Correlation - Definition, Variables and Types

Correlation is defined as the method of creating a relationship or link between two or more than two variables. This method is useful in measuring quantities at a higher or ordinal level. The relation between x and y, or a and b, is known as correlation. It is one of the customary methods approached to calculate the measurements. It incorporates the use of a scattering graph or ‘scatter plot’ to create the relationship between two distinct variables.

Correlation Coefficient: This is a single-valued constant that sums up the relation between any two variables. It is generally denoted by the letter ‘r’. In statistics, the value of the correlation coefficient can reach a minimum of -1 and a maximum of 1, -1 < r < 1. Important results have been deduced due to the range of correlation.

If the correlation coefficient is close to zero (middle), either negative to zero, positive to zero, or zero implies that there is no relationship between the variables.
The correlation coefficient close to +1 means there exists a positive relationship between the two variables.
A correlation coefficient reaching -1 implies the relationship between the two variables to be negative.
Spearman’s Rho is used to calculate the coefficient for ordinal scales, while Pearson’s r is the standard method to find out the coefficient in the case of interval or ratio scales.

The two variables used in making a correlation are generalized as ‘X’ and ‘Y’. They are first related with the help of Pearson’s r method and then represented graphically with the help of a scatter diagram.

Scatter diagram: A scatter diagram is a graphic showing the values of two variables X and Y, and the relationship between these two variables. Variable X values are displayed along the horizontal axis, with Y values displayed on the vertical axis.

Later, one of the parameters is established as an independent parameter when a regression model is employed, and one is specified as the dependent variable. The independent variable X is thought to have some impact or influence on the dependent component Y in regression. The correlation techniques for both variables are symmetric, with no evidence of cause or effect being included in the statistical analysis.

Types of Correlation:

With the help of 'scatter diagram', mathematicians have deduced three types of correlations, which are explained below:

Positive Correlation: Positive correlation occurs due to the values of two parameters moving in the same direction so that a rise or decrease in one variable is followed by an increase or decrease in the other variable. It is further of two types: Strong positive correlation where the points are close to the linear graph, and weak positive correlation where the points are far away from the linear graph.

Negative Correlation: Negative correlation occurs due to the values of two components moving in the opposite direction from each other so that a rise or decrease in one variable is followed by a rise or decrease in the other variable. Negative correlations are also classified into two types. Similar to positive correlation, the negative correlation has a strong negative correlation where the points are close to the linear graph and a weak negative correlation where the points are far away from the linear graph.

No Correlation: when the two variables are not linear or related, it is known as no correlation.