STA-3300 Applied Statistics 2 - 10 stp

The course is intended for students who collect and analyze data, and who need to know more about the use of statistics within their own field. Students attending the course should be familiar with the concepts of random variables and distributions, normal distribution, estimation, confidence interval, and hypothesis testing.

This course teaches the following:

• Single and multiple linear regression analysis

• Single and multiple correlations

• Confounding and interactions

• One- and two-way analysis of variance

• Understanding the regression model and the assumptions behind it

• Testing the significance of the different variables

• Selecting a best-fitting model based on regression diagnostics

• Randomized blocks and model design

The concepts will be presented through many examples and practical exercises with real data from several areas using Python as statistical software. A short introduction to Python will be given as well.

Bachelor of science degree or equivalent.

Application code is 9371.

Knowledge - The candidate will learn:

• The assumptions for using regression- and analysis of variance.

• The two types of independent variables (covariates): measurement- and nominal (categorical) variables.

• The concept of confounding.

• The concept of interaction between independent variables.

• Polynomial regression and transformation of variables.

• Criteria (R2, F, Mallows’ Cp) and strategies for selecting the best model (forward, backward, and stepwise regression, and best subset of variables).

• One- and two-way analysis of variance, balanced design, fixed and/or random factors.

• The general linear model as a framework for multiple regression and Anova models.

Skills - The candidate will become familiar with:

• Prepare data and perform regression analysis in Python.

• Use regression diagnostics, such as analysis of residuals to identify possible outliers, check the model assumptions, and address problems with collinearity.

General expertise - The candidate will learn how to:

• Choose a reasonable linear model to analyze his/her data sets.

• Explain the regression results to a person who is not familiar with statistical thinking.

• Interpret regression results. Understand the significance and the importance of the selected independent variables in explaining the response variable.