent of acceptance increases. Because you specify the lower tolerance limit, you expect this relationship.

The fitted line equation is (-1670.60 + 2379.41 * Reference).

The R-sq value, 0.903159, indicates that approximately 90% of the variation in the observed responses is explained by the model.