k.otieno_stats
TL3 Statistician · Kisumu, KE
Applied statistician. Effect sizes, confidence intervals, and the difference between "no difference" and "not powered to detect one".
Joined 12 September 2024 · last posted 6d · groups: journal-club
115Posts here
0Topics
0Solutions
518Likes received
620Likes given
25Badges
99Days visited
2.9kPosts read
Stated areas
Recent activity
Full stream-
Revisiting: The best-written scientific paper you have read Replied, post #109 · Lounge › Off topic · 22 Jul 2026 · 0 likes For anyone finding this later: the short answer on best-written scientific paper is that it depends on one thing, and the rest of the thread is people identifying which…
-
Revisiting: PCOS and metabolic overlap: what is and is not studied Replied, post #58 · Clinical › Comorbidities · 22 Jul 2026 · 1 like Post #54 and I disagree about the size of the effect, not about the direction. Where somebody is on several medicines, the interaction question and the comorbidity…
-
[2026 update] Detection wavelength and why 214 nm and 280 nm disagree Replied, post #55 · Analytics › HPLC & UHPLC · 21 Jul 2026 · 11 likes Adding the measurement that post #52 says would settle it. Change the wavelength and the proportions change even though the sample has not. That is the reason the…
-
Prior authorisation: what the criteria usually require — a second dataset Replied, post #45 · Access › Insurance & coverage · 19 Jul 2026 · 26 likes The arithmetic on Prior authorisation is the easy part and it is where the errors are, which is an uncomfortable combination. Show your working and someone will catch it.
-
Why a glucagon receptor agonist in a weight-loss compound is not a contradiction Replied, post #39 · Compounds › Retatrutide · 19 Jul 2026 · 22 likes This follows post #38 rather than contradicting it. Since glucagon receptor agonist keeps coming up, it should probably be a maintained page rather than a recurring…
-
A pooled estimate that changed when one trial was added Replied, post #25 · Evidence › Meta-analyses · 18 Jul 2026 · 15 likes That matches what I have seen, for whatever a single anecdote is worth.
-
[2026 update] Currency conversion and the cost you did not see Replied, post #26 · Sourcing › Payments · 18 Jul 2026 · 0 likes Coming back to post #24, because the follow-up matters more than the original answer. Ask what happens if the material fails independent testing before you order, not…
-
What a certificate commits its issuer to Replied, post #38 · Analytics › COA interpretation · 15 Jul 2026 · 6 likes The distinction between release testing and characterisation is one industrial documents make and retail ones usually do not. It is the difference between "we tested…
-
Why retatrutide discussion here is more cautious than elsewhere — one year on Replied, post #59 · Compounds › Retatrutide · 13 Jul 2026 · 0 likes Second-hand on retatrutide discussion, so weight it accordingly — someone whose method I trust told me this and I have not verified it myself.
-
Follow-up: Signs of dehydration worth acting on Replied, post #23 · Safety › Adverse events · 13 Jul 2026 · 9 likes Following this. I have the same question and no better information than the first post.
-
The anatomy of a question that gets answered within an hour Replied, post #106 · Start Here › How to Ask · 9 Jul 2026 · 0 likes Post the arithmetic you already did. Half the questions here are answered by someone spotting which line of your working is wrong, and that is impossible if you only…
-
What I would want from a home test that does not exist yet Replied, post #28 · Analytics › Home & field testing · 5 Jul 2026 · 16 likes False positives and false negatives: no test has 100% sensitivity and 100% specificity. A negative result does not guarantee absence; a positive result does not…