One more thing on this whole business of the “trick” perpetrated in a graph in a climate-change paper. Nate Silver says this:
Actually, what you have is a scientist, Dr. Jones, talking candidly about sexing up a graph to make his conclusions more persuasive. This is not a good thing thing to do — I’d go so far as to call it unethical — and Jones deserves some of the loss of face that he will suffer. Unfortunately, this is the sort of thing that happens all the time in both academia and the private sector — have you ever looked at the graphs in the annual report of a company which had a bad year? And it seems to happen all too often on both sides of the global warming debate (I’d include some of the graphics from An Inconvenient Truth in this category, FWIW.)
But let’s be clear: Jones is talking to his colleagues about making a prettier picture out of his data, and not about manipulating the data itself. Again, I’m not trying to excuse what he did — we make a lot of charts here and 538 and make every effort to ensure that they fairly and accurately reflect the underlying data (in addition to being aesthetically appealing.) I wish everybody would abide by that standard.
Silver’s clearly a really smart guy and knows a lot about how to handle data, but boy is this wrong. A graph in a scientific paper, or for that matter a blog post, is a part of an overall argument and is designed to convey certain information and not other information. The person who makes such a graph has to make choices about what to include and what not to include, and those choices necessarily involve consideration of how the graph fits into the larger argument.
Silver seems to imagine that a graph is an objective representation of The Way Things Are, but it just isn’t. I’m not saying merely that this notion of objectivity is an ideal that can never be fully achieved: I don’t even think it’s meaningful as an ideal to strive for. Any graph you make will be selective. Deal with it.
Of course, I’m not saying that anything goes: you can’t fake your data, and you can’t be deliberately deceptive. But as far as I can tell, there’s precisely no evidence that Jones did anything like that. In his use of words like “sexing up” and “unethical,” Silver is making a serious accusation without supplying evidence. He should know better.
Jones spliced actual temp record data on temperature reconstruction without explaining in the caption that he did so – in other words those reading the graph expected it to be reconstruction only which it was not – this is clearly unethical sexing up of a graph and amounts to scientific misconduct in my book.
Can you supply evidence for this assertion? I don’t even know which Jones paper is being referred to in any of this business — nobody seems to mention the citation. Do you know?
The trick is explained here
http://www.climateaudit.org/?p=7810
Thanks. Now I understand what’s being claimed. That’s a big help.
That graph does look unclear and even misleading. I don’t think there’s necessarily anything wrong with plotting both raw data and reconstructions on the same graph, but as far as I can tell there’s no indication that this was done, so the reader doesn’t know what he’s looking at.
As data-presentation sins go, it’s far from the biggest I’ve seen, and the level of shock and hand-wringing over it seems overblown, but it’s not nothing.