Teaching the Utilization of Quantitative Statistical Techniques for Industrial Application: A Case Study

INTRODUCTIONAs a long time instructor of Quantitative Business courses at a large southeastern University, I am often asked a variety of questions. However, I was recently intrigued by a question posed by a nontraditional student after completing an Applied Statistics course. The student questioned how to solve an actual business problem using the techniques he had just learned. The discussion left me with a nagging question, whether quantitative courses should be taught as stand-alone courses (that is, only statistics should be taught because the textbook only covers statistics) or should they be formatted to answer typical situations encountered in the business environment that utilize several different quantitative techniques. Upon further reflection, I questioned not only the combining of techniques from different analytic courses, but also whether the relationship between related techniques such as ANOVA and regression should be emphasized more fully. In other words, could most undergraduate business students reasonably be expected to connect a series of quantitative techniques to solve a variety of business problems [Yoder & Kurz, 2015]?To keep quantitative courses such as management science and statistics in the Business curriculum, faculty has abandoned teaching algorithms and their circumstance to teaching applications software utilization. This places the burden of tying the relationships of these quantitative methods to student forums/study groups to bridge this gap [Haughton and Kelly, 2015; Palocsay and Markham, 2014].The student, who had previously worked in the quality department of a powder coating operation, explained that over a three year period the powder coating operation had evolved (due mainly to trial and error) to a very acceptable quality level of powder-coated panels. The question formulated became, "would it be possible to apply quantitative methods covered in class to optimize a manufacturing process?" The student provided the early trial data without revealing the final settings for the powder coating equipment as determined at the manufacturing operation. The goal, therefore, was to determine the final equipment settings quantitatively and compare those settings to the final settings employed in the manufacturing facility. In other words, would the techniques taught in class greatly reduce the time required to reach optimality as taught, or is trial and error the only true method for "tweaking" a manufacturing process? To achieve the above-mentioned goal, we needed to duplicate the manufacturing facility's results using a three step process: first, determining which manufacturer of the spray gun to utilize; second, selecting the powder manufacturer; and third, optimizing the spray gun settings. In optimizing the spray gun settings the aim was to reduce the amount of variation in the coating thickness within a panel and from panel to panel to keep consistency in the end product.BACKGROUNDIn powder coating, an electrically charged powder is sprayed onto a grounded metal work piece. Powder is held to the part being coated through electrostatic attraction, until heat is added to hold the powder together and cure it to create a high quality and durable finish.A complete electrostatic powder coating system is comprised of a powder feeder, a control unit, an electrostatic voltage generator, a powder spray gun, and a powder booth/recovery system (see Figure 1).Once in the feeding unit, the powder is spread by compressed air into a liquid state. Next, the powder is siphoned from the feeding unit by the movement of high speed air flowing through a venturi and driven through tubing to the spray gun. The feeder provides a controlled flow of powder to the spray gun. Because air and powder volume are controlled independently of one another, dilution ratios may be adjusted to obtain the desired thickness of coverage that is needed to meet specific coating requirements. …