Hacker News Reader

Hacker News

Comment by ars | Hacker News Reader

Full thread

ars

Um, the real graph looks just as stagnant as the Daily Mail one.

The difference between them is barely even there.

Is anyone actually looking at the graphs on the blog, or are you just reading the title?

View on HN

This. Very odd.

The real problem is that they are showing a tiny period of time. Global warming levels out a bit now and then and then keeps on going. e.g. http://en.wikipedia.org/wiki/File:Global_Temperature_Anomaly... around 1965-1975 it didn't really go up much. And a graph of 1935-1945 might make you think we'd conquered global warming forever and that we could all get on with building the global socialist utopia or whatever. Sadly long term trends in complex systems aren't nearly that easy to form conclusions about.

Yes. Although it wouldn't surprise me that the Daily Mail altered the figures, it doesn't change much of anything to the overall shape of the graph. Therefore it's probably safe to attribute to stupidity rather than malice.

The article is absolutely missing the point.

Did you read the first couple of links in the article? On the day the original publication from DM came, we already debunked it, pointing to the longer time period: http://sargasso.nl/tweespalt-opwarming-plateau-of-niet/ (in Dutch sorry).

And we refer to others who did similar things.

But this specific article is about adjusting the data before presenting it. The first point (missing 8 months) is "understandable". The second point, changing the data at the beginning and end of the graph, is lying on purpose. That's unacceptable. Even though these are small adjustments (about 0,03 degrees), they are done in purpose and to reach a goal (make more people doubt).

I must confess I did not (especially as I don't read Dutch). But I still fail to see how a change of 0.03 degree should be a deliberate attempt at misleading people. I assume that the Daily Mail staff, while not necessarily embarassed by ethics, have at least a modicum of common sense.

It's done at both side of the graph, so it makes 0.06 :-) But that's not the point. It's the principle. If you present a graph based on actual data (verified by many scientist) and you change the data however slightly to fit your point of view, that's malice. Why do it otherwise?

Maybe not enough for this though: http://www.pcc.org.uk/cases/adjudicated.html?article=Nzg5Nw=...

I think what people miss is the fact that many scientists considered a 16 year lull in warming as very unlikely. Not impossible, but unlikely. It doesn't mean they have it all wrong or that it's just a hoax, and says nothing about the long term trend, but it does mean that they have to probably evolve their models and predictions to cater for natural short term variability better.

I find it very entertaining these accusations of cherry picking data for a cause, and then get countered with a different shape of cherry picking. For example proving a trend by using only land based surface temperatures, or only the past 30 years, or only the Northern Hemisphere, or neglecting the warming periods over the past 2000 years, or excluding the uncertainties of past reconstructions. It's a highly polarised debate so there is a lot of bias on both sides of the argument.

DM left out the first 8 months of 1997. Intuitively, it's fair that January is colder than September, but HADCRUT4 is temperature anomalies, so that's probably what went wrong.

That said, I think the article cherrypicks some (it seems to me) very minor issues, adds in an appeal to their own authority, while neglecting to explain why the conclusion that there "was no discernible rise in aggregate global temperatures [in the last 16 years]" is wrong. You can cherrypick any value within one year of the beginning and end of the graph that will show anywhere between ~0.3° rise and ~5.5° fall.

The 1997 slopes are repeated at 3 other times in the graph. Cutting them out makes little difference to the overall conclusion which is that this graph shows random variation and no long term trend.

If the blog author wanted to make some kind of point he should have included MUCH longer graphs, not just half a year.

> If the blog author wanted to make some kind of point he should have included MUCH longer graphs, not just half a year.

They do, in the "looked like" link of the second paragraph (http://sargasso.nl/wp-content/uploads/2012/10/hadcrut4_a.jpg). But that doesn't change the fact that they're beating on DM for misrepresenting the data, rather than making the argument that 16 years to too short.

You're missing the point: The graph is the same for the period of time which the editor selected. But he carefully selected the left (past) cutoff point to be on the same level as the right cutoff point. But in the original data, there's a steep incline at the beginning of the graph.

Now I can't vouch for the data or that the original data doesn't have a steep drop-off just a little before the snippet we're shown, but this is a prime example how choosing the sample properly changes the result dramatically.

That "steep incline at the beginning of the graph" isn't big enough, or over a large amount of time to be worth caring about.

Look at the charts. They're stagnant.

You're missing my point. I trust neither of the two - all I say is that given the numbers you can always choose a timeframe that supports your theory better than the opponents theory. So be wary of the numbers.

Reply on news.ycombinator.com