Hidden Data

John Cook, his co-authors and the University of Queensland have repeatedly claimed all data of any scientific value for the Cook et al consensus paper was released in 2013. This claim is undeniably false. Anyone who has examined what data was released would know it is false.

The authors have released data in the form of Article ID#, title, abstract, authors, journal, publication year, pre-reconciliation endorsement/category rating, post-reconciliation endorsement/category rating and final endorsement/category rating (after "tiebreaks").

That sounds impressive. However, The Consensus Project examined data for 12,465 papers. That data was released for only 11,944 papers. This can be seen in the official data file hosted with the paper, as well as data files posted at Skeptical Science. Each of these contains 11,944 entries, not 12,465. That means data for 521 papers is missing from them.

Some of that data is available from a searachable database posted on Skeptical Science. That database provdies title, abstract, authors, journal, publication year and final endorsement/category ratings. It does not provide Article ID#, pre-reconciliation endorsement/category ratings or post-reconciliation endorsement/category ratings. In other words, it does not provide any information about the individual ratings which led to the results.

Moreover, while the page for it claims to have data for "12,464 papers" (strangely different than the 12,465 listed in the paper), searching it shows it actually contains data for only 12,280 papers. Data for 185 papers is absent from it. Data for those papers is absent from everywhere else as well. It has never been published.

In fact, a person seeking to examine the data for this paper would find it impossible to discern what 185 papers were excluded. There is no listing available anywhere. One could try performing the same search Cook and associates performed, but those results change over time. There's no way to know the results gotten today would be the same as they got for their paper.

There are two possible explanations for the position adopted by Cook, his co-authors and the University of Queensland that don't involve outright lying: 1) They simply never looked to see what data was published; 2) They don't consider the identity of papers analyzed for this project to be of scientific value.


Incidentally, Cook and associates now say:
Timestamps for the ratings were not collected, and the information would be irrelevant.
This is a strange comment as Cook previously told a University of Queensland representative:
ERL said I didnít have to include time stamp info but Iím probably going to anyway, just to show Tolís fatigue theory is all rubbish.
And I have found datestamps for all the ratings published in the data files.