# Spearman and Pearson Correlations – Compare And Contrast Paper

### Grand Canyon –  Sym 506 Weekly Discussion

#### Compare and contrast Spearman and Pearson correlations.

A correlation refers to as a coefficient that evaluates the level to which two variables are likely to change together. This coefficient describes both the direction and the strength of the relationship. There are different forms of correlations. Pearson correlation measures the linear association between two continuous variables. Spearman correlation on the contrary measures the monotonic association between two ordinary or continuous variables. In Pearson, an association is considered linier when one variable change is related with the proportional modification in the other variable. For instance, Pearson can be used to measure whether temperature increase at the facility of production are related with decreasing in the thickness of a coating of a chocolate carried in the pocket. In Monotonic association on the contrary, the variables appear to change together, though not essentially at a constant rate. Therefore the correlation coefficient in Spearman case is founded on the tanked values for every variable instead of the raw data. Spearman rank correlation is referred to as a non-parametric test or a distribution free test which is employed to evaluate the level of relation between two variables. This correlation does not take any assumption regarding the data distribution and it is suitable correlation analysis in situation where the variables are evaluated on an ordinary scale. This is opposed to Pearson correlation that evaluates the association strength between two variables. In Pearson correlation, the two variables are required to be normally distributed. It should also demonstrate homoscedasticity and linearity. Linearity presumes a straight line association between all variables in the analysis while homoscedasticity presumes that data is distributed normally along the line of regression. Based on this analysis, it is evident that the two forms of correlation are considerably different and they are not involved in measuring the same concept in the distribution.