This course takes the participants through the principles of Ordinary Least Squares (OLS) and linear regressions.  Participants will explore relationships between variables and build models using Excel functions and features such as SLOPE, INTERCEPT, LINEST, Solver and the Analysis Toolpak.

Advanced topics include multivariate regression and statistical significance testing using t-tests and F-tests. Guided exercises support the students to ensure practical understanding and immediate applicability of the concepts.

Part I (8 hours)

Introduction to regression

• Exploring relationships: scatter plots, variance, covariance, and Pearson’s correlation coefficient.
• Visualizing relationship : scatter plots.
• The functional form and building a linear regression model: terminology, slope, intercept, residuals, and how to determine the best fit line
• Linear regression with one predictor
• Using Excel tools: Data Tables, Solver, and the functions SLOPE, INTERCEPT, LINEST
• Determining goodness of fit: explained and unexplained variance, coefficient of determination
• Analysing the residuals:  homoskedasticity vs. heteroskedasticity, using QQ plots on residuals as a means to verify normality
• Statistical significance
• The t-statistics, F-statistic, and evaluating coefficient significance
• Interpreting the Analysis Toolpak output
• Testing for Homoskedasticity

Part II (8 hours)

Transforming a nonlinear relationship into a linear model, the power and exponential laws, and approaching a problem parametrically.

Regression with more than one predictor

• Multicollinearity, the correlation matrix from Analysis Toolpak
• Partial effects,
• Adjusted R², model significance and model comparisons (F-test)

Logistic regression and how it differs from linear regression

• Logistic regression functional form
• What are the odds?
• Computing the model parameters

• Build and interpret linear regression models using one or more predictors using Excel’s various features, and learn when each feature is more adequate than others.
• Understand how to assess model performance and goodness-of-fit
• Communicate regression results clearly and confidently
• Learn to apply transformations when relationships appear non-linear
• Perform logistic regression for binary outcomes

Classroom delivery coupled with hands-on application of the concepts through example cases in interactive sessions.