Statistics · Commerce
Regression
Introduction
This guide explains regression within Statistics. It builds from the topic’s central ideas to the way its elements influence decisions and outcomes, giving you a clear structure for applying the concepts to questions and case situations.
Regression models the conditional relationship between
- Regression models the conditional relationship
Regression models the conditional relationship between an outcome and one or more explanatory variables.
- Name the population, variables and
Name the population, variables and purpose before choosing a statistical procedure for regression.
Dependent variable, explanatory variable, slope
- Dependent variable, explanatory variable, slope
dependent variable, explanatory variable, slope, intercept, residual.
- Connect each term to what
Connect each term to what is measured, summarized or inferred; distinguish descriptive results from conclusions about a population.
Fitted line can estimate average sales associated
- Define the question and measurement
Define the question and measurement scale; inspect data quality and choose a suitable method.
- Calculate or display the result
Calculate or display the result, then interpret its units, direction and uncertainty in context.
- A fitted line can estimate
A fitted line can estimate average sales associated with advertising spend within the observed range.
- Use a labeled table, diagram
Use a labeled table, diagram or formula only when it directly answers the question.
Fitted association is not automatically causal
- A fitted association is not
A fitted association is not automatically causal; inspect residuals, assumptions and extrapolation risk.
- Check assumptions, unusual values, missing
Check assumptions, unusual values, missing data and the limits of comparison before reporting a conclusion.
Conclusion
To revise regression, define its main ideas, show how the key elements connect, and apply them to the facts of a relevant situation. Check that each conclusion follows from the evidence and the purpose of the decision.