Correlational Research

DESCRIPTION AND CHARACTERISTICS Yes

Correlational research involves a wide spectrum of non-experimental studies, which begin

with the observation of behaviors produced naturally by the action of variables over which no
manipulation or experimental control has been exercised. These are established behaviors
of existing connections between variables, developing behaviors and behaviors affected by
environmental and biological variables. The correlational method allows collecting evidence
of facts, concentrating on taking measurements and investigating alleged relationships
between them.

This type of descriptive study aims to determine the degree of non-causal relationship or
association between two or more variables. They are characterized because first the
variables are measured and then, through correlational hypothesis testing and the
application of statistical techniques, the correlation is estimated. Although correlation
research does not directly establish causal relationships, it can provide clues about the
possible causes of a phenomenon. This type of descriptive research seeks to determine the
degree of relationship between the variables, the researcher relates the level of one variable
with the corresponding level of another variable, in order to discover relationships between
them. The purpose is to predict the level of a variable based on knowledge; The researcher
has moderate control over the variables.

The existence and strength of this covariation is usually determined statistically by means of
correlation coefficients. It is convenient to keep in mind that this covariation does not mean
that causal relationships exist between the factors, since these are determined by other
criteria that, in addition to covariation, must be taken into account. Examples of correlational
research are the following:

CASES IN WHICH CORRELATIONAL RESEARCH IS USED.

Correlational studies are carried out when the treatment variables cannot be manipulated for
various reasons, of which we will point out three:

YO/. It is impossible to physically manipulate the variables. Let's imagine that a psychologist
wants to study the relationship between two response measures such as intelligence and
academic performance. Intelligence is an individual characteristic, a trait that is defined
based on performance on a standardized test and cannot be physically manipulated.
However, the relationship can be investigated in a correlational study by selecting a group of
students from a school, measuring their IQ and comparing these scores with their
performance at school of these students.

II/. The events have already occurred, but years that a group of university students spent in
the preschool stage and the scores they subsequently obtained in the first stage of EGB

III/. When the manipulation of variables is illegal or unethical, e.g. relationship between
heroin consumption and the number of heart attacks. It is not ethical to administer different
doses of heroin to a group of people and see whether or not a heart attack occurs under
conditions that are as controlled as possible, but we can select heroin addicts from the
population and monitor them for a time.
Characteristics of this type of research:

 It is indicated in complex situations in which it is important to relate variables, but in

which experimental control is not possible.
 It allows multiple variables to be measured and interrelated simultaneously in natural
observation situations, as in the examples offered.
 It allows us to identify associations between variables, but we must prevent them
from being false by introducing the appropriate statistical controls.
 It is less rigorous than the experimental type of research because there is no
possibility of manipulating the independent variables or controlling them rigorously.

This type of research is indicated to determine the degree of relationship and similarity that
may exist between two or more variables, that is, between characteristics or concepts of a
phenomenon. It does not attempt to establish a complete explanation of the cause - effect of
what happened, it only provides clues about the possible causes of an event.

Is characterized
✓ Because they measure the variables first
✓ Then, through Hypothesis testing and the application of statistical techniques, the
correlation is stimulated.

Its main objective is to know how a variable can behave by knowing the behavior of another
or other related variables. This is because if there is a relationship between them, when one
variable varies, the other varies in the same way.

This relationship can be positive or negative:

❖ That is, if one variable shows positive results, the same will happen with the other.
❖ The same happens with a negative variable, the results of the other variable will be
negative.

Research Design: partial

*Non-experimental : It is one that is carried out without manipulating variables, only the
phenomenon is observed, as it occurs in a natural context.

*Transversal: It is responsible for describing relationships between two or more variables at

a given time.

*Longitudinal: These analyze changes over time between variables or their relationship
within a general population.

Quantitative Method

It is one in which quantitative data on variables are collected and analyzed, characteristics or
variables that can take numerical values are measured, statistical research procedures.

FUNCTIONAL RELATIONSHIPS AND COVARIATION HYPOTHESES

In essence, correlational research seeks to identify probable relationships between
measured variables. The term “measured variable” is different from the concept
“manipulated variable”, which corresponds to the experimental method; in fact, in
correlational research, variables are not manipulated in the sense that variations in the
independent variable generate changes in the dependent variable. Rather, measurements
are taken of them in order to observe the direction and degree to which they covary, the
psychological, biological, ecological and social variables, which are often interrelated, are
already predetermined at different levels, the researcher focuses on taking measurements of
them, just as you find them in their natural state.

Under these circumstances the researcher can state conjectures about the functional
relationship between the measured variables. The hypotheses that are tested are called
“covariation hypotheses.” These are propositions that affirm the existence of a functional
relationship between two relevant properties that occur within a universe of clearly
determined individuals (Piscoya 1987).
These hypotheses respond to problems such as the following: What is the relationship
between A and B? and are formulated in these terms: there is a relationship between A and
B, replacing the variables a and b with two behaviors, a problem like this could be stated:
What is the relationship between physical punishment, used as a parenting guideline, and
the aggressive behavior of children? ?. If it were assumed that the punishment and
aggression variables were functionally related, the hypothesis would affirm that “there is
physical punishment and aggressive behavior in children.”
If a statistically significant relationship is found between both variables, it would be possible
to infer that among children who receive punishment there will be a greater probability of
finding aggressive subjects.

Interpreting this relationship with causal arguments would lead to affirming that physical
punishment “always generates aggressive behavior.” This inference, in the framework of
correlational research, is naturally incorrect, since many variables that could account for the
aggressiveness of many children have not been controlled, raising alternative hypotheses
that are certainly relevant. The covariation hypothesis does not postulate causal
relationships in the experimental sense, it only verifies the form of association between
variables. It does not have the intention of proving that variable A unequivocally produces
variable B; it focuses on identifying whether the measured variables are concomitant and
establishing variables measured to the observed events, whose direction and magnitude is
determined statistically.

In the covariation hypothesis, it is not easy to identify which of the variables involved is
responsible for the covariation of the other. Take the example of Briones (1986): if the
scores on a religiosity scale correlate with those of authoritarianism, we could not state with
certainty whether religiosity generates authoritarian behavior or authoritarianism determines
religiosity.
There would be at least two possible explanations:

a) The two variables are indicators of the same dimension or more general variable;
According to this assumption, authoritarianism and religiosity would be indicators of a
common factor.
b) The relationship between the variables is spurious, meaning that the covariation
between the two would be due to a third factor that acts as a cause of the variations
of one variable and the other.

This problem does not occur in experimental research that handles hypotheses that handle
cause – effect that adopts the expression: Y = F (x).
In this formulation, the dependent variable “Y” is a function of the dependent variable . What
it is about is finding a variable “that explains the relationship between two variables. Thus, a
variable will be defined as “dependent” when it is related to others belonging to a previous
temporal order (Glock 1973), likewise a variable will be independent as long as it temporally
precedes another. If the hypothesis to be tested were the following: “separation from parents
produces depression in children”, the independent variable will be “separation from parents”
while “depression” by assumption in the analysis will be the dependent variable. If the
hypothesis is correct, the antecedent variable (separation) would explain depressive
behavior in children. It is not an independent variable in the experimental sense that has
been manipulated, nor have values been taken from it; Its name obeys the temporal order of
location with respect to the one to which it is connected. The constant permanence of the
variables is also used, assuming that the most permanent is the independent variable. If a
relationship is found between age and recreational preferences, the age variable is more
permanent than the preferences variable, therefore it assumes the role of independent
variable.

LIMITATION OF CORRELATIONAL STUDIES

The main limitation of correlational studies is that the results do not indicate whether a
cause-effect relationship exists between the variables considered. There are two reasons for
not being able to validate this type of relationship.

Directionality problem. It refers to the impossibility of demonstrating which variable occurs

first (what was the cause) and which occurs second (the effect). In the correlational study the
variables are simply observed and the research does not allow this directionality to be
established.

Third variable problem. By not creating and controlling the variables as occurs in the
experiment, scientists or researchers cannot know if any unobserved or unconsidered
variable is related to each of the other variables and is the current causal agent.

Pe Let's imagine that a researcher is interested in determining the relationship between

students' beer consumption (the night before the exam) and their scores on the exam. Let's
imagine that the correlation coefficient is r = -.825. This r value indicates a moderately high
negative relationship between the number of pints of beer consumed and test scores.
However, even if we find that the two variables are closely related, we cannot maintain that
one variable causes the other due to the problems of directionality and the third variable.
The data from our hypothetical study indicate a high relationship between the number of
pints of beer consumed the night before the exam and the test score. We may be tempted to
conclude that the number of beers consumed caused the test score, but there are other
possible explanations. Drinking beers takes time. Maybe those who waste their time drinking
beers don't spend it studying. Thus, a third variable (study hours) may be the causal source.
Another explanation may be that the exam causes fear which leads some students to drink
more and study less. This difficulty in inferring causality from correlation is one of the
reasons why many psychologists advocate the use of experiments, where causal inferences
can be made with more confidence.
However, correlational research and other non-experimental approaches such as
observation make important contributions to scientific knowledge. These research strategies
help scientists evaluate descriptive questions that can be the basis for further experiments.
Nonexperimental approaches allow scientists to acquire scientific knowledge about events
that cannot be studied under experimental conditions.
BIBLIOGRAPHY:

- Bernal, C (2006). Research methodology 2nd edition. Mexico: Ed. Pearson