Correlation Analysis

Correlation Analysis

◼ One of the most basic measures of the

association among variables, and a
foundational statistic for several more
complex statistics, is the correlation
coefficient.
Correlation Analysis

◼ there are a number of different types of

correlation coefficients, the most commonly
used in social science research is the Pearson
product-moment correlation coefficient.

◼ The other types of correlation are the

point-biserial coefficient, the Spearman rho
coefficient, and the phi coefficient.
When to Use Correlation and What
It Tells Us
◼ Researchers compute correlation coefficients
when they want to know how two variables
are related to each other

▪ For a Pearson product-moment correlation, both of the

variables must be measured on an interval or ratio scale
and are known as continuous variables
Characteristics of correlation

◼ There are two fundamental characteristics of

correlation coefficients researchers care about.
◼ The first of these is the direction of the
correlation coefficient
▪ A positive correlation indicates that the values on the
two variables being analyzed move in the same direction.
▪ That is, as scores on one variable go up, scores on the
other variable go up as well (on average).
Characteristics of correlation

▪ A negative correlation indicates that the values on

the two variables being analyzed move in opposite
directions.
▪ That is, as scores on one variable go up, scores on
the other variable go down, and vice versa (on
average).
Characteristics of correlation

◼ The second fundamental characteristic of

correlation coefficients is the strength or
magnitude of the relationship.
◼ Correlation coefficients range in strength from
–1.00 to +1.00.
▪ The closer the correlation coefficient is to either
–1.00 or +1.00, the stronger the relationship is
between the two variables.
▪ r is the symbol for the sample Pearson correlation
coefficient
◼ Generally, correlation coefficients stay
between –.70 and +.70.
◼ –.20 and +.20 indicate a weak relation
between two variables
◼ .20 - .50 (either positive or negative) represent
a moderate relationship
◼ larger than .50 (either positive or negative)
represent a strong relationship
Pearson Correlation Coefficients in
Depth
◼ The first step is to notice that we are
concerned with a sample’s scores on two
variables at the same time
◼ it is critical that the scores on the two variables
are paired (score on IV is paired with DV)
Calculating the Correlation
Coefficient
◼ The first step is to standardize your variables
◼ What this does is provide a z score for each
case in the sample.
◼ then we have to find the sum of the cross
products between the z scores on each of the
two variables being examined for each case in
the sample.
◼ When we multiply each individual’s score on
one variable with that individual’s score on the
second variable (i.e., find a cross product),
◼ sum those across all of the individuals in the
sample, and then divide by N
◼ we have an average cross product, and this is
known as covariance.
◼ If we standardize this covariance, we end up
with a correlation coefficient
What the Correlation Coefficient
Does, and Does Not, Tell Us
◼ Correlation Coefficient Tells Us
▪ Strength of the Relationship
▪ Values range between -1 and +1.
▪ +1: Perfect positive relationship (as one variable
increases, the other also increases).
▪ -1: Perfect negative relationship (as one variable
increases, the other decreases).
▪ 0: No linear relationship.
◼ Direction of the Relationship
▪ Positive values indicate a direct (positive)
relationship.
▪ Negative values indicate an inverse (negative)
relationship.
◼ Nature of Linear Association
▪ The correlation coefficient specifically measures
how well data points fit a linear pattern.
Correlation Coefficient Does Not
Tell Us
◼ Causation
▪ Correlation does not imply causation. A strong
correlation does not mean one variable causes the
other to change.
Example
▪ There is a positive correlation between shark attacks and
the number of people at the beach.
▪ This doesn’t mean going to the beach causes shark
attacks. The real reason is more people go to the beach in
summer, which increases the likelihood of shark attacks.
◼ Non-Linear Relationships
◼ It fails to capture relationships that are
non-linear
▪ A child’s happiness might increase with candy
consumption up to a point, but after eating too
much candy, happiness might decrease due to a
stomachache.
◼ Correlation Can Be Misleading Because of
Outliers
◼ A few unusual data points can distort the
correlation value.
Example:
▪ Imagine measuring the relationship between study
hours and exam scores for 10 students.
▪ If 9 students show a strong positive trend, but 1 student
studied a lot and failed because of illness, that one
“outlier” could lower the correlation, making it seem
weaker than it really is.
◼ Correlation Does Not Show How Much One
Variable Affects the Other
◼ It only shows the strength of the relationship,
not the size of the effect.
Example
▪ If height and weight have a high positive
correlation, it doesn’t mean all tall people are
heavy or that height determines exact weight. It
just shows a general trend.
The Coefficient of Determination

◼ The coefficient of determination, denoted as

R2 is a statistical measure that explains the
proportion of the variance (variability) in one
variable that is predictable from the variance
in another variable.

◼ It is derived from the square of the correlation

coefficient r
R2=r2
The Coefficient of Determination
Explained Variance

◼ Explained variance refers to the portion of

the total variability in the dependent variable
(outcome) that can be attributed to its
relationship with the independent variable(s).
◼ Suppose we are studying the relationship
between study hours and exam scores, and
R2=0.7
▪ This means 70% of the variability in exam scores is
explained by study hours.
▪ The remaining 30% (unexplained variance) could
be due to other factors like prior knowledge, sleep
quality, or test anxiety.
Shared variance

◼ When two variables are related, or correlated,

with each other, there is a certain amount of
shared variance between them.
▪ If height and weight are correlated with R2=0.49 it
means 49% of the variability in weight is shared
with height (explained by height). The remaining
51% is influenced by other factors like diet or
genetics.
Other Types of Correlation
Coefficients
◼ Point Biserial Correlation
◼ The Point Biserial correlation is used to
measure the relationship between:
▪ One continuous variable (e.g., test scores).
▪ One binary variable (e.g., gender: male/female or
pass/fail).
▪ Why We Use Point Biserial Correlation:
▪ It helps us understand whether the two groups
(from the binary variable) differ significantly on
the continuous variable.
◼ Spearman's Rho Correlation (ρ)
◼ Spearman’s Rho measures the relationship
between two variables based on their ranked
values (not actual scores).
▪ Used for ordinal data (rankings) or when the data
is not normally distributed.
▪ It measures monotonic relationships (consistent
upward or downward trends).
SPSS activity

◼ Analyze < correlate <bivariate

◼
One-Tailed vs. Two-Tailed Tests in
Hypothesis Testing
◼ When conducting a hypothesis test, choosing
one-tailed or two-tailed depends on the
research question and how you frame the
alternative hypothesis
◼ Two-Tailed Test
▪ A two-tailed test checks for the possibility of a
relationship in both directions.
◼ Question: Does caffeine consumption affect
reaction times?
▪ Null: Caffeine has no effect (μ=0).
▪ Alternative: Caffeine either increases or decreases
reaction time (μ≠0).
▪ A two-tailed test is used because we are
considering effects in both directions.
◼ One-Tailed Test
▪ A one-tailed test checks for a relationship in one
specific direction only.
▪ It is used when we are only interested in whether a
parameter is greater than or less than a specific
value.
◼ Does a new drug lower blood pressure?
▪ Null: The drug has no effect or increases blood
pressure (μ≥0).
▪ Alternative: The drug lowers blood pressure (μ<0).
◼ A one-tailed test is used because we are only
interested in a decrease.