Hitachi: True Experimental Design
As mentioned in previous reports, Hitachi Automotive Products, Los Angeles (HAPLA) is a major supplier and producer of remanufactured (remanned) alternators, starters and other electromechanical automotive parts. A vital part of providing the quality product that has earned HAP-LA a high reputation is the consistent quality assurance testing that takes place throughout the production process. In order to conduct a proper quality assurance test on a product, experimental design research must occasionally be applied to determine if there is a problem in the product and the source of the problem.
True Experimental Design Research
"Experiments are studies involving intervention by the researcher beyond that required for measurement" (Cooper & Schindler, 2002, p. 381). Essentially, this means that the researcher conducting the study manipulates one variable, the independent variable, to observe the effect on the dependent variable and analyzes that data that is provided from this manipulation.
Several different methods of experiments can be utilized to find the outcome, including experiment versus control, random sampling, random assignment, and double blind (Wu, 2002a).
In contrast to the correlational research method, experimental design research provides sound scientific data while correlational studies may not always determine the causal effect between the variables under study (Wu, 2002b). However, experimental design can fall short in objectivity since the researcher is the one who manipulates the variables to prove a hypothesis. This lack of objectivity, if it is not addressed properly, may lead to results that exhibit the Hawthorne Effect, in which the process or outcome is caused by the obtrusive observation during the study (Krippendorf, 1986), or the Pygmalion Effect, in which the belief becomes reality through a self-fulfilling prophecy (McNatt, 1998). The researcher should avoid either of these potential results through ensuring complete objectivity through the use of blinds and/or double...