“When used by scientists, [trick] can mean for example a mathematical approach brought to bear to solve a problem.”Ever since this incident blew up, I am acutely aware of using that term. My fellow scientists or I use it daily, I think, in this way.
As someone else pointed out, "trick" wasn't the real scandal. It was "hide the decline".
As someone else pointed out, "trick" wasn't the real scandal. It was "hide the decline". Exactly.
With one caveat...I haven't taken the time to dig into the specifics of what the reference was on that message, beyond it having to do with the tree-ring data. There are some valid points to consider. Here are a few:
- When you run some of the many smoothing algorithms (e.g., 30-year running mean), you can't run them out into the future. Your last few decades might generate artifacts and not a true running average. It takes a "trick" to prevent the software from running into the future, else you get a downward trendline even if the last few decades are entirely upward. To leave the downward trend at the end would be the irresponsible approach.
- A failure of correlation of tree-ring results to measured temperatures does not necessarily mean the older tree-ring interpretation is wrong. The new growth is different from older rings. It would be irresponsible to include new growth with old.
- etc.
Note that this was
presentation of information that they were discussing, not the actual information. That's important, even if it doesn't justify what they did or reduce the misleading nature of it. Where I see the true problems is how the methods are not revealed clearly (which smoothing algorithm?, what assumptions?, which is proxy data and which is measurements?, etc.).