MPU: HOW to eliminate ?

Alexander Berezin berezin at MCMAIL.CIS.MCMASTER.CA
Thu Jan 25 16:10:13 EST 1996

On 25 Jan 1996, Gregory R. Harriman wrote:

[ snip ]
>      To give you one example of an idea previously proposed by Bill Tivol
> which I agree with wholeheartedly:
> Bill Tivol previously wrote:
>      "If the quality of publications were the foundation for career
> advancement and funding, the motivation would be to publish fewer papers,
> each of which would be more complete--APR or no APR.  I have to agree that
> some quality control is necessary, but if the system rewarded quality
> instead of quantity, the QC problem would be more easily solved."
>      We've probably all known people who are very adept at publishing
> papers which contain the minimal publishable unit "MPU".  Like an assembly
> line, they churn out one paper after another.  Each one contributes
> marginally to what was already known.  Yet, they develop an impressively
> long, if not substantive, CV.  If the focus was on quality rather than
> quantity then the inevitable result will be fewer bad papers being
> published just to add to the length of the CV.  The subtle pressures that
> govern how scientists behave would be changed towards more constructive
> behavior.  This seems like a sensible and more easily accomplished first
> step.
> Greg Harriman


Of course, this is all fine, I agree with everything. 
And this is what I actually meant when said that 90 % of 
all papers are redundant and unnnessary, even if I 
disuse calling them 'garbage' or 'crap' (I do).

The critical point though is how you propose to 
redesign the reward system that publishing many MPU
papers will bring you liability rather than merits.

I personally believe that there are natural limitations
of how much scientist can publish (I mean, quality
stuff) per year. Taking into account that the any good 
work requires a long period of thinking, reflection,
comprehension, etc. I don't believe that normally
GOOD scientist can responsibly produce more t
han 4-5 papers per year.

(this, in fact, is supported by CV of many great 
scientists, most of whom amount to just 100+ lifetime 
papers; Richard Feymnan has about 130 for 40 years 

And 20+ papers per year (many now have these figures !) 
is very much likely excessive, diluted, redundant. There
should be NEGATIVE merit attached to it. Perhaps 
inverted U-curve (instead of linear) is what should
be used for funding (in 'fund researchers, not
proposals' scheme), with max around 4-5 papers
per year.

Alex Berezin 

More information about the Bioforum mailing list