Abstract
In studies with two unpaired samples of continuous data the difference of the two means and their pooled standard error is usually compared to zero. Instead of an unpaired t-test also a Bayesian analysis on the difference of two group means is possible. A traditional unpaired t-test of parallel group study provided a statistically significantly difference in treatment efficacy with t = −3.558, p-value = 0.002. A Bayesian unpaired t-test provided support in favor of the traditional test with a Bayes factor of 0.056. The robustness was assessed.
1. Bootstraps unpaired t-test 95% confidence interval | −2.61536 to −0.71752 |
2. Bayesian unpaired t-test 95% credible interval | −2.8098 to −0.6302 |
3. Gaussian unpaired t-test 95% confidence interval | −2.73557 to −0.79443 |
Obviously the 95% confidence interval of the bootstraps t-test was closer to the traditional Gaussian t-test than it was to the Bayesian t-test. Under the assumption that bootstrap sampling is entirely without overfitting, this would be an argument of overfitting of the Bayesian t-test and an argument in favor of the traditional Gaussian approach. However, with an informed prior this was less a problem.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
To readers requesting more background, theoretical and mathematical information of computations given, several textbooks complementary to the current production and written by the same authors are available.
- 2.
All of them have been written by the same authors and they have been edited by Springer Heidelberg Germany.
Suggested Reading ,
To readers requesting more background, theoretical and mathematical information of computations given, several textbooks complementary to the current production and written by the same authors are available.
All of them have been written by the same authors and they have been edited by Springer Heidelberg Germany.
Statistics applied to clinical studies 5th edition, 2012
Machine learning in medicine a complete overview, 2015
SPSS for starters and 2nd levelers 2nd edition, 2015
Clinical data analysis on a pocket calculator 2nd edition, 2016
Understanding clinical data analysis from published research, 2016
Modern Meta-analysis, 2017
Regression Analysis in Clinical Research, 2018
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer International Publishing AG, part of Springer Nature
About this chapter
Cite this chapter
Cleophas, T.J., Zwinderman, A.H. (2018). Bayesian Unpaired T-Test. In: Modern Bayesian Statistics in Clinical Research . Springer, Cham. https://doi.org/10.1007/978-3-319-92747-3_6
Download citation
DOI: https://doi.org/10.1007/978-3-319-92747-3_6
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-92746-6
Online ISBN: 978-3-319-92747-3
eBook Packages: MedicineMedicine (R0)