I work with a group of kids that attend Carmel Middle School, a top-tier school in Indiana. It’s supposedly one of the best in the Indianapolis area. Carmel Middle School taught their kids recently about what makes great science and science literacy. The book notes that, while simultaneously pointing out that science isn’t perfect and full of unexpected failures, what makes great science is the fact that it is many trials are needed to before a hypothesis can be accepted as true. The authors note and impress that this means that a repeated trial is a repetition of an experiment.
Then, it follows this up by telling students that this doesn’t make any experiment great, but just good. You also need other people in other labs to do the same research. It’s system of checks and balances within the sciences to make sure things were all working. It’s not a matter of distrust, but rather just a simple verification. This is called replication.
The book? Interactive Science by Pearson. It’s reviewed by dozens of people from different universities in America for accurate content.
So, why is this important?
It turns out that some scientists are concerned about having to do exactly what the science textbooks are teaching the kids, what they taught us growing up, and what is even impressed up on all of us as scientists.
Well, after years of us constituents yelling at politicians for being stupid and making horrible decisions, they listened. The GOP is pushing through the HONEST Act, which calls for transparent science so they can make some policy decisions based off of great science ─ not just good science.
What, exactly, does this Act entail? This is according to the NPR summary:
The HONEST Act says the EPA can’t take a particular action based on scientific research unless that research is “publicly available online in a manner than is sufficient for independent analysis and substantial reproduction of research results.”
Frankly, I don’t think this goes far enough. Why just the EPA? Why not with all public policy that would banks heavily on science? Things that bank on health and nutrition, civil engineering, toxic wastes, and more. This would make the promise of our science having broader impacts ─ as stated in our grant applications ─ far more applicable than a vague statement we know will never ring true.
Turns out that Adam Russell of DARPA is onto the same path of thinking I am. His project is to build what Wired has fondly referred to as a bullshit detector. The concept comes out of the problems that some fields of science have had for reproducibility and replicability and P-value hacking that have become more and more widespread.
I don’t know if it would prevent dumb politicians from making even stupider decisions, but we could hope.
Everything has two sides, though.
Thomas Burke says that this is a huge burden to make data available for outsiders. In an editorial placed in Science (pdf), Burke joined up with David Michaels to further note that not all past studies have a lot of documentation (or have lost the raw data associated with it), some have private health data on patience, and there is just not enough budget allocated for this to be fixed.
Burke’s major fear? A challenge to live up to the ideal of what science is supposed to be. A challenge to power. “To say that every study needs to have the data out there — this is code for ‘We are going challenge it — to raise issues of uncertainty and play the delay game’ that was so successfully played, unfortunately, with things like tobacco,” says Burke.
What about businesses?
Turns out a lot of businesses threw their lot in with the HONEST Act. American Chemistry Council, American Farm Bureau Federation, CO2 Coalition, and a lot more. Even the Independent Petroleum Association of America (IPAA). There may be hope for this Act to come to pass, but even if it doesn’t, maybe DARPA will succeed with it’s grading system for science.