[abathur]

I think this is exacerbated by a dual blind-spot in both our thinking and languages regarding the difference between "how something was the last time we checked" and "how something intrinsically/durably/permanently is and will be no matter how many times we check". It probably isn't something we've needed in large quantities for long.

Your question focuses on psychology (and I agree it's important here), but this probably happens to greater and lesser degrees when studying any complex adaptive system.

Say you go study the carbon-uptake potential of a few species of tree, model how planting more trees could impact the climate, and click the publish button.

Your research may already be out of date. Your ability to completely control variables in a complex system has limits. Most of the environmental measurements that went into your model are moving targets. Even if your model is predictive at publication, there's a good chance there are unmodeled dependent variables lurking.

What if your research captures the zeitgeist? What if you inspire the planting of a hundred-billion trees? Do your measurements consider lone trees? Does the density and surface area forested (and eventually the weather-shifts a large forest triggers) impact how it grows and functions as a carbon sink? What if your research inspires someone to plant a massive monoculture that ultimately incubates and spreads a parasite that ultimately destroys the species?

[abathur]

To unpack a smidge, I think the first N steps all involve developing the perception and language to spot, tag, and catalog instances of this problem.

There's a fundamental shortfall in the quality of our thinking relative to the task (and the ultimate answer may just be that we're too time-bound to think this way). It's in the study designs, papers, abstracts, and unavoidable in the popular press.