Scribd
Upload
28 views

PMC 500 Statistical Reasoning in Education: Correlation

The document discusses correlation and how to calculate and interpret the Pearson correlation coefficient (r). It defines correlation as the degree of linear relationship between two variables and provides examples of positive, negative, and zero correlations. It explains how to create scatterplots to visualize correlations and how the Pearson r quantifies both the direction and magnitude of linear relationships between interval/ratio variables. Values of r range from -1 to 1, with higher absolute values indicating stronger linear relationships. The document also covers testing the significance of r using hypothesis testing in SPSS.

Uploaded by

Awi Aziz
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
28 views

PMC 500 Statistical Reasoning in Education: Correlation

The document discusses correlation and how to calculate and interpret the Pearson correlation coefficient (r). It defines correlation as the degree of linear relationship between two variables and provides examples of positive, negative, and zero correlations. It explains how to create scatterplots to visualize correlations and how the Pearson r quantifies both the direction and magnitude of linear relationships between interval/ratio variables. Values of r range from -1 to 1, with higher absolute values indicating stronger linear relationships. The document also covers testing the significance of r using hypothesis testing in SPSS.

Uploaded by

Awi Aziz
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 45

PMC 500

STATISTICAL REASONING IN EDUCATION

CORRELATION

Dr. Nor Asniza Ishak


Dr. Ahmad Zamri Khairani
Correlation
• Thorne and Giesen (2003):
• Correlation:
- the degree of linear relationship between two
variables
• Example:
Mathematics score and Physics score
What patterns do you find in these data?
Student Mathematics Score Physics Score

1 96 90
2 94 87
3 90 84
4 89 81
5 88 80
6 72 70
7 68 65
8 65 60
9 60 58
10 57 53
11 50 48
12 45 40
13 40 38
14 34 30
15 30 25
• What patterns do you find in these data?
• In general,
- students with high Mathematics scores also have high
Physics scores
- students with low Mathematics scores also have low
Physics scores
• These patterns show that there is a correlation between
Mathematics score and Physics score
Scatterplot
• The correlation between Mathematics score and Physics score can be
represented by a scatterplot
• A scatterplot is a visual representation of the relationship between
two variables
• Each point in a scatterplot represents the value for the X and Y
variables for a subject
• Example: A scatterplot of the relationship between Mathematics
score (X) and Physics score (Y)
Ø Creating a scatterplot using SPSS:

- Create a data file


- Click Graphs > Chart Builder... > Scatter/Dot ...
- Drag Simple Scatter to Chart preview box
- Drag the Y variable to the Y-Axis? box
- Drag the X variable to the X-Axis? box
- Click OK
Types of Correlation (Giesen, 2003)

1. Positive correlation
2. Negative correlation
3. Zero correlation
1. Positive Correlation
• In other words, a positive correlation between two variables, X and Y,
implies a direct linear relationship between the variables
- X increases in value, Y increases in value
- X decreases in value, Y decreases in value
• Example:
- Time spent studying for a History test and the score earned on
the test
The points in the scatterplot form
a close approximation to a
straight line slanting upward to
the right

Time spent studying History test (hour)


Time spent studying History test (hour)
2. Negative Correlation
• A negative correlation between two variables is one in which:
- high scores on one variable are associated with low scores on
the other variable
- low scores on one variable are associated with high scores on
the other variable
• In other words, a negative correlation between two variables, X and Y,
implies an inverse linear relationship between the variables
- X increases in value, Y decreases in value
- X decreases in value, Y increases in value
• Example:
- Living off campus and involvement in campus activities
What pattern do you see in this set of
points?
n The points in the scatterplot form a
close approximation to a straight line
slanting upward to the left

Living off campus, X (%)


3. Zero Correlation
• A zero correlation between two variables is one in which:
- a high score on one variable is just as likely to be associated
with a low score as it is with a high score on the other variable
- a low score on one variable is just as likely to be associated with
a high score as it is with a low score on the other variable
• In other words, a zero correlation between two variables, X and Y,
implies no linear relationship between the variables
• Example:
- Order in which students turn in their test papers and their test
scores
What pattern do you see in these points?
n The points on the scatterplot form a
random pattern
n The random pattern of points indicates a
zero correlation

Order in which students turn in their test papers


Order in which students turn in their test papers
The Pearson Correlation Coefficient
• A scatterplot shows the direction of the linear relationship
between two variables
n positive, negative or zero correlation
• But, it does not provide an exact measure of the magnitude of
the linear relationship between two variables
n High, moderate or low correlation
• Thus, the English statistician Karl Pearson (1857-1936)
developed a statistic called the Pearson correlation coefficient or
the Pearson r
• The Pearson r is a statistic that indicates the direction and magnitude
of the linear relationship between two interval or ratio scaled
variables
• Examples of interval or ratio scaled variables:
- test score, height, weight, age, income, time spent studying for
an exam, etc.
• The Pearson correlation coefficient is represented by the small letter r
• The values of r may range from –1.00 to +1.00
• The ‘ + ’ and ‘ – ’ signs represent direct and inverse linear
relationships respectively
• The absolute values of a correlation coefficient represent the
magnitude of the linear relationship
• E.g. r = 0.95 and r = - 0.95 have the same magnitude but different
directions
Computing r using SPSS
Create a data file
• Analyze >Correlate > Bivariate ...
• Double-click on the X variable to move it to the Variables box
• Double-click on the Y variable to move it to the Variables box
• Click OK
Coefficient of Determination
• The square of the correlation coefficient is called the
coefficient of determination
• Coefficient of determination
• The proportion of the variance in a variable that can be
associated with the variance in the other variable
• Is represented by r 2
Coefficient of Determination

Example:
n r = 0.692
n r 2 = 0.6922 = 0.479

n 47.9% of the variance in Y

can be associated with the


variance in X
Testing r for Significance: Assumptions
• The Pearson correlation coefficient, r is used to determine whether
there is a significant correlation between two interval or ratio scaled
variables
• Assumptions of the Pearson correlation coefficient, r :
• The sample is chosen randomly from the population and the scores
for each of the participants should be independent of all other
participants’ scores.
• The scores for each variable are normally distributed in the
population
Testing r for Significance: 5 main steps in
hypothesis testing using SPSS
1. State the null and alternative hypotheses
2. Set the significance level for rejecting the null hypothesis
3. Analyze the data using SPSS
4. Make a decision about the null hypothesis
5. Make a conclusion
Example 1
• A researcher want to find if there is a correlation between
Mathematics scores and Physics scores of 30 Form Four
students.
• Test the hypothesis of whether the correlation between
Mathematics scores and Physics scores is significant at the
0.01 level of significance.
Step 1: State the null and alternative
hypotheses

• Ho : There is no significant correlation between Mathematics


scores and Physics scores in the population.

• H1 : There is a significant correlation between Mathematics


scores and Physics scores in the population.
Step 2: Set the significance level for rejecting
the null hypothesis
Step 3: Analyze the data using SPSS
Step 3: Analyze the data using SPSS
• Create a data file
• Click Analyze > Correlate > Bivariate ...
• Double-click on the variable named Mathematics scores to move it to
the Variables box
• Double-click on the variable named Physics scores to move it to the
Variables box
• Click OK
Step 4: Make a decision about the null
hypothesis

Reject H because p < 0.01


Step 5: Make a conclusion
Example 2
• A researcher want to find if there is a correlation between
Malay Language scores and English Language scores of 30
students
• Test the hypothesis of whether the correlation between
Mathematics scores and Physics scores is significant at the
0.05 level of significance.
Step 1: State the null and alternative
hypotheses

• Ho : There is no significant correlation between Malay Language


scores and English Language scores in the population.

• H1 : There is a significant correlation between Malay Language scores


and English Language scores in the population.
Step 2: Set the significance level for rejecting
the null hypothesis
Step 3: Analyze the data using SPSS
Step 3: Analyze the data using SPSS
• Create a data file
• Click Analyze > Correlate > Bivariate ...
• Double-click on the variable named Mathematics scores to move it to
the Variables box
• Double-click on the variable named Physics scores to move it to the
Variables box
• Click OK
Step 4: Make a decision about the null
hypothesis

Fail to reject Ho because p > 0.05


Step 5: Make a conclusion
Thank You
Enjoy Statistics!!!

You might also like