Scribd
Upload
73 views9 pages

Measuring Relationship Via Regression Analysis and Correlation

The document explains the differences between regression and correlation, highlighting their purposes, outputs, and assumptions. It details regression analysis, including simple and multiple regression, and introduces the Ordinary Least Squares (OLS) method for estimating regression lines. Additionally, it describes the correlation coefficient and its interpretation in measuring the strength and direction of relationships between two variables.

Uploaded by

hadiafiaz932
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
73 views9 pages

Measuring Relationship Via Regression Analysis and Correlation

The document explains the differences between regression and correlation, highlighting their purposes, outputs, and assumptions. It details regression analysis, including simple and multiple regression, and introduces the Ordinary Least Squares (OLS) method for estimating regression lines. Additionally, it describes the correlation coefficient and its interpretation in measuring the strength and direction of relationships between two variables.

Uploaded by

hadiafiaz932
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

Tools for Quantitative Reasoning

Ms. Usma Noor


Topic # 4
MEASURING RELATIONSHIP VIA REGRESSION ANALYSIS AND CORRELATION

Difference between Regression and Correlation


Here’s a concise comparison between regression and correlation in points:
Purpose:
• Regression: Models the dependent variable and one or more independent variables (dependence of one
variable upon other). It is used for prediction or explanation.
• Correlation: Measures the strength and direction of the linear relationship between two variables.
Output:
• Regression: Provides an equation (e.g., Y=a+bX) to predict the dependent variable.
• Correlation: Provides a correlation coefficient (r) to measure the strength and direction of the
relationship.
Number of Variables:
• Regression: Typically involves one dependent variable and one or more independent variables.
• Correlation: Measures the relationship between two variables at a time.
Assumptions:
• Regression: Assumes a dependence and other conditions (normality of errors, independence,
homoscedasticity).
• Correlation: Assumes a linear relationship but does not require other conditions (like residual
independence).
Formula:
• Regression: Typically expressed as a mathematical equation (e.g., Y=a+bX).
• Correlation: Expressed as a single number (correlation coefficient r) ranging from −1 to +1.
Interpretation:
• Regression: Provides specific predictions (e.g., predicting Y for a given X) and model coefficients.
• Correlation: Provides an index of association (e.g., r=0.85 means a strong positive relationship).
What does Regression/ simple regression/ multiple regression mean?

Regression Simple regression Multiple regression.


The dependence of one variable When we study the dependence of a When the dependence of a
(Dependent variable) on one or variable on a single independent variable on two or more than
more other variables (Independent variable then it is called simple two variables are studied then it
Variable) is called regression. regression or two variable regression. is called multiple regression.

Difference between Regression, Regressor and Regressand?

Regression Regressor Regressand


The dependence of one variable The regression process the The regression process the
(Dependent variable) on one or independent variable is called dependent variable is called
more other variables (Independent regressor. regressand.
Variable) is called regression. X variable Y variable

Model Expressions:
If the dependent and independent variates is believed simple linear then the model expressing this relationship
in the form:
̂ = 𝒃𝟎 + 𝒃𝟏 𝒙 𝟏 + ⅇ
𝒚
Estimated Regression line:
̂ = 𝐚 + 𝒃𝟏 𝒙𝟏
𝒚
If the dependent and independent variates is believed multiple linear then the model expressing this relationship
in the form:
̂ = 𝒃𝟎 + 𝒃𝟏 𝒙𝟏 + 𝒃𝟐 𝒙𝟐 + 𝒃 𝟑 𝒙𝟑 + ⅇ
𝒚
Estimated Regression line:
̂ = 𝐚 + 𝒃𝟏 𝒙𝟏 + 𝒃𝟐 𝒙𝟐 + 𝒃𝟑 𝒙𝟑
𝒚
Where:
• ̂ → Dependent Variable, Predictant, Explained, Responsive/Responser, Unknown Variable,
𝒚
Regressand.
• 𝒂, 𝒃𝟎 → Constant Intersept, Constant term, Parameter, value of 𝒚
̂ when x=0
• 𝒃𝟏 , 𝒃𝟐 , 𝒃𝟑 → Slop coefficients for each variable of x, Regression Coefficients, Rate of change ( how much
y changes for each unit of x )
• 𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 → Independent, Predictors, Explanatory, Known variables, Regressor
• ⅇ → Error term, Unknown (actual value or observed values that are Y – estimated value or predicted values
that are 𝒚
̂)
Ordinary Least Square Method/ OLS Method for estimated regression line 𝒚
̂ = 𝐚 + 𝒃𝟏 𝒙𝟏 :
The OLS method finds the values of a and b that minimize the sum of squared residuals (errors).

Correlation:
The correlation coefficient (also called Pearson's correlation coefficient, denoted as r) measures the strength and
direction of the linear relationship between two variables x and y. (height and weight of child). It ranges from −1
to 1, where:

• R=1 then perfect positive correlation


• R=0.8 then highly positive correlation
• R=0.3 then low positive correlation
• R=0 the no correlation
• R= -0.3 then low negative correlation
• R= -0.8 then highly negative correlation
• R= -1 then perfect negative correlation
CamScanner
CamScanner
CamScanner
CamScanner
CamScanner
CamScanner

You might also like