How to enter prior information (Mean age and SD) into JASP for Bayesian two sample t-test
I am a Pediatrician interested in clinical application of statistics principles.
I understand to some extent frequentist statistics basic to intermediate level as a non-statistician.
Now as a novice to Bayesian statistics, I want to do Bayesian two sample independent t-test.
Study hypothesis: Is there age difference between malaria and dengue children?
There is difference (H1) & No difference (H0).
I have to input prior age observations of malaria and dengue children from previous studies.
How and where to input the prior data of age: mean and Sd of malaria and dengue children in JASP test input area for Bayesian analysis?
I have present data that I wish to import to JASP from a CSV file.
I request step by step guidance to learn doing the test with JASP.
Thanking you.
Comments
Hey! In JASP, you need to go to the "Input" section under the Bayesian test. There, where it says priors, you enter the mean and standard deviation for the malaria and dengue children. You input them as "informative priors". Be careful with your parameters! It's crucial. [url=https://projektowanie.art/]😉[/url]
Noah
Thanks for the tip,
I tried following the instructions.
But in input section I find the following:
I can only enter prior information of either malaria children or dengue children.
How can I enter the two groups' prior age information: mean and Sd (location and scale are given for one entry)?
The prior information that is test-relevant is the expected effect size under H1. This is usually expressed as Cohen's d, that is, the difference in age divided by the standard deviation. You can certainly specify a prior based on previous research (an "empirical prior") but it needs to be on the scale of Cohen's d.
EJ
Thank you.
I try learning and doing it and seek further help if needed.
Is there a comprehensive help file or user's manual of JASP that explains the components of input and output of Bayesian analysis for a clinician who is not a statistician?
Thank you once again for the guidance.
On https://jasp-stats.org/jasp-materials/ you will find Mark Goss-Sampson's manual Bayesian Inference in JASP. My website has several tutorial articles, and each year we organize a hybrid workshop on Bayesian hypothesis tests at the University of Amsterdam. I am about to start work on a course book but that project is likely to take a few years to finish.
EJ
Thank you for the inputs for learning.