Correlational Research

CORRELATIONAL RESEARCH
Professor: Dr. Amir Karimnia

Prepared by: Fatemeh Yousefi
Correlational Research
Definition:
 Correlational research design investigates statistical
relationship between 2 variables (or more) without
controlling or manipulating by the researcher.
Objective:
 The aim of this research is to explore whether a
relationship between 2 variables exists.

Correlational Research
Examples:
 Is there any relationship between watching violent
movies and aggressive behavior?
 Is smoking associated with Alzheimer’s disease?
 Research that focuses on a particular language skill
(reading, listening, speaking and writing also use

correlational analysis.
Types of Correlational Research
 Negative correlation (variables are in opposite
directions)
Example: as coffee consumption increases, tiredness
decreases
 Zero correlation (no relationship between the variables)
Example: coffee consumption is not correlated with

height
 Positive correlation (variables are the same direction)
Example: as height increases, weight increases

Characteristics of Correlational Research
 Non-experimental (the investigator just analyzes

and examines the relationship among variables, not
changing or modifying them in any way)
 Dynamic (it is not static, another research might
disapprove it)
Correlational vs. Experimental Research
 Both are quantitative methods and investigate

relationship between variables.
 Important differences in how data is collected & the
types of conclusion you can draw.
Correlational vs. Experimental Research
 Purpose: Ex. R→ cause & effect relationship between

variables
C R→ relationship between variables
 Variables: Ex. R→ manipulated
C R→ only observed with no manipulation
 Validity: Ex. R→ high internal validity (certain
conclusion about causation)
C R→ high external validity (confidently
generalize your conclusions to other
populations or settings)
When to use Correlational Research
 To investigate non-causal relationships

Example: is any correlation between the number of
children people have and which political party they
vote for?
(A strong correlation could be useful for making
predictions about voting patterns).
 To support the theory of experimental research

When it is impractical, time consuming or too costly
Example: whether greenhouse gas emission cause
global warming: impossible to do an experiment but
through observation and analysis you can show a
strong correlation that supports the theory.
 To test new measurement tools

When you need to test its reliability or validity, collect
data, using the new scale, test the degrees of
correlations between different measurements. Finding
high correlation means that your scale is valid.
How to collect data for Correlational Research
 Surveys (questionnaires can be used to measure the

variables, it can also be online, by phone, by mail or
in person), e.g. correlation between income and
vegetarianism
 Observation (gathering data in the natural

environment, about phenomenon or behavior by
recording, counting, describing and categorizing
actions and events), e.g. relationship between
gender and class participation
 Secondary data (using previous data that may be

collected for other purposes), merits: inexpensive
and fast, because data collection is complete
demerits: data may be unreliable, incomplete or not
entirely relevant, and you have no control over the
reliability or validity of the data collection
procedures
How to analyze Correlational data
 After collecting data, you can statistically analyze
the relationship between variables using regression
analyses.
 Correlational coefficient ‘r’ → -1 to +1
Pearson Correlation vs. Spearman Correlation
Pearson correlation Spearman correlation
 Parametric  Non-parametric
 Monotonic
 Linear Spearman correlation is
suitable for:
 Ordinal data (e.g.
measuring economic status
using hierarchy: ‘wealthy’ ,
‘middle income’ or ‘poor’
 Interval data (e.g.
temperature in degrees
Fahrenheit, 20-40
degrees/40-60 degrees)
Main types of Correlation
Used in applied linguistics research
1) Pearson product-meant correlation
Standard type of correlation → ‘r’
Measuring a linear correlation→ -1 to +1
Pearson product-moment correlation
r=-1
Perfect negative correlation
r<-. /5
Strong negative correlation
0>r>-. /3
Weak negative correlation
r=0
No-correlation
When to use the Pearson correlation coefficient
 Both variables are quantitative

 Both variables are normally distributed
 The data have no outliers (using scatterplot)
 The relationship is linear (using scatterplot)
2) Point-biserial correlation and phi coefficients
When one or both variables are dichotomous (2
values/binary) e.g. ‘gender’
Variables are binary or continuous, in binary variables
we have 2 possible values such as sex (male/female),
smoker (yes/no), but continuous variables can take on
any reasonable value such as age, weight, height, test
scores)
3) Partial correlation
It is a very useful technique. It is relationship between 2
variables, while controlling for the effect of one or more
other variables.
Example: amount of food eaten & blood pressure while
controlling for weight or amount of exercise
 Continuous independent variable: ‘amount of eaten’
 Continuous dependent variable: ‘blood pressure’
 Continuous control variables: ‘weight’ & ‘amount of

exercise’
4) Multiple correlation
‘R’→ the correlation between one variable and a
group of variables e.g. composite effects of a range
of learner characteristics (motivation, creativity,
anxiety) on achievement: more than one or two
variable is usually not recommended because:
the more control variables, the less reliable your test.

Correlational Research

Uploaded by

Copyright:

Available Formats

Correlational Research

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Correlational Research

Uploaded by

Copyright:

Available Formats

CORRELATIONAL RESEARCH

Professor: Dr. Amir Karimnia

relationship between 2 variables exists.

 Research that focuses on a particular language skill

(reading, listening, speaking and writing also use

Example: coffee consumption is not correlated with

Example: as height increases, weight increases

 Non-experimental (the investigator just analyzes

 Both are quantitative methods and investigate

 Purpose: Ex. R→ cause & effect relationship between

 To investigate non-causal relationships

 To support the theory of experimental research

 To test new measurement tools

 Surveys (questionnaires can be used to measure the

 Observation (gathering data in the natural

 Secondary data (using previous data that may be

Pearson correlation Spearman correlation

 Both variables are quantitative

 Continuous dependent variable: ‘blood pressure’

 Continuous control variables: ‘weight’ & ‘amount of

You might also like