Statistical approaches

Scientists frequently use statistics to determine the validity of their results. Usually a statistician gets involved and then there are statistic tests that are run to test the hypothesis. Most frequently these statistics tests are run with one specific goal. Test to determine whether these observed phenomenon are just random events i.e. signal vs. noise. In statistics as most scientists know it, this is testing the null hypothesis. However, very few scientists realize that there are two schools of statistics Bayesian and Frequentism, with the Fisher frequentism being the one used more often for data analysis. There has been some discussion in the literature in fields such as clinical trials, Wildlife analysis, Psychology that maybe we have not been giving Bayesian enough importance and instead following the trend with conventional analysis. Future discussion in this blog will discuss these dilemma’s. For an interesting read try this article “The False Dilemma: Bayesian vs. Frequentist” by Jordi Vallverdu in Electronic Journal for Philosophy, 2008.