The Peptide CommonsEst. May 2024
Independent. We sell nothing and are affiliated with no manufacturer or pharmacy. Every moderation action is logged in public
Evidence · Study critique · continued

[2026 update] Confounding by indication, explained with a concrete example posts 31–60

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

BF
b.friskTL26 Feb 2026#31
TK
t.kulkarniTL3Regular6 Feb 2026#32
k.redgrave, post #10: A criticism that would apply equally to every trial in the field is worth stating once and is not a reason to discount a particular paper. Go to post

Post #30 put the caveat in the right place and I want to underline it.

Multiple comparisons: if a paper reports many outcomes, the chance of a spurious association by random chance is real. Pre-specification of primary outcomes matters and secondary analyses are weaker evidence.

This has been discussed before and I could not find the thread, so, again.

1 like in reply to #10 6mo
VB
v.bergstromTL27 Feb 2026#33

Surrogate endpoints are not automatically bad and their validity is compound-specific and population-specific. The question is whether this surrogate has been validated for this use.

6 likes 6mo
RJ
r.jhannsdttirTL3Regular7 Feb 2026 · edited#34

A run-in period that excludes non-responders before randomisation changes what the trial is estimating. It is legitimate design and it must be stated in any summary.

17 likes 6mo
JP
j.palaciosTL28 Feb 2026#35

Statistical significance and clinical importance are different and both are needed. A significant difference below the minimal important difference is a real finding of no practical consequence.

0 likes 6mo
VT
vial_tableTL2Member8 Feb 2026#36
Ridgeway, post #8: Generalisability and validity are separate axes. A trial can be internally impeccable and still tell you nothing about the person asking. The conclusion is tentative; the arithmetic underneath it is not. Go to post

Where I part company with post #34, and it is a narrow parting.

A criticism that would apply equally to every trial in the field is worth stating once and is not a reason to discount a particular paper.

Not disagreeing with anyone above, just adding the bit I keep having to look up.

3 likes in reply to #8 6mo
GO
g.oyelaranTL28 Feb 2026#37

Adding the measurement that post #36 says would settle it.

Generalisability: do the inclusion/exclusion criteria narrow the population so much that results do not apply to real people asking about it? This is a fair criticism but requires specificity about which real people and why the difference matters.

Adding a source would improve this post and I do not have one to hand.

11 likes 6mo
IL
integrator_logTL3Regular9 Feb 2026#38

Defending a paper against criticism: if the authors respond, they might clarify something the paper explained poorly. Their response might also miss your point. Either way, the exchange in public is more useful than quiet disagreement.

Not the answer, but possibly the question that gets there.

23 likes 6mo
SG
s.girardTL29 Feb 2026#39

Criticism is more useful when it is narrower. "The trial answers a different question from the one being asked" is actionable; "the trial is flawed" is not.

1 like 6mo
CO
c.okaforTL39 Feb 2026#40
ST
s.teixeiraTL210 Feb 2026#41

Post #39 and I disagree about the size of the effect, not about the direction.

Attrition is the failure mode most likely to invalidate a result and the least likely to be discussed. Differential attrition between arms is the specific thing to look for.

0 likes 6mo
K
KAnderssonTL3Regular10 Feb 2026 · edited#42
t.kulkarni, post #32: Post #30 put the caveat in the right place and I want to underline it. Multiple comparisons: if a paper reports many outcomes, the chance of a spurious association by random chance is real. Pre-specification of primary outcomes matters and secondary analyses are weaker evidence. This has been discussed before and I could not find the… Go to post

Per-protocol and intention-to-treat analyses answer different questions and neither is the honest one by default. Reporting both is the practice worth insisting on.

0 likes in reply to #32 6mo
EN
e.ndiayeTL211 Feb 2026#43

Generalisability and validity are separate axes. A trial can be internally impeccable and still tell you nothing about the person asking.

I am confident about the direction and much less about the magnitude.

13 likes 5mo
DS
d.szymanskiTL3Wiki editor11 Feb 2026#44

Narrowing post #43, because the general version has more than one answer.

Surrogate endpoints are not automatically bad and their validity is compound-specific and population-specific. The question is whether this surrogate has been validated for this use.

Caveat: everything above assumes the paperwork is what it says it is.

4 likes 5mo
BC
b.correiaTL211 Feb 2026#45
i.grimaldi, post #5: The opening post describes the usual case. This is about the unusual one. Per-protocol and intention-to-treat analyses answer different questions and neither is the honest one by default. Reporting both is the practice worth insisting on. The step people skip is the one I have spelled out. Go to post

Per-protocol and intention-to-treat analyses answer different questions and neither is the honest one by default. Reporting both is the practice worth insisting on.

2 likes in reply to #5 5mo
GV
g.valckenaereTL3Regular12 Feb 2026#46
c.okafor, post #40: Hold a trial to the standard something could actually have met. A criticism that no achievable design could have answered is a criticism of the field rather than of the paper. Go to post

A criticism that would apply equally to every trial in the field is worth stating once and is not a reason to discount a particular paper.

If that is already documented somewhere, ignore me and link it.

0 likes in reply to #40 5mo
SD
st.dialloTL212 Feb 2026#47

I read post #43 twice before replying, because I had assumed the opposite.

Statistical significance and clinical importance are different and both are needed. A significant difference below the minimal important difference is a real finding of no practical consequence.

19 likes 5mo
H
HHidalgoTL2Member12 Feb 2026#48

Thank you for taking the time. That was more work than a reply usually is.

8 likes 5mo
AJ
a.jansenTL213 Feb 2026#49
KForsberg, post #11: Taking post #10 at face value and following it one step further. Building consensus on which criticisms matter: if everyone agrees that the sample size is small but only you think that affects the conclusion, maybe your criticism is more idiosyncratic. That does not make it wrong but it is worth noticing. Old habit: I write down the… Go to post

Generalisability: do the inclusion/exclusion criteria narrow the population so much that results do not apply to real people asking about it? This is a fair criticism but requires specificity about which real people and why the difference matters.

Where I would look next, rather than where I would stop.

4 likes in reply to #11 5mo
MD
m.dalgaardTL3Regular13 Feb 2026#50

Confounding: in observational data, is there a third variable that explains the apparent association? In randomised data, randomisation should balance unknown confounders, though known confounders can be adjusted for.

It is the kind of thing that is obvious once and never again.

0 likes 5mo
NL
n.laurentTL213 Feb 2026#51

Multiple comparisons: if a paper reports many outcomes, the chance of a spurious association by random chance is real. Pre-specification of primary outcomes matters and secondary analyses are weaker evidence.

2 likes 5mo
CR
compounding_ruthTL4Pharmacist14 Feb 2026#52

Building consensus on which criticisms matter: if everyone agrees that the sample size is small but only you think that affects the conclusion, maybe your criticism is more idiosyncratic. That does not make it wrong but it is worth noticing.

9 likes 5mo
IN
i.norgaardTL214 Feb 2026#53

Confirming post #50 from a second method, which matters more than confirming it from a second person.

Criticism is more useful when it is narrower. "The trial answers a different question from the one being asked" is actionable; "the trial is flawed" is not.

Someone will know this better than I do and I hope they say so.

28 likes 5mo
IT
impurity_tableTL3Analytical chemist15 Feb 2026#54
e.ndiaye, post #43: Generalisability and validity are separate axes. A trial can be internally impeccable and still tell you nothing about the person asking. I am confident about the direction and much less about the magnitude. Go to post

I had written a reply contradicting post #52 and deleted it. Here is what survived.

Hold a trial to the standard something could actually have met. A criticism that no achievable design could have answered is a criticism of the field rather than of the paper.

I would want a second opinion before relying on that.

0 likes in reply to #43 5mo
SO
s.ostergaardTL215 Feb 2026#55

Seconded. It reads as careful rather than confident, which is the right register.

0 likes 5mo
BV
bias_varianceTL4Biostatistician15 Feb 2026#56

Defending a paper against criticism: if the authors respond, they might clarify something the paper explained poorly. Their response might also miss your point. Either way, the exchange in public is more useful than quiet disagreement.

That is all the detail I have. Someone else will have more.

5 likes 5mo
MS
m.steinerTL216 Feb 2026#57

Post #54 is right about the mechanism and I think understates the practical bit.

The pre-specified endpoint being a weaker proxy than you would like is a real criticism. It is a smaller one than saying the result was chosen after the fact.

20 likes 5mo
B
batchlogTL3Regular16 Feb 2026 · edited#58
compounding_ruth, post #52: Building consensus on which criticisms matter: if everyone agrees that the sample size is small but only you think that affects the conclusion, maybe your criticism is more idiosyncratic. That does not make it wrong but it is worth noticing. Go to post

Attrition is the failure mode most likely to invalidate a result and the least likely to be discussed. Differential attrition between arms is the specific thing to look for.

0 likes in reply to #52 5mo
VK
v.krastevTL216 Feb 2026#59
f.villalobos, post #20: Building consensus on which criticisms matter: if everyone agrees that the sample size is small but only you think that affects the conclusion, maybe your criticism is more idiosyncratic. That does not make it wrong but it is worth noticing. Go to post

A run-in period that excludes non-responders before randomisation changes what the trial is estimating. It is legitimate design and it must be stated in any summary.

Reading it again, the caveat matters more than the finding.

8 likes in reply to #20 5mo
MA
m.achebeTL217 Feb 2026#60

Post #56 and I disagree about the size of the effect, not about the direction.

Per-protocol and intention-to-treat analyses answer different questions and neither is the honest one by default. Reporting both is the practice worth insisting on.

This is the sort of thing the wiki should carry and currently does not.

19 likes 5mo