Dataset: steel_rough.csv Source: R.S. Peruchi, P.P. Balestrassi, A.P. de Paiva, J.R. Ferreira, M. d. S. Carmellozi (2013). "A New Multivariate Gage R&R Method for Correlated Characteristics," International Journal of Production Economics," Vol. 144, pp. 301-315 Description: Results of a 2-Way Random Effects Model (aka Gage Repeatability and Reproducibility). The study had 12 parts, 3 operators, and 4 replicate measurements. There are the following variance components: Model y = mu + alpha_i + beta_j + (ab)_ij + e_ijk V(P) = V(alpha) V(Repeatbility) = V(e) V(Reproducibility) = V(beta) + V(ab) V(MS) = V(Repeatability) + V(Reproducibility) V(Total) = V(P) + V(MS) %R&R = 100*sqrt(V(MS)/V(T)) Signal-to-Noise Ratio: SNR = sqrt(2V(P)/V(MS)) Variables/Names Part (part) Replicate (replicate) Operator (operator) Arithmetic Average Roughness (roughave)