Developing Models for ADV Statistics
Craig Lazarski, US Math, Adv. Statistics
Tinkerplots is an interactive software tool that allows students to design simulations to test hypotheses. In Advanced Statistics students are tasked with developing a model that they can use to evaluate the likelihood of certain events occurring. For example, we posed a question regarding a man trying to cheat the health care system by pretending he was deaf. The students developed a model using Tinkerplots that illustrated what the results would be for someone that was truly experiencing hearing loss:
The model consists of a distribution of responses that occurred under the assumption that a person is truly guessing when they respond to 100 questions regarding their hearing. The students are quickly able to deduce that the theoretical value is 50 questions answered correctly but they also see the natural variation in responses. They were then able to ask questions about when do we believe the person is truly deaf or misleading and they arrived at guidelines such as 40 to 60 correct responses would indicate a person that is deaf but if a person responded to fewer than 40 we have suspicion that they are purposely misleading us since the likelihood of that occurring under our assumption is not high. As we develop more mathematical tools we can step away from Tinkerplots and use the normal distribution and formal hypothesis testing but the foundation for inferential thinking was developed through these explorations using Tinkerplots. Furthermore, as we develop more complicated hypotheses such as those related to multiple proportions and sample means we can return toTinkerplots to test our theories before exploring the formal mathematics that make them work.
Collaborative Paper written with ADV Biotechnology Skill Set: harris mccopin report-151iweo
Tinkerplots Video Showing How Data Can Be explored with the Tool
