3 ms·
Thou shalt not make a graph without error bars. In that respect, the error here belongs entirely to whoever made that graph in the first place. They plotted al
by cjensen 13y ago
Thou shalt not make a graph without error bars.
In that respect, the error here belongs entirely to whoever made that graph in the first place. They plotted all those models together to try to convince the reader that all models point to doom. But in doing so without error bars, they implied a level of confidence for specific years that was entirely absent from the models.
- spindritf 13y agoIn principle, ok, but you wouldn't be able to see anything if that graph also included error bars for 100 climate models.
- coherentpony 13y agoYou can't really see anything with 100 climate models without error bars. While your statement may be correct, the point is entirely moot.
- swombat 13y agoThe "article" relies entirely on this faulty graph, and then draws spurious conclusions... The graph by itself may be forgivable, but surrounded by this pattern of faulty thinking it is just one more piece of nonsense.
- retrogradeorbit 13y agoHow do you measure the uncertainty in predictive models? From my physics I remember that measurements have uncertainty. And then you propagate the uncertainty into derived values with differential equations. But how do you ascertain the uncertainty in a computer model? Did the original graphs actually have error bars?
- jkrems 13y agoYes, see the debunk links in here. Those models normally all have ranges. Since he's in expert, it's a reasonable assumption that he either intentionally choose a graph without errors or he intentionally removed the error bars. Don't know which one would be worse.
- khawkins 13y agoIt would be redundant to include error bars. Each path should be representing a random draw from the space of possible futures.