"Scientists often learn more from studies that fail. But failed studies can mean career death. So instead, they’re incentivized to generate positive results they can publish."<p>Not in the "hard" sciences: my background is physics, and there are lots of papers on "failed studies", which serve to constrain the domain of applicability of some theory or other. Or, better yet, indicate new science to be found.<p>The authors note the bias in the survey: "Our survey was not a scientific poll. For one, the respondents disproportionately hailed from the biomedical and social sciences and English-speaking communities."
Science just costs too much, and scientists have paid too much to be considered scientists. The Ph.Ds that I know just want jobs that pay the bills, that pay off their loans. They do research because it is what gives them grants and enables the climb of the hierarchcy, not because they're driven by it. The system has beat out their lust of curiosity and exploration.<p>The root problem is cost. We have to make science cheaper. We have to put it back in the hands of the curious and adventurous. Science should be possible by anyone - even teenagers. If science can't be done by the young, the poor, the autodidacts, what's the point?<p>Currently, everyone is on the 'teach everyone programming' kick. U.S. states are now starting to require that everyone learns programming. But, what about science? Let's create the 'github' of science - where anyone with a hypothesis can create a notebook, gather up like-minded people to collaborate, gather data, analyze it, 'fork' others' research into new areas. That's how we will make it cheaper and accessible.
Yes, yes, yes - these are the things we ought to be talking about! In academia, in government, in society! Great to see these problems summarized here, usually one only sees an article on one of the seven.<p>It gives me hope that we <i>have</i> started talking about these things - at least in academia (as numerous think-pieces in Nature et al. testify). We need to continue this discussion, make the public aware of it, and then start taking steps to solve it. No, science is not doomed, but boy do we still have a lot of work to do to get it to where it should be...
I had hoped to read one of those occasional lists of the hard, outstanding problems scientists still need to solve, in biology, say, or physics. Instead, it's a list of meta "problems" like science funding, or poor study designs.
I think that this trend of decreasing funding for science research will only continue unless some fundamental changes are made to the system.<p>For one, the government is becoming increasingly weak while corporations are becoming increasingly powerful.<p>Unlike the government, corporations in general (and their investment strategies) are focused on short term results - That's how executive pay/bonuses are structured - CEOs don't want to invest in something that will only bear fruit in 10 years so that some future CEO will get all the credit for it. Humans are terrible at allocating credit/praise because we like to pretend that the universe is simple and that all actions have simple, predictable effects without unexpected side-effects.<p>Science research cannot exist in a corporate environment. Science can only rely on government or philanthropy.
Daunting as this list seems, of these points and fixes are of course interrelated. Even small tweaking can bring big improvements. Take open science. Improvements to funding are needed to promote long-term projects, but in absence of those kinds of grants, open science is all the more important. Sharing partial results and details of studies that might not make it into publications leaves a baton for others to pick up and make the most of their limited funding time.<p>And speaking of open science, making research public as it's happening (not even results, just what you're studying) can help prevent redundant concurrent studies in multiple labs and facilitate collaboration, also making the most of the limited funding scientists have.
This list is great! Reproducibility and verification of results is key!<p>Pachyderm (disclaimer: my company) is building infrastructure tools to help data scientists reproduce results by offering "Git for Data."
Could the "citizen science" movement help with the paywall problem?<p>What about legislation requiring that any unclassified government funded science be available outside a paywall? Last I saw a figure for it, the combined budgets of all of the US military bands was ~$270 million a year. I bet a lot of science publication could be "nationalized" for that amount.
1. Money. Economical issue.<p>2. Poor design. Competence issue.<p>3. Replication. Incentive issue.<p>4. Peer review. Bad old analog system.<p>5. Paywalls. Capitalism.<p>6. Poor communication. Competence issue.<p>7. Stress. Personal issue.<p>Nothing gets in the way of science except ourselves.
It is an obvious propaganda piece for the benefit of those who profit on the fossil fuel industry.<p>Psychology and Climate change are in different categories. The crisis of reproducibility in <i>soft</i> sciences has nothing to do with the overwhelming scientific consensus on the reality of global warming.<p>Many mentioned issues are real in hard science too but it just an example that you can lie with the truth. Perverse incentives, publication bias, imperfection of peer review, etc can't invalidate established results e.g., Newtonian physics continues to work in the domain it is applicable for.<p>It is infuriating that the planet (planetary habitability) is destroyed for the benefit of the very few.