Research Methods · Statistics · continued

Regression to the mean in progress reports posts 31–35

Research Methods › Statistics number needed to treat heterogeneity confounding effect size

This is a continuation of a long topic, addressed by post number rather than by page. Start at post 1.

n.kuuselaTL211 Jul 2025#31

a.molnar, post #2: The thing about regression that took me longest to accept is that a plausible mechanism is not evidence of an effect. It is a reason to look, not a result. Go to post

Answering the question post #29 raises rather than the one it answers.

Worth stating the null on regression before we explain it: the observation may be nothing. That possibility deserves a sentence and usually does not get one.

2 likes in reply to #2 13mo

PharmNotes_WhitfieldTL413 Jul 2025#32

Collapsed as off-topic by two members at trust level 3 or above

The arithmetic in post #29 is right; the assumption feeding it is the part to check.

The most common statistical error in this community is not technical: it is treating a self-selected collection of reports as a sample from a population.

That is the honest state of it as of this week.

s.cabreraTL216 Jul 2025#33

Regression to the mean explains a large share of apparent improvements in anything selected for being extreme. It is not a statistical curiosity; it is the default explanation.

The reasoning is more useful than the number, which is why I have shown it.

27 likes 12mo

dr_okonkwoTL4 Moderator18 Jul 2025#34

My position on regression is current rather than settled. I have revised it once already and I expect to again, so treat it accordingly.

13 likes 12mo

e.halonenTL220 Jul 2025 · edited#35

The honest answer on regression is that it depends, and the useful part is the list of what it depends on. Four items, in rough order of how much they matter.

Most people get the first two right and then argue about the fourth.

4 likes 12mo

This topic was referenced in

About the Statistics categoryResearch Methods › Statistics · 8 replies

Suggested topics

Topic	Participants	Replies	Views	Activity
About the Statistics category Research Methods › Statistics effect size number needed to treat heterogeneity Pinned Wiki Effect sizes, intervals, multiplicity, and the difference between absent and undetected. This post is a community wiki: any member at trust level 3 or above can edit it, and every edit is recorded with its…	+4	8	2.3k	7mo
Second pass at: Regression to the mean in progress reports Research Methods › Statistics effect size worked example heterogeneity Second pass at: Regression to the mean in progress reports Writing it up because I had to work it out twice and would rather nobody else did. Posting a small dataset on Regression. It is mine, it is…	+71	75	4k	17mo
Sample size intuition for a personal experiment Research Methods › Statistics number needed to treat heterogeneity confounding effect size On the subject in the title: Sample size intuition for a personal experiment Working notes rather than a conclusion. A follow-up question about sample size intuition that I did not know to ask the first time.…		4	62k	18mo
Second pass at: Multiplicity when you track fifteen variables Research Methods › Statistics number needed to treat confounding effect size worked example Posting this under the heading it deserves: Second pass at: Multiplicity when you track fifteen variables Everything below is what sits behind that. Two things I would like separated before anyone answers on…		2	12k	18mo
Absence of evidence and evidence of absence Research Methods › Statistics effect size worked example number needed to treat Absence of evidence and evidence of absence Writing it up because I had to work it out twice and would rather nobody else did. Asking about absence of evidence on behalf of the question I keep seeing asked…	+16	20	3.9k	5mo

Related topics — sharing the tags number needed to treat, heterogeneity, confounding

Topic	Participants	Replies	Views	Activity
Pooling trials with different estimands Evidence › Meta-analyses publication bias risk of bias effect size meta-analysis Solved On the subject in the title: Pooling trials with different estimands Working notes rather than a conclusion. Asking about pooling trials with different estimands on behalf of the question I keep seeing asked…	+120	132	2.3k	1d
What TRIUMPH is designed to answer, and why we should not pre-empt it — does this still hold? Compounds › Retatrutide randomised trial effect size TRIUMPH programme retatrutide Solved Asking directly, because I could not find a straight answer: What TRIUMPH is designed to answer, and why we should not pre-empt it — does this still hold? An honest uncertainty about TRIUMPH rather than a…	+43	47	5.5k	just now
Immortal time bias in a claims-database study — one year on Evidence › Study critique effect size risk of bias confounding Solved Posting this under the heading it deserves: Immortal time bias in a claims-database study — one year on Everything below is what sits behind that. I would like to disagree carefully with the settled view on…	+41	46	3.7k	22h
Coming back to: Half-life, steady state, and accumulation worked through Pharmacology › Pharmacokinetics tirzepatide liraglutide worked example semaglutide On the subject in the title: Half-life, steady state, and accumulation worked through Working notes rather than a conclusion. Working through the kinetics rather than the pharmacology, because I think the…	+14	18	15k	3mo
Gauge and length selection for subcutaneous injection Devices › Syringes & needles syringe units needle gauge & length worked example subcutaneous injection Solved Gauge and length selection for subcutaneous injection Writing it up because I had to work it out twice and would rather nobody else did. A follow-up question about gauge and length selection that I did not…	+50	55	1.3k	8mo