Fun with Data Analysis
A butterfly develops through a process known as metamorphosis. The term metamorphosis comes from the Greek word transformation or change in shape (Butterfly lifecycle). A butterfly starts out as a caterpillar. When the caterpillar becomes full grown, it then transforms into a pupa. This pupa hatches the butterfly completing the cycle of change. The constant throughout this process is transformation or change. Research is similar in that information or data is transformed in support of new knowledge. The cycle begins with the development of a theory, followed by the formation of a testable hypothesis. Information is collected and transformed in support of new ...view middle of the document...
Conducting an appropriate amount of sampling helps create a well-balanced set of information. This information can be used to support the creation of new knowledge. Sampling a population in support of drawing conclusions is not only conducted in research but in manufacturing facilities throughout the world. ANSI/ASQ Z1.4 is an attribute sampling guidance document. It guides the user on sampling selection. The selection of sampling is predicated on a user’s desired quality or confidence level. Using this confidence level helps determine the number of units necessary to draw conclusions surrounding the entire population of interest. Table 1 is an example of a table supporting accepts on zero defects found. In the event of one defect the population of interest would be considered defective.
In the above table AQL represents acceptable quality level. Let’s assume a firm desires 95% confidence that a population of 10,000 units is free of defects. Using Table 1 the firm would be required to evaluate 29 units. The 95% confidence interval is represented by an AQL of 2.5. This AQL sampling assumes normal distribution and therefore each tail of the distribution is represented by 2.5 or 2.5%. Taking 2.5% multiplied by 2 equals 5% probability of falsely accepting bad material. In the event all 29 units met the predefined criteria, the firm would be able to claim the entire population is acceptable with 95% confidence.
The above example demonstrates using a proportion of the population can allow for conclusions regarding the entire population. Adequate sampling alone will not facilitate appropriate conclusions. Measurement trust is also required. A researcher could evaluate the entire population of interest, however if the measures are not appropriate or are not reliable and valid, then the information cannot support conclusions. Drawing good research conclusion requires the combination of both measurement trust and adequate sampling.
An Experimental Analysis of Dynamic Incentives to Share Knowledge
Deck and Erkal (2013) stated that collaboration with rivals can assist with speed and innovation of new products. Their research concluded collaboration unravels as the product moves towards its initial stages of the productive lifecycle. As a product moves closer to commercial exposure, the profitable opportunities become clearer and firms attempt to sever relationship ties in order to protect future revenue. To study this phenomenon, Deck and Erkal modified a measurement system used by Erkal and Mineheart (2008), to study the dynamics of competitor sharing during the innovation process.
This study used a within-subjects design to evaluate and predict success of the model. A total of 96 subjects were selected from Melbourne University to participate in computer simulated models. These simulations were designed to represent the product development cycle and commercial decision points. Each cycle contained...