Regression Analysis
A quantitative method that studies how one or more input variables relate to an outcome variable and derives a statistical equation that explains and predicts that relationship.
Key Points
- Examines independent (input) variables against a dependent (output) variable to quantify their relationship.
- Produces coefficients and fit metrics (for example, R-squared) that support prediction and forecasting.
- Can be simple (one predictor) or multiple (many predictors), and may use linear or nonlinear forms.
- Relies on quality historical data; check assumptions, manage outliers, and remember correlation is not causation.
Example
A construction project team analyzes past jobs where inputs were square footage, number of floors, and material grade, and the output was total cost. They fit a regression model to estimate cost for a new building based on its planned size and complexity.
PMP Example Question
During estimating, the team analyzes historical data on number of features, code complexity, and team experience to quantify their effect on testing effort and to build an equation that predicts future testing hours. Which technique are they using?
- Analogous estimating
- Regression analysis
- Critical path method
- Monte Carlo simulation
Correct Answer: B — Regression analysis
Explanation: Regression analysis develops a statistical relationship between multiple inputs and an output to create a predictive equation.