## how many academics does it take to change… a p-value threshold?

Posted in Books, pictures, Running, Statistics, Travel with tags , , , , , , , , on August 22, 2017 by xi'an

“…a critical mass of researchers now endorse this change.”

The answer to the lightpulp question seems to be 72: Andrew sent me a short paper recently PsyarXived and to appear in Nature Human Behaviour following on the .005 not .05 tune we criticised in PNAS a while ago. (Actually a very short paper once the names and affiliations of all authors are taken away.) With indeed 72 authors, many of them my Bayesian friends! I figure the mass signature is aimed at convincing users of p-values of a consensus among statisticians. Or a “critical mass” as stated in the note. On the next week, Nature had an entry on this proposal. (With a survey on whether the p-value threshold should change!)

The argument therein [and hence my reservations] is about the same as in Val Johnson’s original PNAS paper, namely that .005 should become the reference cutoff when using p-values for discovering new effects. The tone of the note is mostly Bayesian in that it defends the Bayes factor as a better alternative I would call the b-value. And produces graphs that relate p-values to some minimax Bayes factors. In the simplest possible case of testing for the nullity of a normal mean. Which I do not think is particularly convincing when considering more realistic settings with (many) nuisance parameters and possible latent variables where numerical answers diverge between p-values and [an infinity of] b-values. And of course the unsolved issue of scaling the Bayes factor. (This without embarking anew upon a full-fledged criticism of the Bayes factor.) As usual, I am also skeptical of mentions of power, since I never truly understood the point of power, which depends on the alternative model, increasingly so with the complexity of this alternative. As argued in our letter to PNAS, the central issue that this proposal fails to address is the urgency in abandoning the notion [indoctrinated in generations of students that a single quantity and a single bound are the answers to testing issues. Changing the bound sounds like suggesting to paint afresh a building on the verge of collapsing.

## ASA’s statement on p-values [#2]

Posted in Books, Kids, Statistics, University life with tags , , , , , , , , on March 9, 2016 by xi'an

It took a visit on FiveThirtyEight to realise the ASA statement I mentioned yesterday was followed by individual entries from most members of the panel, much more diverse and deeper than the statement itself! Without discussing each and all comments, some points I subscribe to

• it does not make sense to try to replace the p-value and the 5% boundary by something else but of the same nature. This was the main line of our criticism of Valen Johnson’s PNAS paper with Andrew.
• it does not either make sense to try to come up with a hard set answer about whether or not a certain parameter satisfies a certain constraint. A comparison of predictive performances at or around the observed data sounds much more sensible, if less definitive.
• the Bayes factor is often advanced as a viable alternative to the p-value in those comments, but it suffers from difficulties exposed in our recent testing by mixture paper, one being the lack of absolute scale.
• we seem unable to escape the landscape set by Neyman and Pearson when constructing their testing formalism, including the highly unrealistic 0-1 loss function. And the grossly asymmetric opposition between null and alternative hypotheses.
• the behaviour of any procedure of choice should be evaluated under different scenarios, most likely by simulation, including some accounting for misspecified models. Which may require an extra bit of non-parametrics. And we should abstain from considering further than evaluating whether or not the data looks compatible with each of the scenarios. Or how much through the mixture representation.

## two, three, five, …, a million standard deviations!

Posted in Books, Statistics, University life with tags , , , , , , , on September 26, 2014 by xi'an

I first spotted Peter Coles’ great post title “Frequentism: the art of probably answering the wrong question” (a very sensible piece by the way!, and mentioning a physicist’s view on the Jeffreys-Lindley paradox I had intended to comment) and from there the following site jumping occured:

“I confess that in my early in my career as a physicist I was rather cynical about sophisticated statistical tools, being of the opinion that “if any of this makes a difference, just get more data”. That is, if you do enough experiments, the confidence level will be so high that the exact statistical treatment you use to evaluate it is irrelevant.” John Butterworth, Sept. 15, 2014

After Val Johnson‘s suggestion to move the significant level from .05 down to .005, hence roughly from 2σ up to 3σ, John Butterworth, a physicist whose book Smashing Physics just came out, discusses in The Guardian the practice of using 5σ in Physics. It is actually induced by Louis Lyons’ arXival of a recent talk with the following points (discussed below):

1. Should we insist on the 5 sigma criterion for discovery claims?
2. The probability of A, given B, is not the same as the probability of B, given A.
3. The meaning of p-values.
4. What is Wilks Theorem and when does it not apply?
5. How should we deal with the `Look Elsewhere Effect’?
6. Dealing with systematics such as background parametrisation.
7. Coverage: What is it and does my method have the correct coverage?
8. The use of p0 versus p1 plots.

## Adaptive revised standards for statistical evidence [guest post]

Posted in Books, Statistics, University life with tags , , , , , , , on March 25, 2014 by xi'an

[Here is a discussion of Valen Johnson’s PNAS paper written by Luis Pericchi, Carlos Pereira, and María-Eglée Pérez, in conjunction with an arXived paper of them I never came to discuss. This has been accepted by PNAS along with a large number of other letters. Our discussion permuting the terms of the original title also got accepted.]

Johnson [1] argues for decreasing the bar of statistical significance from 0.05 and 0.01 to 0:005 and 0:001 respectively. There is growing evidence that the canonical fixed standards of significance are inappropriate. However, the author simply proposes other fixed standards. The essence of the problem of classical testing of significance lies on its goal of minimizing type II error (false negative) for a fixed type I error (false positive). A real departure instead would be to minimize a weighted sum of the two errors, as proposed by Jeffreys [2]. Significance levels that are constant with respect to sample size do not balance errors. Size levels of 0.005 and 0.001 certainly will lower false positives (type I error) to the expense of increasing type II error, unless the study is carefully de- signed, which is not always the case or not even possible. If the sample size is small the type II error can become unacceptably large. On the other hand for large sample sizes, 0.005 and 0.001 levels may be too high. Consider the Psychokinetic data, Good [3]: the null hypothesis is that individuals can- not change by mental concentration the proportion of 1’s in a sequence of n = 104; 490; 000 0’s and 1’s, generated originally with a proportion of 1=2. The proportion of 1’s recorded was 0:5001768. The observed p-value is p = 0.0003, therefore according to the present revision of standards, still the null hypothesis is rejected and a Psychokinetic effect claimed. This is contrary to intuition and to virtually any Bayes Factor. On the other hand to make the standards adaptable to the amount of information (see also Raftery [4]) Perez and Pericchi [5] approximate the behavior of Bayes Factors by,

$\alpha_{\mathrm{ref}}(n)=\alpha\,\dfrac{\sqrt{n_0(\log(n_0)+\chi^2_\alpha(1))}}{\sqrt{n(\log(n)+\chi^2_\alpha(1))}}$

This formula establishes a bridge between carefully designed tests and the adaptive behavior of Bayesian tests: The value n0 comes from a theoretical design for which a value of both errors has been specified ed, and n is the actual (larger) sample size. In the Psychokinetic data n0 = 44,529 for type I error of 0:01, type II error of 0.05 to detect a difference of 0.01. The αref (104, 490,000) = 0.00017 and the null of no Psychokinetic effect is accepted.

A simple constant recipe is not the solution to the problem. The standard how to judge the evidence should be a function of the amount of information. Johnson’s main message is to toughen the standards and design the experiments accordingly. This is welcomed whenever possible. But it does not balance type I and type II errors: it would be misleading to pass the message—use now standards divided by ten, regardless of neither type II errors nor sample sizes. This would move the problem without solving it.

## a refutation of Johnson’s PNAS paper

Posted in Books, Statistics, University life with tags , , , , , , , on February 11, 2014 by xi'an

Jean-Christophe Mourrat recently arXived a paper “P-value tests and publication bias as causes for high rate of non-reproducible scientific results?”, intended as a rebuttal of Val Johnson’s PNAS paper. The arguments therein are not particularly compelling. (Just as ours’ may sound so to the author.)

“We do not discuss the validity of this [Bayesian] hypothesis here, but we explain in the supplementary material that if taken seriously, it leads to incoherent results, and should thus be avoided for practical purposes.”

The refutation is primarily argued as a rejection of the whole Bayesian perspective. (Although we argue Johnson’ perspective is not that Bayesian…) But the argument within the paper is much simpler: if the probability of rejection under the null is at most 5%, then the overall proportion of false positives is also at most 5% and not 20% as argued in Johnson…! Just as simple as this. Unfortunately, the author mixes conditional and unconditional, frequentist and Bayesian probability models. As well as conditioning upon the data and conditioning upon the rejection region… Read at your own risk. Continue reading

## Statistical evidence for revised standards

Posted in Statistics, University life with tags , , , , , , , , , on December 30, 2013 by xi'an

In yet another permutation of the original title (!), Andrew Gelman posted the answer Val Johnson sent him after our (submitted)  letter to PNAS. As Val did not send me a copy (although Andrew did!), I will not reproduce it here and I rather refer the interested readers to Andrews’ blog… In addition to Andrew’s (sensible) points, here are a few idle (post-X’mas and pre-skiing) reflections:

• “evidence against a false null hypothesis accrues exponentially fast” makes me wonder in which metric this exponential rate (in γ?) occurs;
• that “most decision-theoretic analyses of the optimal threshold to use for declaring a significant finding would lead to evidence thresholds that are substantially greater than 5 (and probably also greater 25)” is difficult to accept as an argument since there is no trace of a decision-theoretic argument in the whole paper;
• Val rejects our minimaxity argument on the basis that “[UMPBTs] do not involve minimization of maximum loss” but the prior that corresponds to those tests is minimising the integrated probability of not rejecting at threshold level γ, a loss function integrated against parameter and observation, a Bayes risk in other words… Point masses or spike priors are clearly characteristics of minimax priors. Furthermore, the additional argument that “in most applications, however, a unique loss function/prior distribution combination does not exist” has been used by many to refute the Bayesian perspective and makes me wonder what are the arguments left in using a (pseudo-)Bayesian approach;
• the next paragraph is pure tautology: the fact that “no other test, based on either a subjectively or objectively specified alternative hypothesis, is as likely to produce a Bayes factor that exceeds the specified evidence threshold” is a paraphrase of the definition of UMPBTs, not an argument. I do not see we should solely “worry about false negatives”, since minimising those should lead to a point mass on the null (or, more seriously, should not lead to the minimax-like selection of the prior under the alternative).

## Revised evidence for statistical standards

Posted in Kids, Statistics, University life with tags , , , , , , , , on December 19, 2013 by xi'an

We just submitted a letter to PNAS with Andrew Gelman last week, in reaction to Val Johnson’s recent paper “Revised standards for statistical evidence”, essentially summing up our earlier comments within 500 words. Actually, we wrote one draft each! In particular, Andrew came up with the (neat) rhetorical idea of alternative Ronald Fishers living in parallel universes who had each set a different significance reference level and for whom alternative Val Johnsons would rise and propose a modification of the corresponding Fisher’s level. For which I made the above graph, left out of the letter and its 500 words. It relates “the old z” and “the new z”, meaning the boundaries of the rejection zones when, for each golden dot, the “old z” is the previous “new z” and “the new z” is Johnson’s transform. We even figured out that Val’s transform was bringing the significance down by a factor of 10 in a large range of values. As an aside, we also wondered why most of the supplementary material was spent on deriving UMPBTs for specific (formal) problems when the goal of the paper sounded much more global…

As I am aware we are not the only ones to have submitted a letter about Johnson’s proposal, I am quite curious at the reception we will get from the editor! (Although I have to point out that all of my earlier submissions of letters to to PNAS got accepted.)