What This Document Is
These are lecture notes from STAT 572: Statistical Methods for Bioscience II, offered at the University of Wisconsin-Madison. The notes cover advanced statistical modeling techniques, specifically focusing on hypothesis testing and model selection within a biological context. The material builds upon foundational statistical knowledge and delves into more complex scenarios involving multiple predictor variables and the challenges that arise when those variables are interconnected. It explores methods for comparing different statistical models to determine the best fit for a given dataset.
Why This Document Matters
Students enrolled in advanced biostatistics or related bioscience courses will find these notes particularly valuable. They are ideal for reinforcing concepts presented in lectures, preparing for assignments, and building a strong foundation for more advanced statistical analysis. Researchers and practitioners needing to apply statistical modeling to biological data will also benefit from understanding the principles discussed within. These notes are most useful when studied *in conjunction* with course lectures and assigned readings.
Common Limitations or Challenges
These notes represent a specific instructor’s presentation of the material and do not substitute for a comprehensive textbook or independent study. They are focused on the theoretical underpinnings and application of statistical methods, and do not include detailed computational code or step-by-step instructions for performing analyses. The notes assume a prior understanding of basic statistical concepts like regression and hypothesis testing. Access to statistical software (like R) is also beneficial for fully grasping the practical implications of the discussed methods.
What This Document Provides
* An overview of significance testing procedures when dealing with multiple predictor variables.
* Discussion of methods for comparing statistical models, including those based on nested structures.
* Exploration of the consequences of correlated predictor variables (multicollinearity) on statistical inference.
* Introduction to model selection criteria, designed to balance model fit and complexity.
* Examination of how different statistical tests can yield varying results when applied to the same data.
* Conceptual understanding of measures of model fit, such as adjusted R-squared.
