What is the correlation range

correlation

Range of values ​​of correlations

Range of values ​​from -1 to +1

Correlations usually relate to linear relationships and have a range of values ​​from -1 to +1. Unless there is a linear relationship between the variables, the value of r is zero. In this case, however, the two variables could also be related to one another in a non-linear form (i.e. non-linear). If the correlation is +1, there is a perfect relationship between the variables. The same applies to a correlation of -1. In this case, the relationship is opposite. For example, in a performance test, the percentage of errors should correlate perfectly in the opposite direction to the percentage of correctly solved tasks. The higher the percentage of mistakes, the lower the percentage of correctly solved tasks. Fig. 11 illustrates some selected correlations between two variables in the form of point clouds.

Figure 11: Representation of selected correlations between two variables.

Low, medium, and high correlation

Among other things, Cohen (1988) specified a convention for correlations that states at which value a correlation should be classified as low, medium or high:

  • r = 0.1 for a low correlation
  • r = 0.3 for a mean correlation
  • r = 0.5 for a high correlation

The exact amount depends on the question

However, it always depends on the question of whether the determined characteristic value can be classified as low or high. For example, one could calculate the relationship between the results from the same multiple-choice test that was determined by two evaluators. This correlation is referred to as evaluation objectivity and should be well above r = 0.5 in order to still be able to speak of an acceptable agreement. Even a correlation of r = 0.9 would still be unsatisfactory in this specific case.