9. CORRELATION Flashcards
1
Q
- What Measure of Association do we use when we are assessing the association between 2 numeric variables?
A
- this means that both our exposure and our outcome
are numeric - we do not use the Mean or the Mean Difference
- INSTEAD:
- we use a specific mathematical model
- this is an equation
2
Q
- What is the benefit of using a mathematical model?
A
- we can predict the change in the Outcome (Y)
for a standard change in the Exposure (X) - this is known as the Regression Coefficient
- we can also quantify the strength of the association
- this is known as the Correlation Coefficient
3
Q
- What determines the accuracy of the Mathematical Model?
A
- the relationship between the Correlation Coefficient
and the Regression Coefficient
4
Q
- What are the 3 steps that we need to follow in order to fully investigate the association between 2 numeric variables?
A
- Derive the Scatter Plot
- Perform a Correlation Analysis
- Perform a Linear Regression Analysis
5
Q
- What is a Scatterplot?
A
- it is a graph
- it portrays the relationship between any two numeric
variables - the entire sample of the participants is plotted in this
2- Dimensional Plot
EACH INDIVIDUAL:
- has a value for the exposure (Independent Variable)(X)
- it has a value for the outcome (Dependent Variable)(Y)
∴ each participant has an Xi, Yi
6
Q
- What can Scatter Plots provide?
A
- they can provide an overall, graphical impression of
the association between the 2 numeric variables of
interest
7
Q
- What can Scatterplots reveal?
A
- they can reveal trends
- such as a direct (positive association)
- or an inverse (negative) association
8
Q
- What is a Direct (Positive) Association?
A
AS THE EXPOSURE (X) INCREASES:
- the outcome (Y) increases
9
Q
- What is an Indirect (Negative) Association?
A
AS THE EXPOSURE (X) INCREASES:
- the outcome (Y) decreases
10
Q
- What kind of association do we see in this Scatterplot?
A
- Direct (Positive) Association
11
Q
- What kind of association do we see in this Scatterplot?
A
- Indirect (Negative) Association
12
Q
- Define Correlation.
A
- it is a term that is usually used interchangeably with
the term: “association” - it refers to the association between numeric variables
13
Q
- What is a Positive Correlation?
A
- this is when an increase in the Exposure (X Variable)
leads to an increase in the Outcome (Y Variable)
14
Q
- What is a Negative Correlation?
A
- this is when an increase in the Exposure (X Variable)
leads to a decrease in the Outcome (Y Variable)
15
Q
- Can you give one examples in which we do not have typical exposure and outcome variables?
A
- the Correlation between the LDL and HDL
Cholesterol
16
Q
- What is a Correlation Coefficient?
A
- this is a measure of association
- it describes the strength of the correlation between 2
numeric variables
17
Q
- What values can a Correlation Coefficient have?
A
- the Correlation Coefficient ranges between -1 to +1
- it cannot take any value outside of this range
18
Q
- What do the positive or the negative sign of the Correlation Coefficient indicate?
A
- they indicate the direction of the association
- this association can either be positive or negative
19
Q
- What indicates the strength of the association?
A
- the actual value of the Correlation Coefficient
20
Q
- What does it mean when we have a Correlation Coefficient (r) of +1?
A
- this is known as a Perfect Positive Correlation
- this means than an increase in X results in an increase
in Y
21
Q
- What does it mean when we have a Correlation Coefficient (r) of -1?
A
- this is known as a Perfect Negative Correlation
- this means than an increase in X results in a decrease
in Y
22
Q
- What does it mean when we have a Correlation Coefficient (r) of 0?
A
- this means that there is no correlation
23
Q
- What is the Correlation Coefficient for this graph?
A
- r = 0
- there is no correlation
- the X,Y points are randomly positioned all over this plot
- they have no clear increasing or decreasing trend
24
Q
- What is the Correlation Coefficient for this graph?
A
- r = -1
- there is a Perfect Negative Correlation
- the X,Y points are perfectly positioned one behind the
other along a diagonal line - they have a decreasing trend
25
Q
- What is the Correlation Coefficient for this graph?
A
- r = +1
- there is a Perfect Positive Correlation
- the X,Y points are perfectly positioned one behind the
other along a diagonal line - they have an increasing trend
26
Q
- What is the Correlation Coefficient for this graph?
A
- 0 < r < 1
- there is a Positive Correlation
- the X,Y points are positioned one behind the other
along a diagonal line - they are not positioned perfectly
- they have an increasing trend
27
Q
- What is the Correlation Coefficient for this graph?
A
- -1 < r < 0
- there is a Negative Correlation
- the X,Y points are positioned one behind the other
along a diagonal line - they are not positioned perfectly
- they have a decreasing trend
28
Q
- What indicates the strength of the correlation?
A
- how close the correlation coefficient is to +1 or -1
29
Q
- Fill in the titles for the following coefficient values:
29.1. r = -1
29.2. r ≥ -0.7
29.3. r = -0.7 to -0.5
29.4. r = - 0.5 to -0.3
29.5. r < -0.3
29.6. r = 0
29.7 r < +0.3
29.8. r = +0.3 to +0.5
29.9. r = +0.5 to +0.7
29.10. r ≥ +0.7
29.11. r = +1
A
29.1. Perfect Negative Correlation
29.2. Strong Negative Correlation
29.3. Moderate Negative Correlation
29.4. Weak Negative Correlation
29.5. Very Weak Negative Correlation
29.6. No Correlation
29.7. Very Weak Positive Correlation
29.8. Weak Positive Correlation
29.9. Moderate Positive Correlation
29.10. Strong Positive Correlation
29.11. Perfect Positive Correlation
30
Q
- What are the two main types of Correlation?
A
- Pearson’s Correlation
- Spearman’s Correlation
31
Q
- What is Pearson’s Correlation?
A
- this is a type of correlation
- it is most commonly used type (99.9% of the time)
- it denotes the correlation between 2 variables
- it does this using the original values of these variables
- it is used when data is continuous
32
Q
- What is Spearman’s Correlation?
A
- this is a type of correlation
- it is also known as Spearman’s Rank Correlation
- it denotes the correlation between 2 variables
- it does this by first ranking the values
- it ranks them usually from lowest to highest
- it then assesses the correlation between the 2 ranks
33
Q
- In which 3 situations do we make use of Pearson’s Correlation?
A
- Continuous Data
- Normally Distributed Data
- Linear Relationship
34
Q
- In which situations do we make use of Spearman’s Correlation?
A
- when we do not have Continuous Data
- when we do not have Normally Distributed Data
- when we do not have a Linear Relationship
- when the data is Ordinal
