2. Balanced One-Factor Random Models
Society for Industrial and Applied Mathematics eBooks · 2005
2.1 Introduction The first model we consider is the balanced one-factor random design. This model describes a gauge R&R study in which a single operator selects a random sample of p parts and measures each part r times using the same measurement gauge. Table 2.1 presents a subset of data reported by Houf and Berman [32]. In this example, the monitored parts are power modules for a line of motor starters. The response is the thermal performance of the module measured in C° per watt. Measurements are taken on the device using a thermal resistance measuring instrument. The data shown in the table represent measurements taken by a single operator. Each response has been multiplied by 100 for convenience of scale. The p = 10 parts were sampled at random from the manufacturing process, and r = 2 replicate measures were made on each part. We now present the model used to analyze the data in Table 2.1. The completed analysis is reported in Section 2.5.