The unequal variance t-test is an underused alternative to Student's t-test and the Mann–Whitney U test

By Graeme D. Ruxton. 2006

1,945 citations recorded by the source corpus.

Subfields: Food Science, Statistics and Probability, Water Science and Technology

View Paper: https://www.semanticscholar.org/paper/f51a91f31e8f67edbdf4e33b8661c7b1bffffc7e

How researchers see this paper

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

In Food Science

In Statistics and Probability

In Water Science and Technology

Citations in the corpus

Blackbirds sing higher-pitched songs in cities: adaptation to habitat acoustics or side-effect of urbanization? (2009, methodology): “1) and thus the Mann–Whitney U test was appropriate to test group differences (Ruxton 2006).”
Five-Point Likert Items: t test versus Mann-Whitney-Wilcoxon ( Addendum added October 2012 ) (2010, methodology): “In such cases, separate-variance procedures such as the Welch test are recommended as being more Type I error robust (Cribbie & Keselman, 2003; Ruxton, 2006; Zimmerman, 2006).”
Rules of engagement for resource contests in a social fish (2011, methodology): “We compared winner and loser size and behaviour using paired t tests and the content of male and female contests using Welch’s t tests (as recommended by Ruxton 2006).”
Why Psychologists Should by Default Use Welch's t-test Instead of Student's t-test with Unequal Group Sizes (2017, methodology): “6 For example, many statistical users believe that the Mann-Whitney non-parametric test can cope with both normality and homosedasticity issues (Ruxton, 2006).”
A practical guide for using statistical tests to assess randomized algorithms in software engineering (2011, methodology): “used? Based on the discussions in this section, in contrast to [54] and in line with [38], we suggest to use Mann-Whitney U-test rather than t-test and Welch test.”
Neglected biological patterns in the residuals (2011, background): “Similarly, Ruxton (2006) reported that the unequal variances t-test has been neglected by behavioural ecologists in the past.”
Fold change and p-value cutoffs significantly alter microarray interpretations (2012, methodology): “Due to small sample size, equal variance across data could not be assumed and data was analyzed using an unpaired TTest with unequal variance with no statistical corrections [21].”
Equivalence Tests (2017, neutral): “Even though Student’s t test is by far the most popular t test in psychology, there is general agreement that whenever the number of observations are unequal across both conditions, Welch’s t test (1947), which does not rely on the assumption of equal variances, should be performed by default (Delac”
Five-Point Likert Items: t test versus Mann-Whitney-Wilcoxon (2010, methodology): “In such cases, separate-variance procedures such as the Welch test are recommended as being more Type I error robust (Cribbie & Keselman, 2003; Ruxton, 2006; Zimmerman, 2006).”
Why Welch’s test is Type I error robust (2016, methodology): “Ruxton (2006) advocates the routine use of Welch’s test.”
Using the Student's "t"-Test with Extremely Small Sample Sizes. (2013, neutral): “Similar results for unequal sample sizes (Ns between 6 and 25) and unequal variances were reported in a review article about the Welch test by Ruxton (2006).”
Biotic homogenization at the community scale: disentangling the roles of urbanization and plant invasion (2013, methodology): “We tested for significant differences in alpha-diversity indices of Robinia and Betula plots using Welch t-tests which is frequently used for data sets with unequal sample variances (Ruxton, 2006).”
Recent advances in differential evolution: a survey and experimental analysis (2010, methodology): “In order to prove the statistical significance of the results in Table 2, a two tail unequal variance t-test has been applied according to the description given in Ruxton (2006) (see also NIST/SEMATECH 2003) for a confidence level of 0.95.”
Song amplitude affects territorial aggression of male receivers in chaffinches (2011, methodology): “Comparisons of the PCA scores, as well as comparisons of the raw response variables, were made with Welch 2-sample t-tests, following the recommendations of Ruxton (2006).”
Changing philosophies and tools for statistical inferences in behavioral ecology (2009, methodology): “In carefully designed experiments, balanced or randomized group assignment allows the data to be analyzed with straightforward statistical tests like analysis of variance (ANOVA), t-test, regression or mixed-models (Ruxton 2006; Stephens, Buskirk, and del Rio 2007).”
Test statistics for the comparison of means for two samples which include both paired observations and independent observations (2017, result): “This is in line with empirical evidence for the performance of Welch’s test, when only independent samples are present, which leads to many observers recommending the routine use of Welch’s test under normality (e.g. Ruxton, 2006).”