t-tests, non-parametric tests, and large studies--a paradox of statistical practice? https://t.co/CI8pFmUklm
@NahanniFinanci1 https://t.co/BHRoutyWcP. This paper is saying so. Reading it as of now.
RT @AFredston: friends, does our field have an analog to this (medical) paper showing that t-tests are more appropriate than non-parametric…
RT @AFredston: friends, does our field have an analog to this (medical) paper showing that t-tests are more appropriate than non-parametric…
friends, does our field have an analog to this (medical) paper showing that t-tests are more appropriate than non-parametric alternatives even for non-Gaussian data when n is large? https://t.co/CXVSoPazSi
An apparent paradox of statistics: For large sample sizes, a t-test may still be the best answer EVEN WHEN distributional assumptions are not met: https://t.co/tqVmwijFjm #DataScience Statistics
RT @JLuis_FG: "Non-parametric tests are most useful for small studies. Using non-parametric tests in large studies may provide answers to t…
"Non-parametric tests are most useful for small studies. Using non-parametric tests in large studies may provide answers to the wrong question, thus ... (1/2) https://t.co/eDe0ZnplrB #epitwitter
"For studies with a large sample size, t-tests and their corresponding confidence intervals can and should be used even for heavily skewed data"
Great paper on why you should probably be using a t-test: https://t.co/XwoPfJboxR