Testing the significance of a correlation with nonnormal data: comparison of Pearson, Spearman, transformation, and resampling approaches.

By Anthony J. Bishara, James B. Hittner. Psychological methods, 2012

720 citations recorded by the source corpus.

Subfields: Computer Networks and Communications, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/1dc5b2114d1ff561fc7d6163d8f4e9c905ca12c4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

In Statistics and Probability

Citations in the corpus

Transparency and replicability in qualitative research: The case of interviews with elite informants (2019, background): “So, although Pearson's r is the most frequently used correlational test, results can be biased when variables are from distributions with heavy tails (Bishara & Hittner, 2012; de Winter, Gosling, & Potter, 2016).”
Comparing the Pearson and Spearman correlation coefficients across distributions and sample sizes: A tutorial using simulations and empirical data. (2016, methodology): “There are, of course, a large number of other types of data transformations, such as a logarithmic, multiplicative inverse, or power transformation, that can be successfully applied prior to calculating the Pearson correlation coefficient (Bishara & Hittner, 2012).”
Violating the normality assumption may be the lesser of two evils (2018, methodology): “…regression coefficients (Williams, Grajales & Kurkiewicz 2013) and usually do not impair hypothesis testing (no inflated type I error rate, e.g. Bishara & Hittner 2012; Puth, Neuhauser & Ruxton 2014; Ives 2015; Szöcs & Schäfer 2015; Warton et al. 2016) even at relatively small sample sizes,…”
The generality-specificity of creativity: Exploring the structure of creative potential with EPoC (2016, methodology): “…of all variables were homogenized into normal distributions using the Rankit method (e.g.,Solomon & Sawilowsky, 2009), a highly applicable transformation that can approximately normalize any distribution shape (Bishara & Hittner, 2012), a particularly desirable feature for planned analyses.”