Skip to main content Skip to course navigation

Correlation and Regression Analysis

Learning Objective

By the end of this lesson, the learner should be able to:

  • Explain the purpose of correlation and regression analysis.
  • Use a scatter diagram as an initial visual check for a relationship.
  • Understand the strength and direction of linear association.
  • Understand how regression models a response using predictor variables.
  • Recognize the importance of residual analysis.
  • Connect statistical relationships with practical process improvement.

1. Introduction

Correlation and Regression Analysis are statistical techniques used to study relationships between variables.

Correlation measures the strength and direction of a relationship, while regression models a response in relation to one or more predictor variables and can support prediction.

For a Black Belt, these methods can help investigate whether changes in one variable are associated with changes in another and whether a useful model can be developed for the process.

2. Scatter Diagram — The First Visual Check

A scatter diagram plots observations for two variables.

It provides an initial visual check for a possible relationship before statistical calculations are performed.

For example, a Black Belt may plot:

Machine Temperature → Defect Rate

The scatter diagram can help reveal whether the data appear to show a relationship and whether further correlation or regression analysis may be appropriate.

3. Correlation

Correlation describes the strength and direction of linear association between two variables.

The correlation coefficient is commonly represented by r.

The Black Belt can use the correlation coefficient to describe the observed linear relationship between variables.

However, correlation should be interpreted carefully.

Important Principle

Correlation does not by itself prove causation.

Two variables may show a strong association without one variable necessarily causing the other.

Therefore, the Black Belt should combine statistical analysis with process knowledge and appropriate investigation.

4. Regression Analysis

Regression models the response in relation to predictor variables.

In a process-improvement context, this can help the Black Belt understand how a response changes in relation to a predictor and can support prediction.

For example:

Predictor: Machine Temperature
Response: Defect Rate

A regression model can be used to study the relationship between these variables.

5. Residual Analysis

A residual is the difference between an observed value and its corresponding fitted value from the regression model.

Residuals help the Black Belt assess model adequacy.

Residual plots can therefore provide information about whether the selected model adequately represents the observed data and whether further investigation may be required.

A Black Belt should not rely only on the regression equation or correlation coefficient. The adequacy of the model should also be considered.

6. Relationship Between the Tools

The analysis can follow a logical sequence:

Scatter Diagram → Correlation → Regression → Residual Analysis → Process Interpretation

Each step provides additional information.

Scatter Diagram

Visual examination of the relationship.

Correlation

Description of the strength and direction of linear association.

Regression

Model the response using predictor variables.

Residual Analysis

Assess the adequacy of the fitted model.

7. Tools & Techniques

The source identifies the following tools:

  • Scatter diagrams
  • Correlation coefficient r
  • Regression equation
  • Residual plots
  • Minitab
  • Excel

8. Application Example

Machine Temperature and Defect Rate

A manufacturing company wants to investigate whether machine temperature is related to defect rate.

The Black Belt:

  1. Collects temperature and defect-rate data.
  2. Creates a scatter diagram.
  3. Examines the relationship.
  4. Calculates the correlation coefficient.
  5. Uses regression to model the response.
  6. Reviews residuals to assess model adequacy.

This provides a structured way to investigate the relationship before deciding what process action may be appropriate.

9. Case Study

Injection Molding Plant

An injection molding plant investigated the relationship between process conditions and defects.

The analysis found a strong relationship with:

r = 0.85

The company then controlled temperature within the optimal range.

As a result, defects were reduced by 25%.

Black Belt Learning

The case demonstrates how relationship analysis can support process improvement when combined with process investigation and appropriate action.

The statistical relationship provides evidence to investigate; the improvement action must still be connected to actual process understanding.

10. Black Belt Perspective

A Black Belt should ask:

  1. What are the two variables?
  2. What does the scatter diagram show?
  3. What is the strength and direction of the linear association?
  4. Is regression appropriate for the process question?
  5. Does the model adequately represent the data?
  6. What do the residuals indicate?
  7. Does the statistical relationship have a meaningful process interpretation?

The Black Belt should remember:

A relationship is evidence for investigation, not automatic proof of cause.

11. Lesson Practice

  1. Why should a scatter diagram precede correlation?
  2. Why does correlation not by itself prove causation?
  3. What can residual analysis tell a Black Belt?
  4. What is the difference between correlation and regression?
  5. Why is model adequacy important when using regression?
  6. Give one example of a process where correlation and regression could be useful.

12. Key Learning Points

  • A scatter diagram provides an initial visual check for a relationship.
  • Correlation describes the strength and direction of linear association.
  • The correlation coefficient is represented by r.
  • Regression models a response in relation to predictor variables.
  • Residuals are useful for assessing model adequacy.
  • Minitab and Excel can support correlation and regression analysis.
  • Correlation alone does not establish causation.
  • Statistical analysis should be connected to process understanding and measurable improvement.

13. Lesson Conclusion

Correlation and Regression Analysis provide the Black Belt with structured methods for investigating relationships between process variables.

The essential sequence is:

Visualize → Measure Relationship → Model → Check Residuals → Interpret → Improve

The objective is not simply to calculate r or produce a regression equation, but to use the analysis appropriately to understand the process and support evidence-based improvement.