0% found this document useful (0 votes)
53 views17 pages

CIVE 5015: Research Data Analysis: Lecture 9: Correlation and Regression Analysis

This document provides an overview of correlation and regression analysis. It defines correlation as determining the size and direction of association between two variables, but not causation. Scatter plots can show positive, negative, or no correlation. Correlation coefficients like Pearson's r and Spearman's rho measure correlation from -1 to 1. Regression finds the relationship between an independent and dependent variable and allows for prediction. The document demonstrates performing correlation and regression tests in SPSS and interpreting the outputs.
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
0% found this document useful (0 votes)
53 views17 pages

CIVE 5015: Research Data Analysis: Lecture 9: Correlation and Regression Analysis

This document provides an overview of correlation and regression analysis. It defines correlation as determining the size and direction of association between two variables, but not causation. Scatter plots can show positive, negative, or no correlation. Correlation coefficients like Pearson's r and Spearman's rho measure correlation from -1 to 1. Regression finds the relationship between an independent and dependent variable and allows for prediction. The document demonstrates performing correlation and regression tests in SPSS and interpreting the outputs.
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/ 17

CIVE 5015: Research Data Analysis

Lecture 9: Correlation and Regression Analysis

Dr Faisal Ahammed
Lecturer of Civil Engineering
Program Director – Civil Engineering Postgraduate Programs
University of South Australia
Correlation
• The purpose of correlation is to determine the size and direction
of association between two variables.

• Correlation doesn’t determine the causation.

• If two variables, x and y, are correlated; we can’t say x causes y.


It can be y causes x or other known or hidden variables may
produce the variation between x and y.
Scatter Plot
A scatter plot represents each case on a graph.
Positive Correlation
Negative Correlation
Drawing Scatter Plot
Graphs → Legacy dialogs → Scatter / dot → Simple scatter → X-
axis variable and Y-axis variable → OK→ Double click the scatter
plot → Elements → Fit line at total
Types of Correlation Coefficient
Two common types:
- Spearman’s rank
- Pearson’s product moment

• Spearman’s rank and Pearson’s r correlations are based on linear


relationships between variables.

• Non-linear relationships may be analysed using “eta”.

• Both Spearman’s rank and Pearson’s r are measured between -1


and +1.

• A correlation closer to zero suggests no (or weak) correlation,


whereas a value closer to 1 (plus or minus) suggests a strong
correlation.
Conditions of Correlation Coefficients
• Pearson’s r is a test of correlation between two parametric, scale
level variables.

• Spearman’s rho is a test of correlation between two scale level


variables, when at least one of them is not parametric.

• Both tests are rely on linear association between variables (it can
be checked by drawing scatter plot).
Performing Correlation Tests in SPSS
Analyse → Correlate → Bivariate → Data (variables) input → Pearson,
Spearman, Two tailed, Flag significant correlations → OK
Interpretation of SPSS Outputs
• Using Spearman’s rho, the correlation between WC and GSP is
0.91, which is strong positive correlation and the test is
statistically significant, as p value is 0.000 ( <0.01).

• The correlation between Ratio (= GSP/WC) and GSP is -0.166,


which is weak negative correlation, but the test is NOT
statistically significant, as p value is 0.189 (>0.01).
Regression Analysis
Regression is a powerful tool for:

- Summarising the nature of relationship between two variables

- Prediction of dependent variables.


Regression Equation
Y= 𝑎 + 𝑏𝑋 𝑜𝑟 𝑌 = 𝑎 − 𝑏𝑋

X = independent variable
Y = dependent variable
b = slope of line of best fit
a = the point at which line crosses the Y-axis.
Regression Analysis Using SPSS
Analyse → Regression → Linear → Dependent and independent
variable → OK

Independent variable
= WC
Dependent variable
= GSP
Regression Test Results

• The level of significance is less than 5% (0.000); the test is


statistically significant.
• The Pearson's r value is 0.939; it shows strong positive associations
between WC and GSP.
• The regression equation is GSP = 40078.90 + (74.57×WC)
• If WC is 50 GL, GSP = 40078.90 + (74.57×50) = $43807.4 m
• IF WC is 7500 GL, GSP =40078.90 + (74.57×7500) = $599353.9 m
Multiple Regression
A Case of Multiple Regression

Please see the note “a case of multiple regression”.


Thank You

You might also like