4 ms·

This sounds like academia. People writing papers with the only purpose to be cited (not read) and improve the currently fashionable index (h-index, whatever). P

by pqs 8y ago

This sounds like academia. People writing papers with the only purpose to be cited (not read) and improve the currently fashionable index (h-index, whatever). People presenting in conferences that are organized so everyone can add a conference in their CV. Projects that organize stakeholder meetings, which are a wast of time for everyone, because nowadays projects must organize stakeholder meetings in order to be funded.

neuronic 8y ago
This is why in other threads I have often observed people asking for a more holistic approach to evaluate scientific output. The problem, of course, is the establishment of true independence for the evaluating entity. Ideally a "score" or index should incorporate reproducibility way way way over which journal is publishing the research. Include funding sources, track record, institutions, and also reviewers! In bioinformatics, I tried reproducing models from Cell papers and others, noticing that it's pretty much impossible to get the same results with the same data (!) according to described methods. Things like arbitrarily configured sets of R-scripts look more like someone tried to push certain results over actually investigating natural phenomena with the help of software. The paramount experience is convincing authors to reveal their source code. Most are suspiciously reserved because they know they published "pretty" bullshit that journals like Cell were glad to publish, because the biology part of the paper has some "big names" on it.
- phkahler 8y ago
  >> This is why in other threads I have often observed people asking for a more holistic approach to evaluate scientific output. In many cases engineers evaluate scientific output by making things. Some results are too theoretical to be useful in the near term and that's fine, but if all the science is that abstract we have a problem. We're at the point where science is feeding into science rather than into the real world.
croon 8y ago
While we have Resume Driven Development to contend with, I'm sure there are similar practices in almost any other field.

by pqs 8y ago

neuronic 8y ago

This is why in other threads I have often observed people asking for a more holistic approach to evaluate scientific output. The problem, of course, is the establishment of true independence for the evaluating entity. Ideally a "score" or index should incorporate reproducibility way way way over which journal is publishing the research. Include funding sources, track record, institutions, and also reviewers! In bioinformatics, I tried reproducing models from Cell papers and others, noticing that it's pretty much impossible to get the same results with the same data (!) according to described methods. Things like arbitrarily configured sets of R-scripts look more like someone tried to push certain results over actually investigating natural phenomena with the help of software. The paramount experience is convincing authors to reveal their source code. Most are suspiciously reserved because they know they published "pretty" bullshit that journals like Cell were glad to publish, because the biology part of the paper has some "big names" on it.

phkahler 8y ago
>> This is why in other threads I have often observed people asking for a more holistic approach to evaluate scientific output. In many cases engineers evaluate scientific output by making things. Some results are too theoretical to be useful in the near term and that's fine, but if all the science is that abstract we have a problem. We're at the point where science is feeding into science rather than into the real world.

croon 8y ago

While we have Resume Driven Development to contend with, I'm sure there are similar practices in almost any other field.