Public Consequences of Sloppy Science

Unless you’ve been comatose, camping in the Everglades, hiking the South Georgia Islands or lying wasted in a tiki hut outside Cabo Wabo, this shouldn’t be news by now: James Hansen’s NASA Goddard Institute for Space Studies (GISS) has been revealed to have used a subset of bad (temporally displaced September) thermal data in averaging worldwide surface temperatures for October.

Oops? Sure, we’ve all made mistakes in science, at some level. But errors have degrees of importance, and given its context, this was a major whopper. While this blunder doesn’t necessarily discredit or debunk previous work, as the most zealous of skeptics would desire, it does raise doubts. Any such basic, preventable quality control error does call into question their oversight, and as such, their devotion to accuracy in analysis — past as well as present and future. Because what they do is so high in profile, this error has and will garner adverse media attention.

This is where it goes from “oops” to “oh sh|t.” Whether or not in the public eye, time-checking one’s observational input data is so fundamental to analytic integrity that neglecting to do so is a scientific blunder of the highest order. This applies whatever the data source. It is the responsibility of the scientist(s) in charge of the analysis to verify its veracity.

Most of the time, the scientists, reviewers or oversight body notices such an error, the data is corrected and re-analyzed, and life goes on, the humiliation limited to a small and often forgiving scientific circle of as little as one member. However, for “leading scientists” in a highly visible, heavily scrutinized and controversial specialty to fail to QC new data before analyzing, posting and promoting it in public…absolutely inexcusable! If outright falsification is the scientific equivalent of murder (e.g., the disbarred Korean geneticist), failure to QC your data (especially for something so simple as time range) is tantamount to scientific negligent homicide.

There is a lesson here. If you “don’t have the resources” to QC your results, don’t post them until you do! When it comes time to publish actual papers in the future, I do hope for their sake that the formal review process does its job with utmost rigor, and any such blunders are caught. This is one reason that “Let it publish before letting it public” is a good idea. As a fellow scientist of the atmosphere, I recommend Dr. Hansen and crew do just that, or risk further embarrassment, skepticism (warranted or not) and damage to credibility.



Comments

Leave a Reply

You must be logged in to post a comment.