If you go via the ordinal regression route, you are essentially dividing your “active treatment” (X = 1) population into 4 slices: subjects who give you a rating of 1, 2, 3 or 4, respectively. Similarly for the “control treatment” (X = 0).

@vdveenb.bsky.social Bert, I have to be creative about sharing photos since the DM is text only. Here’s a “spike plot” that shows the (discrete) number of fish caught by the two of us and illustrates where the mean of the Poisson distribution falls. Sorry about my primitive drawing skills. 🤭

I did eventually manage to find this table (substitute the word “likely” with the word “probable” for the purpose of our discussion) in this article:

efsa.onlinelibrary.wiley.com/doi/full/10....

But don’t use this abomination if you can help it. Using that much ink to display the means as bars is not necessary. 🤭

Your book topic sounds really interesting! Do you have anything up online? Would love to check it out.

This was on my Twitter timeline - would make for a great book cover: Art by Duy Huynh. 👇

David, I am so excited to discover you here. My niece from Romania is an aspiring designer of characters (self-taught) so I will share your website with her so that she can draw inspiration from your artistic journey.

Wild flowers in the rain…The vibrant colours are so lovely…

The last standing wild flowers at our local dog park…