[MORPHMET] Re: PREPRINT: between-group principal components analysis (bgPCA)

2019-05-18 Thread MORPHMET
[Final, summary post.]

I thought I should post some notes about the responses to Andrea’s recent 
postings. Several people seem to have extrapolated beyond what Andrea said.

After a few months of the Cardini, O’Higgins, Rohlf, and Bookstein 
collaboration (Polly was CC:ed on the very early emails but I do not 
remember him engaging), I suggested to Andrea that there should be two 
publications not one. I could see that while writing about the same flaws 
in the same method the approaches Fred and I were taking would be difficult 
to combine in the same paper. Andrea and Paul agreed and Andrea then 
suggested to Fred that he write a paper separate from ours.  Thus, Fred did 
not unilaterally jump ahead and just write his own paper as some seem to 
have assumed. The understanding was that we expected the two papers to be 
published together as “companion papers” in some journal yet to be 

I know that Fred was concerned about the ethics of waiting a long time to 
warn people that they were using a severely flawed (I would say strongly 
biased) method.  The usage of BG-PCA seems to have increased lately and it 
did not seem fair to let people continue to write papers and dissertations 
based on this method once we knew how bad it was. His biorxiv upload and 
his announcement of it on morphmet were not a surprise to us. There were 
emails exchanged about the need to warn users. At the time, Fred told me 
that he needed to go ahead with the biorxiv upload as it was unclear how 
long it would take our ms. to be completed due to other demands on our time.

Reading Fred’s papers can take time but if one just looks at his Fig. 1 
(Google "biorxiv 627448" if you lost the link that Fred posted) you will 
see the magnitude of the problem. It is not subtle! Incidentally, the 
Cardini et al. draft also has a more extensive Figure illustrating the same 
problem as a function of n but the rest of the paper is very different. In 
fact, I found it interesting how two papers about the same defect in the 
same method and reaching the same conclusion could have so little overlap. 
Thus, Fred’s paper does not infringe on the content of the Cardini et al. 
manuscript or interfere with its publication – in fact I think it makes it 
more important as it will show the problem does not require an 
understanding of an abstract theorem. I believe the Cardini et al. paper 
will show that the defects in the method are very easy to understand and 
obvious once you think about it in the right way.

F. James Rohlf, Distinguished Professor, Emeritus. Ecology & Evolution
Research Professor, Anthropology
Stony Brook University

MORPHMET may be accessed via its webpage at http://www.morphometrics.org
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to morphmet+unsubscr...@morphometrics.org.

[MORPHMET] Re: PREPRINT: between-group principal components analysis (bgPCA)

2019-05-16 Thread morphmet_moderator
This and related threads have now been locked.

Non-morphometric issues should be taken up elsewhere.

Non-morphometric posts/replies are subject to deletion from the web site.

Violators are subject to being put back on the "moderation" list.

MORPHMET may be accessed via its webpage at http://www.morphometrics.org
You received this message because you are subscribed to the Google Groups 
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to morphmet+unsubscr...@morphometrics.org.

[MORPHMET] Re: PREPRINT: between-group principal components analysis (bgPCA)

2019-05-16 Thread MORPHMET
Dear MorphMetters,

Yesterday David Polly reminded us of the existence of a 3.5-page draft 
fragment emailed by Andrea Cardini on September 5, 2018 and headed as by 
"Cardini, O'Higgins, Polly, Rohlf, Bookstein." So, yes, David Polly was 
indeed present at the creation. It begins with paragraphs of an 
introduction acknowledged to be "largely borrowed" from me and continues 
with a summary of some of Cardini's simulation results, which are 
consistent with mine but reported using mostly different simulations and 
quantifications, together with a sketch of a discussion quite different 
from mine.  This fragment, particularly its style of simulating and 
reporting, may have served later as the basis of the second of the two 
papers as I referred to it both in my Monday morphmet posting and in my 
biorxiv preprint. (I say "may have" because I have not seen any more recent 
state of that second manuscript.) So my Monday note to morphmet pointing to 
my biorxiv preprint should have mentioned the September document as part of 
the history of our project before it split into two. I apologize to David 
for not realizing the role he played in this early history, and apologize 
to Andrea for overlooking this early draft of the less algebraic half of 
our joint project.  I hope the authors of this second paper will post it to 
biorxiv as urgently as mine was posted once theirs, too, is in final draft.


On Tuesday, May 14, 2019 at 1:32:24 PM UTC-4, MORPHMET wrote:
> Dear MorphMetters,
> Some of you may have been in the auditorium in the Department of Botany, 
> University of Vienna, back in March when Philipp Mitteroecker and I were 
> the two scheduled discussants for the conference "GMAustria19" on 
> applications of geometric morphometrics.  Several of the papers delivered 
> there used between-group principal components analysis (bgPCA), and after 
> each of those papers I mentioned in the course of my commentary that bgPCA 
> was fatally flawed in applications to most GMM data sets and should NEVER 
> be used here. In my keynote address, which closed the meeting, I had one 
> cryptic slide about this assertion, with an example that flashed on the 
> screen but was immediately replaced by the next slide.
> The typical response to both my own talk and my criticism of the talks of 
> others, as far as bgPCA was concerned, was along the lines of "Hunh?" or 
> sometimes "What are you blathering about this time? Isn't bgPCA in the 
> standard toolkit?" I answered that the Bookstein paper they should read was 
> just then being written, as one of a pair jointly arising from 
> conversations with Andrea Cardini, Jim Rohlf, and Paul O'Higgins following 
> an original hunch of Cardini's, and that my argument would be pretty 
> convincing once it was actually written down.  The claim isn't that people 
> are using bgPCA incorrectly. They're using it according to the published 
> formulas, yes, but the method itself yields biological nonsense much too 
> often.
> That was March.  In April, two different articles in Nature (one by 
> Detroit et al., one by Chen et al.) buttressed claims about sister species 
> of Homo sapiens using the bgPCA method, and so suddenly it became clear 
> that we authors had to do something quickly lest this become an epidemic of 
> bad biometrics. So we accelerated our writing. My paper was the first to be 
> finished, probably because it is a single-authored item by an emeritus with 
> no other obligations, and it seemed like a good idea to upload the final 
> draft to https://www.biorxiv.org even before submitting the paper, so 
> that any letter to the editors of Nature could include a link to  the 
> argument as to exactly WHY bgPCA is nearly always unsound and its 
> inferences invalid for applications in contemporary GMM. 
> That is the draft that has just appeared as 
> https://www.biorxiv.org/content/10.1101/627448v1
> For those of you who were at the March meeting, this is the argument 
> (complete with formulas) defending my stern condemnation there. I won't try 
> to summarize it in this morphmet note -- if you're interested, just read 
> the abstract on page 1 of the link.  For those of you who have already 
> published bgPCA analyses, you know who you are -- my paper argues strongly 
> that you need to go back and revisit the inferences of those papers in a 
> mood of much more intense multivariate skepticism.  For the rest of you, 
> please consider this draft manuscript to be a wake-up call. A technique 
> that has appeared in dozens of papers and that was, alas, specifically 
> praised by Mitteroecker and Bookstein personally (back in 2011) could 
> nevertheless, when examined closely (for the first time!), turn out to be 
> algebraic garbage when applied to data sets where there are far more shape 
> coordinates than specimens. But isn't that the usual situation in GMM these 
> days?  
> As always, I welcome all responses, both positive and