… which can mean at least two reasonable things: (1) a particular set of measurements stopped increasing so fast; (2) the underlying process stopped or slowed. It seems clear that #2 is the more interesting of these.
I’m a big fan of the Intraocular Trauma significance test
(Interocular: between the eyes.) It makes for a good soundbite, but I don’t think it’s usually the best criterion. There’s a reason why fancier and more objective significance tests have been developed!
It seems clear that #2 is the more interesting of these
It also seems clear that we don’t have a good handle on the underlying process so claims about what it does or does not should not be expressed in plain and simple phrases.
I don’t think it’s usually the best criterion
I didn’t say it was—I said I liked it. Fancier significance tests are fancier, but also easier to trick oneself with.
claims about what it does or does not should not be expressed in plain and simple phrases
It appears that from this you draw the conclusion that any given plain and simple phrase can and should and will be clearly understood to refer to something easier to make such claims about with confidence. I draw a different conclusion: we shouldn’t make claims with plain and simple phrases that are liable to be understood in terms of things we don’t have a good handle on.
easier to trick oneself with
I am not at all convinced. It is very, very easy to trick oneself into seeing patterns that aren’t there, and they will quite often appear to hit you between the eyes. Have a look at some random noise:
These are twelve randomly generated datasets with statistics crudely resembling those of the global warming data from 1960 to 2014. None of them has any sort of hiatus in the underlying process; they’re all ramp + white noise. I’d say at least half have “hiatuses” inflicting at least as much interocular trauma as the actual global mean surface temperature graph’s “hiatus” does.
If you have MATLAB you can generate similar graphs yourself:
n=55; f=7.5; x=1:n; for i=1:12; y=(1:n)+f*randn(1,n); subplot(3,4,i); plot(x, y, 'r-'); bestj=0; bestm=1; bestk=0; for j=1:(n-14); x1=j:(j+14); y1=y(x1); c=lscov([0*x1'+1 x1'], y1'); if c(2)<bestm; bestj=j; bestm=c(2); bestk=c(1); end; end; if bestm<=0.5; x1=bestj+(0:14); hold on; plot(x1,bestk+bestm*x1,'b-','LineWidth',3); hold off; end; end;
(This only plots the 15-year trend lines when the gradient over those 15 years is ⇐ half the underlying gradient. You will notice that in my plots, every subplot has a trend line plotted. Yours probably will too.)
For a better simulation of the interocular trauma from actual climate data, I did the same as above but after finding the best “hiatus” in the 55-year data I extended the data on the left (same ramp, same-distribution white noise) to give us 55 years with that “hiatus” at the end. Here are the results:
I reckon that numbers 4,5,7,8,9,10,12 are about as impressive as the “hiatus” in the actual data. That’s just over half.
None of your plots satisfy my Interocular Trauma test (by the way, you’re right that it’s interocular, though the intraocular might be a Continental variation, coup d’oeil and all that :-D). Even the bright blue LOOK AT ME! lines don’t help.
And if we’re throwing pictures around and talking about “objective” statistical metrics, I give you the Anscombe’s quartet.
None of your plots satisfy my Interocular Trauma test
Several of them are as convincing to me as the “hiatus” in the actual temperature data.
Anscombe’s quartet
I’m familiar with Anscombe’s quartet, but what’s its relevance here? I mean, I take it you’re saying something more sensible than “Knowing a few statistics computed from a dataset may tell you far less than everything there is to know about it; therefore we should judge whether or not global warming has slowed or stopped by eyeballing the graph rather than applying any statistical tests”, but what?
Nope. I’m merely convinced that the existence of the hiatus in the measured temperatures isn’t very strong evidence of anything beyond itself. Very similar effects can be produced by noise; therefore seeing such an effect isn’t good evidence of anything more than noise. Of course it might have some more interesting cause, but if want to see better evidence to be convinced that it does.
Eh?
The trouble with merely pointing at things and saying “Behold!” rather than making an actual argument is that teen your readers need to guess what argument you’re hinting at. In this case the best guess I could come up with seemed unlikely, which is why I wrote “I take it you’re saying something more sensible than …, but what?”. Perhaps you might explain what you did have in mind?
I’m merely convinced that the existence of the hiatus in the measured temperatures isn’t very strong evidence of anything beyond itself.
So, in this thread, who are you arguing against? Did someone say “this hiatus certainly means X”?
The trouble with merely pointing at things and saying “Behold!” rather than making an actual argument
If you were to bother looking at the start of this subthread, you would have seen that the original issue was
the disagreement was just over the existence of a recent hiatus in land-ocean surface temperature warming
Questions about existence are adequately answered by merely pointing at things and saying “Behold!”
I have a feeling you are searching for an opponent who would claim something along the lines of “The hiatus is a incontrovertible proof that global warming isn’t happening” and are disappointed that such an opponent is unwilling to present himself.
who are you arguing against? [...] just over the existence of a recent hiatus in land-ocean surface temperature warming
OK, so let’s try to be really clear about this. I suggest that there are three possible claims here. GRAPH: “if you look at the temperature graph, its gradient is lower circa 2005 than circa 1990”. SIGNIFICANT: “GRAPH, and furthermore the difference in gradients is too large to be adequately explained by noise”. MECHANISM: “SIGNIFICANT, and furthermore the best explanation is that something has changed in whatever underlying warming phenomenon may have been going on”.
What were the original questions at issue? Well, in passive_fist’s comment three papers (one “pro-hiatus”, two “anti-hiatus” are cited. The first argues for SIGNIFICANT and suggests two possible explanations, one of which is MECHANISM. The second argues both against GRAPH (it claims that the data need adjusting) and against SIGNIFICANT (it points out that the reduction gets smaller if you include the latest data, including the very warm 2014, and if you don’t start at the cherry-picked El Niño year of 1998). The third argues against SIGNIFICANT on the basis that if you do the statistics right there isn’t actually evidence for a reduction in warming, and explains that the question is important because of possible implications for MECHANISM.
So it doesn’t look to me as if the question was only ever about GRAPH.
Now, perhaps you were only ever talking about GRAPH. But if so, your comments were (I’m sorry to have to say) entirely irrelevant to the points actually at issue.
I have a feeling [...]
Nope, nothing of the sort. Sorry to be less made-of-straw than you might like.
… which can mean at least two reasonable things: (1) a particular set of measurements stopped increasing so fast; (2) the underlying process stopped or slowed. It seems clear that #2 is the more interesting of these.
(Interocular: between the eyes.) It makes for a good soundbite, but I don’t think it’s usually the best criterion. There’s a reason why fancier and more objective significance tests have been developed!
It also seems clear that we don’t have a good handle on the underlying process so claims about what it does or does not should not be expressed in plain and simple phrases.
I didn’t say it was—I said I liked it. Fancier significance tests are fancier, but also easier to trick oneself with.
It appears that from this you draw the conclusion that any given plain and simple phrase can and should and will be clearly understood to refer to something easier to make such claims about with confidence. I draw a different conclusion: we shouldn’t make claims with plain and simple phrases that are liable to be understood in terms of things we don’t have a good handle on.
I am not at all convinced. It is very, very easy to trick oneself into seeing patterns that aren’t there, and they will quite often appear to hit you between the eyes. Have a look at some random noise:
These are twelve randomly generated datasets with statistics crudely resembling those of the global warming data from 1960 to 2014. None of them has any sort of hiatus in the underlying process; they’re all ramp + white noise. I’d say at least half have “hiatuses” inflicting at least as much interocular trauma as the actual global mean surface temperature graph’s “hiatus” does.
If you have MATLAB you can generate similar graphs yourself:
(This only plots the 15-year trend lines when the gradient over those 15 years is ⇐ half the underlying gradient. You will notice that in my plots, every subplot has a trend line plotted. Yours probably will too.)
For a better simulation of the interocular trauma from actual climate data, I did the same as above but after finding the best “hiatus” in the 55-year data I extended the data on the left (same ramp, same-distribution white noise) to give us 55 years with that “hiatus” at the end. Here are the results:
I reckon that numbers 4,5,7,8,9,10,12 are about as impressive as the “hiatus” in the actual data. That’s just over half.
None of your plots satisfy my Interocular Trauma test (by the way, you’re right that it’s interocular, though the intraocular might be a Continental variation, coup d’oeil and all that :-D). Even the bright blue LOOK AT ME! lines don’t help.
And if we’re throwing pictures around and talking about “objective” statistical metrics, I give you the Anscombe’s quartet.
Several of them are as convincing to me as the “hiatus” in the actual temperature data.
I’m familiar with Anscombe’s quartet, but what’s its relevance here? I mean, I take it you’re saying something more sensible than “Knowing a few statistics computed from a dataset may tell you far less than everything there is to know about it; therefore we should judge whether or not global warming has slowed or stopped by eyeballing the graph rather than applying any statistical tests”, but what?
So are you convinced that the “hiatus” is just an artifact of noise in the data?
Eh? Where is this lovely piece coming from?
Nope. I’m merely convinced that the existence of the hiatus in the measured temperatures isn’t very strong evidence of anything beyond itself. Very similar effects can be produced by noise; therefore seeing such an effect isn’t good evidence of anything more than noise. Of course it might have some more interesting cause, but if want to see better evidence to be convinced that it does.
The trouble with merely pointing at things and saying “Behold!” rather than making an actual argument is that teen your readers need to guess what argument you’re hinting at. In this case the best guess I could come up with seemed unlikely, which is why I wrote “I take it you’re saying something more sensible than …, but what?”. Perhaps you might explain what you did have in mind?
So, in this thread, who are you arguing against? Did someone say “this hiatus certainly means X”?
If you were to bother looking at the start of this subthread, you would have seen that the original issue was
Questions about existence are adequately answered by merely pointing at things and saying “Behold!”
I have a feeling you are searching for an opponent who would claim something along the lines of “The hiatus is a incontrovertible proof that global warming isn’t happening” and are disappointed that such an opponent is unwilling to present himself.
OK, so let’s try to be really clear about this. I suggest that there are three possible claims here. GRAPH: “if you look at the temperature graph, its gradient is lower circa 2005 than circa 1990”. SIGNIFICANT: “GRAPH, and furthermore the difference in gradients is too large to be adequately explained by noise”. MECHANISM: “SIGNIFICANT, and furthermore the best explanation is that something has changed in whatever underlying warming phenomenon may have been going on”.
What were the original questions at issue? Well, in passive_fist’s comment three papers (one “pro-hiatus”, two “anti-hiatus” are cited. The first argues for SIGNIFICANT and suggests two possible explanations, one of which is MECHANISM. The second argues both against GRAPH (it claims that the data need adjusting) and against SIGNIFICANT (it points out that the reduction gets smaller if you include the latest data, including the very warm 2014, and if you don’t start at the cherry-picked El Niño year of 1998). The third argues against SIGNIFICANT on the basis that if you do the statistics right there isn’t actually evidence for a reduction in warming, and explains that the question is important because of possible implications for MECHANISM.
So it doesn’t look to me as if the question was only ever about GRAPH.
Now, perhaps you were only ever talking about GRAPH. But if so, your comments were (I’m sorry to have to say) entirely irrelevant to the points actually at issue.
Nope, nothing of the sort. Sorry to be less made-of-straw than you might like.
Irrelevant to the debate you were having inside your mind, probably. Unfortunately, I was not part of it.
Do you, seriously, think you are being reasonable in this discussion?
I am the very embodiment of reasonableness, am I not? :-P
OK, tapping out now. (By the way, none of the downvotes you’ve received in this thread come from me.)