Hacker News Reader

Comment by jerryhuang100 | Hacker News Reader

It's interesting to see this story on the front page at the same time with another one titled "We Should Not Accept Scientific Results That Have Not Been Repeated". It's easy to solve both problems (killing two birds with one stone?) by implementing protocols of purging scientists with track records of non-repeatable results from "personal opinions" and fraudulent scholars gaming publishing systems, including titles, positions, degrees and/or grant money ever awarded. This should at least release some vacancy in the academia (given a 11% repeatable rate from the Amgen study).

View on HN

Removing scientists whose results aren't replicated is firing by lottery. Results fail to repeat all the time.

And while you're at it, hire some of those underemployed post-docs to reproduce new results full time.

I concur this with a proposition of a system named as repeatability evaluation and digital tracking, or r.e.d.d.i.t., with the crowd-sourcing efforts.

The point of the Nautilus article was repeatability is a must before a scientific result is accepted. However pressure to publish new results often lead to scientists losing motivation in doing replicating studies. It's a bug in the scientific ecosystem, particularly in the allocation of rewards. A significant number of experiments will always fail to replicate for various random reasons. That doesn't mean the concerned scientist is at fault.

one or two publications without repeatable results might be acceptable - maybe he or she is just out of luck or some how choose the wrong hypothesis. but if there are five or six non-repeatable papers in the publication laundry list, esp some used for degrees, promotions to tenured positions, or grant applications, this makes him or her looks like fraud with a track record of publishing "personal opinions."

As a recent comment pointed out - the issue isn't solved by dividing the pie, but by increasing the size of it.

These arguments lead to one historically known outcome.

"increasing the size of (pie)"

this is relatively vague as if u do read the article, there are more and more adjunct and temp faculty positions at universities, not more tenure-track positions. besides, the pie of r&d grants in term of nih or nsf budgets did not grow much in the past decade. also i dont see there is major possibility for anothe UC system or ivy league to be duplicated in the next decade or so - it takes much longer time to build an oven for a bigger pie.

Reply on news.ycombinator.com