Search for: All records

“Simulation-based inference” is often considered a pedagogical strategy for helping students develop inferential reasoning, for example, giving them a visual and concrete reference for deciding whether the observed statistic is unlikely to happen by chance alone when the null hypothesis is true. In this article, we highlight for teachers some implications of different simulation strategies when analyzing two variables. In particular, does it matter whether the simulation models random sampling or random assignment? We present examples from comparing two means and simple linear regression, highlighting the impact on the standard deviation of the null distribution. We also highlight some possible extensions that simulation-based inference easily allows. Supplementary materials for this article are available online. 

Learning standards for biology courses have called for increasing statistics content. Little is known, however, about biology students’ attitudes towards statistics content and what students actually learn about statistics in these courses. This study aims to uncover changes in attitudes and content knowledge in statistics for students in biology courses. One hundred thirty-four introductory biology students across five different instructors participated in a pre-post study of statistical thinking and attitudes toward statistics. Students performed better on the statistics conceptual inventory at the end of a biology course compared to the beginning. Student attitudes showed no change. These preliminary results suggest the potential importance for laying a conceptual foundation in statistics prior to taking biology courses with little formal statistical instruction. 

The rapid acceleration of microbial genome sequencing increases opportunities to understand bacterial gene function. Unfortunately, only a small proportion of genes have been studied. Recently, TnSeq has been proposed as a cost-effective, highly reliable approach to predict gene functions as a response to changes in a cell’s fitness before-after genomic changes. However, major questions remain about how to best determine whether an observed quantitative change in fitness represents a meaningful change. To address the limitation, we develop a Gaussian mixture model framework for classifying gene function from TnSeq experiments. In order to implement the mixture model, we present the Expectation-Maximization algorithm and a hierarchical Bayesian model sampled using Stan’s Hamiltonian Monte-Carlo sampler. We compare these implementations against the frequentist method used in current TnSeq literature. From simulations and real data produced by E.coli TnSeq experiments, we show that the Bayesian implementation of the Gaussian mixture framework provides the most consistent classification results.