In this detailed article about academic plagiarism are some interesting details about how to do data forensics on Excel files. It really needs the graphics to understand, so see the description at the link.
(And, yes, an author of a paper on dishonesty is being accused of dishonesty. There’s more evidence.)
EDITED TO ADD (7/13): Guardian article.
Clive Robinson • June 26, 2023 12:21 PM
@ ALL,
The UK Guardian Newspaper had a piece yesterday,
https://www.theguardian.com/education/2023/jun/25/harvard-professor-data-fraud
Winter • June 26, 2023 12:58 PM
The question is often asked why these people did not produce convincing fake data? The theory and practice is not that difficult and you can easily find out how to do it right.
The answer I hear from everyone is that if you do that much work and effort, you could more easily do the experiments. And if you delve into it you realize early on that the risks of being discovered is increasing with every step.
My impression is that the fraudsters think no one will ever check.
But every paper is there forever. Someone will check them some time in the future. And that probability increases with the strength of the paper’s claims.
As many German ministers and politicians have found out, even a little scientific fraud can haunt you decades later when you yourself have long forgotten it. And scientists and bankers do not forgive fraud. And I am not so sure about bankers.
Steve • June 26, 2023 1:20 PM
I happened to see the same posting yesterday and was amazed at how incredibly inept the (alleged) fraudulent manipulation seems to have been.
Having lived on the fringes of academia for most of my working career, I have frequently been astonished at the number of absolute dunderheads who seem to get awarded PhDs and faculty positions.
Clive Robinson • June 26, 2023 1:24 PM
@ All,
A quick DuckDuck gives the reseachers website the home page of which has this snippet,
“the author, most recently, of “Rebel Talent: Why it Pays to Break the Rules in Work and Life.“
Yes that strange noise you hear is my hollow ironic laugh…
Clive Robinson • June 26, 2023 1:28 PM
@ Winter,
“And scientists and bankers do not forgive fraud. And I am not so sure about bankers.”
Neither do insurance companies…
modem phonemes • June 26, 2023 1:44 PM
Obligatory (at least for statistics amateurs) –
Does a Benford analysis apply here ?
Phillip • June 26, 2023 2:33 PM
All: I had read another article about this case; the practice is named “P-Hacking” or something like it.
Fred Bush • June 26, 2023 3:09 PM
More than one author! Multiple studies from that same paper, with different authors collecting the data, have now been implicated in fraud.
The other co-author who’s been called out, so far, is Dan Ariely, author of The Honest Truth About Dishonesty.
@Winter:
@All:
“My impression is that the fraudsters think no one will ever check.”
But they do check.
I did a simple lookup on Wikipedia which gives her husbands name.
A search on both their names pulls up as the second item, this document from the Cambridge Historical Commission,
https://www.cambridgema.gov/-/media/Files/historicalcommission/pdf/chcmeetingfiles/D1557_memo.pdf
Now knowing that both Husband and Wife were commiting fraud in their proffessional lives and that most people that think they are smarter than others think they can get away with fraud in other asspects of their life.
I suspect that an arson investigator on reading the document in light of what is now known would treat the 1-2million loss of the first house with deep suspicion.
Made worse by the fact they want to level abother historical building and rebuild with out of character for the historical area replacment of the adjacent property, would suggest that the first loss by fire may not have been in any way an accident.
Thus we may see the pair of them having to find accessories for silver bracelets.
Rombobjörn • June 26, 2023 5:30 PM
@Winter:
And scientists and bankers do not forgive fraud. And I am not so sure about bankers.
Bankers forgive fraud every day. They’re quite happy to let fraudsters siphon money out of the payment card system as long as they can cover the cost by raising their fees. We all effectively pay taxes to the carding industry through payment card fees. The card system couldn’t exist in its current fundamentally flawed design if bankers were unforgiving about fraud.
Ted • June 26, 2023 6:44 PM
Hmm. Fascinating. The forensics make a lot more sense when you play around with an Excel file (.xlxs).
The Data Colada team notes that Excel files are actually Zip files. You can unbundle them by changing the file extension from .xlxs to .zip.
There is a file in the bundle called calcChain.xml that provides a historical log of when formulas were added to a worksheet.
Really this makes a lot more sense when you play around with it. Here is a calcChain.xml excerpt from an Excel file I was playing around with:
<c r=”C7″ i=”1″ l=”1″/>< c r=”C8″ i=”1″/>< c r=”C5″ i=”1″/></calcChain>
In this example, I added the first formula in box C5. Then I added a formula in C6, and then C7. I cut and pasted the formula from C6 to C8. There is a very good explanation in the first article as to how you can tell what original data may have been moved.
lurker • June 26, 2023 9:13 PM
@Ted
Yes, the gubbins inside a .xlsx bundle are very handy for forensics, starting from Author and PathTo, on upwards. The bloat of a .docx vs. the old .doc includes the infinite undo edit history. But as they show in the Datacolada Pt.3 even a simple .csv file reveals its dirty secrets when you know what you’re looking for.
Winter • June 27, 2023 5:22 AM
@Rombobjörn
Bankers forgive fraud every day.
From clients. Not exactly when you are a banker yourself caught with your hand in the till. At least, that was my impression.
Winter • June 27, 2023 5:46 AM
@Rombobjörn
Bankers forgive fraud every day
That banks have, indeed, few problems with fraud as long as it is clients who are defrauded is also obvious from my other comment:
‘https://www.schneier.com/blog/archives/2023/06/friday-squid-blogging-giggling-squid.html/#comment-423484
Ulf • June 27, 2023 6:01 AM
Just to make it explicit what others have hinted at: while some of the points concern the underlying data (and are thus unrelated to the file format being used to store them), only part 1 makes use of Excel features, and specifically features of XLSX that are not in XLS files. So the forensics are either done on XLSX files, or are independent of the file format (similar to what @lurker said about DOCX vs DOX).
A.C. • June 27, 2023 3:58 PM
This is not p-hacking. P-hacking is when you collect a bunch of data then squint at them until you find a significant correlation. This is fabrication. They knew what they wanted the data to say and changed values until the data said it.
WhizzMan • June 28, 2023 3:09 AM
Bankers forgive if it costs more to act than to forgive. Bankers just do the calculation and if forgiveness is more profitable, that is what they will choose.
Phillip • June 30, 2023 11:15 AM
@A.C.
I looked into it and saw the distinction, thank you. Possibly, in this case, the actor was “too lazy for P-Hacking” and decided upon some Expert-Texpert macro hacking.
Mended • July 4, 2023 6:06 PM
Genuinely fascinating. Who knew that Excel stores this kind of information? Must be really useful for forensic accountants.
@anon, June 26, 2023 5:05 PM, that’s very, very interesting, but I was able to find online a lawsuit regarding the 2009 fire which names the then-owner of the building, and it’s not either this paper’s author or her husband.
Subscribe to comments on this entry
Sidebar photo of Bruce Schneier by Joe MacInnis.
Clive Robinson • June 26, 2023 12:10 PM
@ Bruce,
Such is the way of the world where,
“The best form of defence is offence”
And,
“The bigger the lie, the more it becomes the truth”
As the old saying has it,
“Fling enough mud and some of it will stick, where is anybodies guess”.
The only advice that can be given is,
1, Duck and cover.
2, Become a champion mud slinger.
As you and I both know having people take your work without credit or acknowledgment can be anoying. But with age comes the view that,
“Imitation Is the Sincerest Form of Flattery.”
After all if an idea does not have real value why steal it…
But the thing is as I’ve said before I don’t mind people using my ideas as long as they,
1, Give me a hat tip.
2, Buy you two drinks.
But hey some are to tight to even do that. Thus I hope,
“The cow-bird of doom, over flies them and…” 😉