W6 Correlation Analysis

Question 1

How can we use correlation analysis and what does it measure?

Answer 1

• We can use it to examine the relationship between any two things or any two variables to see the extent to which they are related.
• It measures strength and direction of data

Question 2

• What is negative correlation?
• What is positive correlation?
• What is little systematic tendency?

Answer 2

• Positive correlation = bottom left to top right of the graph
• Negative correlation = top left to bottom right of the graph
• This is 0 or no relationship between the data and is a horizontal line.

Question 3

If there are lots of data points on a graph clumped together around the line of best fit what does that tell you about the strength of the data?

Answer 3

• The more data that is clumped together around the line of best fit the stronger the data is

Question 4

What is the definition of correlation coefficient?

Answer 4

• A numerical value that indicates the extent to which two variables are related.
• Thus, it is a numerical summary of a bivariate (two variables) relationship

Question 5

Give the 7 steps to correlation coefficient:

Answer 5

1 & 2: Consider the Null (H0) & Alternative (H1) Hypothesis for any relationship.

3: Level of significance (P-value/Pearsons coefficient)
4: Collect & Summarise data
5: Assumptions to be carried out before attempting correlation coefficient
6: Run statistical test
7: Interpret significance of results

Question 6

What are the 5 assumptions that need to be tested before attempting a correlation coefficient?
(if our data meets any of these assumptions then it allows us to do a correlation coefficient)

Answer 6

To make sure the data is parametric:

1. Normal distribution
2. Homogeneity of variance
3. Interval/ratio
4. Independence
5. Linear relationship

Question 7

What is the parametric test that we will be using?

Answer 7

• The parametric test we will be using is called a Pearson’s Correlation test

Question 8

Assumptions:

What is normal distribution?

Answer 8

• We are looking to see if our data is normal.

- Does it fit between 1.0 & -1.0 which relates to skewness and kurtosis

Question 9

Assumptions:

Can you describe what Independence is?

Answer 9

• This is checking participants aren’t influencing other participants’ data

Question 10

Assumptions:

What is meant by Interval/ratio?

Answer 10

• You have to make sure that the data is either interval or ratio data
  Interval = Equal units or intervals between data points on a scale but there is no absolute zero point E.g. Body temperature
  Ratio = Equal units of measurements and has an absolute zero point E.g. Speed

Question 11

Assumptions:

Can you describe what homogeneity of variance is?

Answer 11

• This tests the variance of one variables stability compared to to the other at all levels
• The way to see if data has good homogeneity of variance is to ask “Does the data follow/stay close to the line of best fit?”
• Homogeneity = The same
• Heterogeneity = Different

Question 12

Assumptions:

What is Linear relationship?

Answer 12

• You can check this by plotting all the data on a scatter plot and try to see any obvious curves in the relationship or straight lines

Question 13

Interpreting SPSS:

• If you are looking for the number of participants what is that represented as in SPSS?
• If you are looking for the Correlation Coefficient how will that be represented in SPSS?
• Significant Value/ P value is represented as what in SPSS?

Answer 13

• ‘N’ represents the number of participants
• Correlation coefficient is represented as ‘Pearson’s Correlation’ or ‘r’ in SPSS
• Significant value/P value is represented as ‘Sig.’ in SPSS

Question 14

Interpreting SPSS:
- SPSS provides you with the following data. Write a small paragraph to describe the data you are provided with.
'Pearson Correlation = .659'
'Sig. (2-tailed) = .000'
'N = 25'

Answer 14

Looking at the following data I would interpret a highly significant positive linear relationship between both variables (r=.66 p<0.001)