Editing Design for Six Sigma (section)

===Critics===
Response surface methodology and other DFSS tools uses statistical (often empirical) models, and therefore practitioners need to be aware that even the best statistical model is an approximation to reality. In practice, both the models and the parameter values are unknown, and subject to uncertainty on top of ignorance. Of course, an estimated optimum point need not be optimum in reality, because of the errors of the estimates and of the inadequacies of the model. The uncertainties can be handled via a Bayesian predictive approach, which considers the uncertainties in the model parameters as part of the optimization. The optimization is not based on a fitted model for the mean response, E[Y], but rather, the posterior probability that the responses satisfies given specifications is maximized according to the available experimental data.<ref>{{Cite journal |last=Peterson |first=John J. |date=2004-04-01 |title=A Posterior Predictive Approach to Multiple Response Surface Optimization |url=https://doi.org/10.1080/00224065.2004.11980261 |journal=Journal of Quality Technology |volume=36 |issue=2 |pages=139–153 |doi=10.1080/00224065.2004.11980261 |s2cid=116581405 |issn=0022-4065|url-access=subscription }}</ref>
 
Nonetheless, response surface methodology has an effective track-record of helping researchers improve products and services: For example, [[George Box]]'s original response-surface modeling enabled chemical engineers to improve a process that had been stuck at a saddle-point for years.<ref>{{Cite web |title=Response Surfaces, Mixtures, and Ridge Analyses, 2nd Edition {{!}} Wiley |url=https://www.wiley.com/en-us/Response+Surfaces%2C+Mixtures%2C+and+Ridge+Analyses%2C+2nd+Edition-p-9780470053577 |access-date=2022-04-09 |website=Wiley.com |language=en-us}}</ref>