What I find more disturbing than problems with the publishing/peer review process (of which there are many), is that even in a world as small as the beekeeping/bee research world, astonishing claims are not challenged.

What does one say about bees (controls and experimental colonies) that:

Started from packages (on foundation) March 28
Have 20 frames drawn and 14 filled by May 21
Fed 2.6kg of HFCS weekly (some doesed, some not) from July 1 to September 30
After September 20th, hives are fed plain HFCS to bring them up to at least 15 frames of stored feed
October 5th, hives are fed another 7.6 liters of HFCS with fumagillin
December 22nd, hives were fed HFCS/sugar patties

...that's a lot of feed going into hives that are fully drawn out and well established/provisioned by May 21 in New England...especially if one considers that there there is no note about honey supers, swarming, spinning out frames, etc (which would be super important if contaminated stores were at play).  This is, of course from the now famous "Harvard Study"...we often get caught up in the data and miss the obvious.  One might consider that all the HFCS used (including for the controls) was impacting the bees,  but there is too much data on HFCS feeding to make that believable...or to make the data believable without some kind of explanation.
==================================

What does one say about a study that claims:
1.  By inspecting hives and destroying any drone comb once a month, the production of drones is halted.
and
2.  That hives (both experimental and controls) inspected once a month  for drone comb didn't try to rear a single drone all season.
...this is from the Seeley small cell study.
=================================

What about a study where bees are being exposed to and tested for sensitivity to nosema...but bees that died in the incubator (even an entire cage of controls that died) were discarded from the study and not tested for nosema infection?  Might it not be important _if_ the bees that died in cages either had nosema and died of another cause, or died of nosema?
...that one is from the  "undetectable levels of imidacloprid increases susceptibility to nosema" paper from Pettis, et al.
=================================

None of these claims pass the smell test...yet they did pass peer review in one way or another (of course there are many more, these are the three that come to mind immediately).  All three of these papers are, in their own way, of a high "pedigree" (or "pedigogree")...the reputations of the researchers and/or the institutions they represent are of the highest order.

I firmly reject the claim that only scientists can properly read and understand a paper,  and I don't really think people should be using math/statistics to analyze their own data that they don't understand themselves (this isn't a criticism of people who do, I just think that there is a point of diminishing returns when we abstract data statistically, and once it is beyond the pay grade of the person making the claims I think we are on thin ice).

I think if we let the institution of science dictate to us (from within it's own systems) what is proven (publishable?, robust?) then science itself suffers.

Some things are very technical and hard to understand for the lay person....but if you can't explain a thing, you don't understand it.  If you don't understand it, I'm pretty sure you don't know it.

...but what do you do with "the rest of the paper" when you see claims like those cited above?

Two things come to mind...neither of them is a solution to anything...but perhaps some progress.

1.  Create some mechanism where papers can be corrected without the stigma of a retraction...get the record in order.  It would be nice (and probably un-human) if researchers were rewarded for fixing things after the fact without the fear of never getting funded again.

2.  WRT bee research, I've said it before on Bee-L, I'd like to see 2 sets of controlls instead of one (and not a bigger pool of controls to be averaged).  If the data between the controls isn't significantly closer that between both of the controls and the experimental colonies, it is clear that no amount of statistical analysis will show any kind of real world correlation (I'm sure that is an overstatement,  but I think it is at more risk of being understated).  I don't think researchers should be doing research where two sets of controls can't be matched with fairly simple math....what could such results possibly mean?

deknow

             ***********************************************
The BEE-L mailing list is powered by L-Soft's renowned
LISTSERV(R) list management software.  For more information, go to:
http://www.lsoft.com/LISTSERV-powered.html