During your time studying Research Methods, correlations are something that will come up frequently. We may even state something in our everyday life, which is a predictive correlation. For example, the co-variable 'a hot day' will be positively correlated with 'sweating a lot'; it is hot today so I will sweat a lot.
If the hot day scenario was to be tested, a researcher might record the temperature changes and how much the participant sweats. Or, the researcher may measure how much participants sweated on a hot day. We expect to find a positive correlation between the Variables. Let's take a look at how correlations are studied in psychology.
- Let's take a look at correlational research in psychology.
- We will start by looking at the correlation meaning, correlation formula and the different types of correlation.
- To finish, we will evaluate correlational research, including the advantages of correlation in psychology and its disadvantages.
Correlational Research Psychology
Correlations are a standard statistical test used in psychology.
Researchers use many types of Statistical Tests, such as correlations, to identify if their data supports the null or alternative hypothesis proposed at the start of their study.
If a correlation is found, this indicates the results support a relationship between the Variables and potentially the alternative hypothesis, a predictive statement suggesting that the results expect to see a relationship between Variables. However, if no correlation is found, then the analysis supports the null hypothesis, a predictive statement that the researcher expects to find no relationship between the Variables.
Correlation Meaning
The correlational research design is a non-experimental technique that does not require the researcher to manipulate the variables. Instead, they measure the variables and then carry out a correlational analysis.
A correlation is a statistical test that tests whether there is an association and relationship between two variables.
An example of an alternative hypothesis that predicts a correlation between two variables is that students who spend more time studying are more likely to perform better in their exams.
An example of a null hypothetical hypothesis that predicts no correlation between two variables is that the amount of milk drank is unlikely to be associated with how tall people grow.
The example above is a hypothesis that can be tested using correlational analysis, as the research can use the test to see if there is a relationship between how long students spent studying and the percentage scores that students received in an exam.
Correlation Formula
In statistical terms, the correlation coefficient is expressed as Pearson's r.
A correlation coefficient is a figure representing the magnitude, i.e., how strong the relationship and association is between two variables.
A positive coefficient suggests a positive relationship between the two variables, and a negative coefficient indicates a negative relationship between the two variables.
The relationship, strength and direction of a correlation can also be visually represented on a scatter diagram. We will use the example above to understand how a scatter diagram can be plotted. To do this, the researcher would need to plot how long each student spent studying against the percentage score they received.
You do not need to learn the Computation correlation formulae for your GCSE studies.
Types of Correlation
When it comes to learning about the types of correlation in psychology, there are two things that we need to keep in mind:
- The magnitude of the correlation (how strong the correlation is)
- The direction of the correlation (positive, negative or no)
Let's start with looking at how you can identify the magnitude of the relationship between two variables. As you may remember, this can be determined from the correlation coefficient. The coefficient can range from -1 to +1, and the negative or plus sign indicates whether the relationship is positive or negative.
The table below summarises which coefficient values represent substantial, moderate, weak or no magnitudes.
Coefficient value (+) | Coefficient value (-) | Magnitude of association |
+1 | - 1 | Perfect correlation |
more than 0.7 but less than 0.9 | more than -0.7 but less than -0.9 | Strong correlation |
more than 0.4 but less than 0.6 | more than -0.4 but less than -0.6 | Moderate correlation |
more than .01 but less than 0.3 | more than -.01 but less than -0.3 | Weak correlation |
0 | 0 | No correlation |
From scatter diagrams, we can interpret the magnitude of correlations. The researcher can estimate a strong positive correlation when each data point is clustered close together. If they are moderately close together, the relationship can be assumed as moderate. And if the data points are widely dispersed or randomly plotted on the scatter diagram, then the correlation can be interpreted as weak or nonexistent.
Sometimes we may use scatterplots instead of coefficient values to interpret whether a correlation is positive, negative or nonexistent. Let's look at examples of how each would be displayed and analysed.
The following data used and shown are completely hypothetical and Vaia Originals.
Positive Types of Correlation
The graph below shows a positive correlation. From the graph, it can be inferred that one co-variable would increase as the other co-variable increases; this is evident as the data points direct upwards. The graph can be interpreted as a positive correlation that indicates that as time spent studying increases, the test scores students receive also increases.
Figure 1: The scatterplot infers a positive correlation between time spent studying and test scores.
Negative Types of Correlation
The graph below shows a negative correlation. From the graph, it can be inferred that as one variable increases, the other decreases; this is evident as the data points direct downwards. The graph can be interpreted as a negative correlation indicating that anxiety scores decrease as time spent sleeping increases.
Figure 2: The scatter plot indicates a negative correlation between time spent sleeping (hrs) and anxiety scores (GAD; lower scores are reflective of low anxiety levels).
Nonexistent Types of Correlation
The graph below shows no correlation or association between the two variables when the chart displays no pattern in the direction of data points. The graph findings will be reported as there is no association between the amount of milk drank and the participants' height.
Figure 3: The scatter plot suggests no correlation between the amount of milk drank (ml in a year) and height grown (cm in a year).
Advantages of Correlation in Psychology
The advantages of correlations in psychology are:
- A correlational research design does not require the researcher to manipulate the variables, so there is less likely that the researchers' bias will affect the study. The advantage of this is that it increases the validity of the research.
- Correlational research is simple to replicate, so it is relatively easy to identify if the study is reliable.
- Correlations can provide many details about how two variables are related, such as the relationship's direction and magnitude. These details are helpful because they allow researchers to identify to what extent two variables are associated.
- When analysing correlational data, it can be easily plotted on a scatterplot; this makes it easier for the researcher and reader to visualise and interpret the study's findings.
- It can be used as a starting point in research, e.g. to help researchers identify if further investigations are required. Further research can help researchers understand why a correlation or no correlation was found, which can't be established with correlations.
Disadvantages of Correlations in Psychology
The disadvantages of correlations in psychology are:
- As correlational research is non-manipulative, it is difficult for the researcher to control confounding factors that may affect the study's validity.
Confounding factors in correlational research is when other factors affect one or both of the investigated variables.
- A correlational analysis is restrictive as it can only be used to analyse Quantitative Data that can be measured on a scale. For example, it isn't easy to use a correlation when analysing data from a Likert scale.
- The cause and effect of correlations can't be established - from the correlation results, we cannot identify which variable is the cause and effect of a phenomenon.
- From correlational research, we can't identify if one variable has more of an effect on the other. Therefore, this analysis has limited utility.
Correlation - Key takeaways
- The correlational research design is a non-experimental technique that does not require the researcher to manipulate the variables. Instead, they measure the variables and then carry out a correlational analysis.
- When it comes to learning about the types of correlation in psychology, there are two things: correlations can tell us the magnitude of the correlation (how strong the correlation is) and the direction of the correlation (positive, negative or no direction).
- Correlation coefficients and scatter plots can tell us the magnitude and direction of correlations.
- There are three main types of correlation: positive, negative and no direction. These can be further subdivided into perfect, strong, moderate, weak or no magnitudes.
- There are many advantages of correlation in psychology and disadvantages. Correlations help visualise data, for instance, allowing for easy interpretation, but the interpretation cannot provide cause-and-effect data.
Explanations 
Textbooks 
Exams 
Magazine 
