As a researcher myself, I try to keep a good record of everything I do. This usually works for a period of time (generally until submission!), though notes are invariably less than 100% complete (even if it’s 99.9%). Some very honest and thorough researchers are not as good as they should be at record-keeping… it’s far too much like admin.

And yes, files tend to replicate themselves in a dozen different places, but in my experience this happens for similar versions of code, with no way of telling where the file you actually used is. Also, labs get periodically cleared out. Old formats (7″ floppies, DAT, microfiche) get archived for a while, then disposed of. Sometimes they’re backed up by a student /technician onto a general “Old” disk somewhere, but existing doesn’t make people aware of it, or where it is.

As #41 says, the information required to reproduce the work (though the data would need to be unearthed in this case) should be in the paper. If it’s not, then that’s the fault of both referees and the journal editor. If there’s insufficient information to reproduce the work, then there should be no problems in getting a fully open-source attempt to reproduce it published (as #56 rightly states).

Maybe that’s the way to go?

Get the new data and analze it using the techniques originally used by Jones. This will provide an out-of-sample test of Jones’ results.

It the resuls are different, then Jones 1990 is obsolete and the paleoclimate community will, of course, hail the new results as cutting edge, and demand that to move on from Jones’ antiquated work.

In an empirical sense, doesn’t this mean the data you are looking for is unimportant and uninteresting? Or perhaps I should say the author, empirically, found his own work unimportant and uninteresting?

There is an underlying underlying assumption here that data decays or evaporates spontaneously unless it is carefully sheparded. In my experience the problem with data storage is just the opposite. Data records are hardy and resistant – rather like an antrax spore. If data is in fact gathered and stored, office workers will instinctively preserve it. The originals will get filed and duplicates will be created. If you need to get rid of it you must expend some effort.

That’s right, even if the *papers* were reviewed, we are assured by the University that the data itself is not and never has been available to anyone other than Dr Jones.

(emphasis added.)

I am gobsmacked. Completely gobsmacked. One person’s say-so & officials require millions of people to spend heaven only knows how much — ?! This well beyond funny. How much do people outside ClimateAudit know? Shouldn’t they be alerted?

Now, backups can’t be found. Time to delete all research based on that data, since IT WAS NEVER AUDITED OR PEER REVIEWED. That’s right, even if the *papers* were reviewed, we are assured by the University that the data itself is not and never has been available to anyone other than Dr Jones.

You state the perfect answer to the “peer reviewed” mantra of climate researchers. The answer is: THE DATA WAS NEVER AUDITED OR PEER REVIEWED. You point out the distinction between “peer reviewed papers” and “peer reviewed data”.

Folks at ClimateAudit are leaders in exposing the lack of auditing and peer review of data used in benchmark research papers in climate studies. Folks here have made similar statements, and I think your statement is one of the best.

In trying to explain to the public, to the media, and to other scientists the fatal flaw in the foundation of climate research, this type of short clear statement is ideal for cutting through the fog of antropogenic hot air (AHA).

In 1990, I worked for a scientific/demographic research organization. We were jazzed to have a bookshelf with 30 10mb portable cartridge disks — 300+MB of data on a shelf!!!

How much could we store in that? An unbelievable amount of data (for the time). Textual measurement data is incredibly compressible. That’s how, in 1990, we were able to fit a detailed street address map of the entire USA on a single CD.

This situation is quite simple actually:

* Data were not properly archived

* Therefore, a less-than-adequate amount of backup copies were made (Dr Jones, if keeping the data only on his personal machine, was putting his entire career at risk, whether he knew that or not!)

* Now, backups can’t be found. Time to delete all research based on that data, since IT WAS NEVER AUDITED OR PEER REVIEWED. That’s right, even if the *papers* were reviewed, we are assured by the University that the data itself is not and never has been available to anyone other than Dr Jones.

My only positive hope and suggestions, seriously given:

— Run a Google Desktop search of all computers in Dr Jones’ posession. Perhaps the data is hiding in an obscure place

— Use Restorer2000 to search the “deleted” areas of all disk drives that are now or ever have been used by Dr Jones

If these attempts to recover the data fail, then the research results should be withdrawn. Science abhors non-repeatable “research”.