Using structural equations modeling as a tool in six sigma methodology

The problem of not completing the production in planned time has taken account in this study. The completion of production after the planned due date is pointed out as "delay". Various improvements should be performed to decrease delay to customer expacted value. For effectiveness and long lasting of improvements, actions should not be planned before the root cause is clearly defined. In this study classic six sigma tools are used to define the problem and measure the current process.Problem causes can be observed or unobserved variables. Observed variables are that the effects of are directly measured. Unobserved (latent) variables are that can be measured indirectly by one or more observed variables.A method is most commonly used in literature to determine the effects of and relations among unobserved (latent) variables, Structural Equations Modeling (SEM).When potantial root causes determined by using classic cause and effect tools of six sigma methodology, it is realized that these causes are latent variables that are categorized under some root causes which are also latent variables. As a result of literture search, it is observed that SEM is a common used tool for social sciences, but the usage of SEM in engineering improvement projects and root cause analysis could not be detected. This study is the first in usage of SEM at improvement project root cause analysis.In this study cause and effect tools and FMEA are used to predict the model (determining the potantial root causes and potantial relations among them). Confirmatory Factor Analysis (CFA), which a kind of SEM analysis, is used to proove that the correlations of predicted model conforms with data. To test the relationship model, which is formed based on cause and effect tools and FMEA results, by CFA provides determining the base improvement areas and relations among causes.