Pearson v. Spearman Correlations

There are diverse reasons why people examine for correlations between variables; further, there are diverse correlational statistics that they can use when examining for correlations. Two of the most often used correlational statistics for computing the association between two variables are the Pearson correlation statistic and the Spearman correlation statistic (Woodbury, 2001). However, there are theoretical and functional differences between these two statistics (Gravetter & Wallnau, 2006). The purpose of this paper is to discuss these differences.

Appendix A presents two data sets of 20 observations each, along with a brief description of how the data were collected. Examination of these two data sets illustrates the differences that are being discussed. The first difference between the Pearson and Spearman correlations is that they have been developed for different levels of data. Specifically, the Spearman correlation is customarily used when data are ordinal data which is t

. . .
elation does not. However, the inferences allowed on the basis of the Pearson statistic rest on the assumption that the two variables are jointly normally distributed. When this assumption is not justified such as it often is not in a non-linear situation, then a non-parametric measure such as the Spearman correlation statistic is probably going to be more appropriate because it is not based on the assumption that both variables are sampled from populations that follow a normal (Gaussian) distribution. In other words, despite the nature of the collected data, if joint normal distribution cannot be assumed, the powerful Pearson statistic should not be selected. In terms of real world conclusions and inferences made, one point that should be noted is that the Pearson r typically deals with a direct measurement of the variable of interest. In this regard, examination of Appendix A clearly shows that Data Set 1 is based on a test that is a direct measure of mathematical ability. Therefore, any conclusions about students' math ability in relation to their final grade in the class would be specifically about these variables. The foregoing cannot fully be claimed regarding Data Set 2. These data are not direct measures of studen
. . .

Some common words found in the essay are:
Data Set, Gravetter Wallnau, McClave Sincich, Student Team, Course Grade, , Specifically Spearman, Pearson Spearman, Differences Appendix, data set, Data Sets, data set 2, spearman correlation, set 2, set 1, data set 1, woodbury 2001, mathematical ability, true zero, data sets, appendix data, pearson correlation, mcclave sincich 2006, gravetter wallnau 2006, pearson spearman correlations,
Approximate Word count = 1526
Approximate Pages = 6 (250 words per page)

More Essays on Pearson v. Spearman Correlations