Taguchi's robust parameter design is an experimental procedure based on reducing the system variability that noise factors cause. The results, obtained under unrealistic assumptions about noise, may mislead practitioners when it comes to improving quality in robust design. For example, many hydrological data and the multi-path fading of a signal in wireless communication systems are positively skewed and cannot be modeled by any normal distribution. This manuscript focuses on the case where noise factor follows gamma distribution and investigates its true effects on the following: the response, the choice of an estimator in modeling, and the estimation of optimum factor settings. Then, a new density function is proposed for a given response under the gamma effect. A design of simulated experiments with gamma noise is conducted and two related examples are presented to illustrate the findings.