False consensus

Another long and rambling article I'm afraid. Concision isn't my strong point. There's just something about the idea that scientists are closed-mined that really winds me up the wrong way.

This time I look at the idea that a scientific consensus doesn't mean anything because scientists all want to / have to agree with each other. In astronomy at least, this isn't true. There's an abundance of examples, past and present, of people publishing very non-mainstream ideas in reputable journals. One reason it might appear so is that there is a media obsession with making things exciting. This tends to happen at the slightest provocation without good evidence. It's great for generating enthusiasm, but that turns sour if astronomers have to keep saying, "nope, that's not true" all the time.

There's nothing wrong with being excited, it's just that scientists hate getting excited about things which aren't true.

On the other hand there are some scientific trends which should start warning bells ringing. Astronomy is a competitive subject made of many different small groups each trying to outdo the other. This means there's very little incentive for everyone to agree and lots of incentive to claim new, unexpected discoveries. With projects that requires hundreds of people, this rivalry is by necessity much reduced. We don't need to eliminate huge project groups by any means, but it would be a mistake to rely exclusively on large groups.

There's also a "publish or perish culture" which is potentially more serious. You cannot sensibly evaluate someone's abilities in a multifarious subject like astronomy using only the number of papers they've published. If your ranking as an astronomer depends only on this, there's a strong motivation to publish a lot of mediocre papers rather than a few good ones. Which encourages you to go for small, easy, boring projects.

Finally, the "research grant" funding system. It doesn't exactly encourage a false consensus, but it comes close. Grants typically expect a certain number of publications relating to their given topic. This is hardly a sensible way to encourage original thinking and innovation. The postdoctoral system (which is where most of the research is done) is ever-more reliant on this model of funding, and that's dangerous.

https://astrorhysy.blogspot.com/2015/10/false-consensus.html

Scientists are pirates now, and with good cause

"In many countries, it's against the law to download copyrighted material without paying for it - whether it's a music track, a movie, or an academic paper. Published research is protected by the same laws, and access is generally restricted to scientists - or institutions - who subscribe to journals. But some scientists argue that their need to access the latest knowledge justifies flouting the law, and they're using a Twitter hashtag to help pirate scientific papers."

"...Elsevier wouldn't comment on the case, but did give a statement to BBC Trending saying that they recognise that access and publishing options are key for researchers. The company says it provides open access journals, rental options, individual article purchases and other means of disseminating research papers."

They do, but open access is much more expensive for authors than the standard license (>$1000-2000). Individual article purchases are not a sensible alternative : it's rare to read a paper in its entirety, far more often you just need to know one particular measurement or conclusion. It's not expensive per paper but it very quickly adds up since most papers cites >~30 papers each (currently I'm working on one that cites 90, which would come to $2700 at $30 per article EDIT : And that doesn't even begin to count the number of papers I had to look at to determine that they weren't relevant, which at a minimum would double this).


The great thing about astronomy is that we have http://arxiv.org/ which provides free access to pretty much everything these days, regardless of journal license.
http://www.bbc.com/news/blogs-trending-34572462

Open reviewing isn't that much better than normal reviewing

I'd be wary of disclosing the identity of reviewers. There's a risk that they won't want to be seen as supporting unconventional research, thereby establishing a false consensus. However I do think that publishing the author-reviewer correspondence as well as the research itself might be a good idea.

I'm surprised that the difference in quality is so low (5% better using the alternative method), but I haven't read the original report yet.

https://www.timeshighereducation.com/news/open-peer-review-better-quality-than-traditional-process

A thousand galaxies just isn't enough

A Thousand Galaxies Discovered In The Coma Cluster... but it's still not enough

In this article linked below I cover a few different papers. At least 300 galaxies have been recently discovered in the Coma cluster which are about as large as the Milky Way, but a thousand times fainter. Hundreds more smaller galaxies have also been found there, and also in the closer Virgo cluster by another team. Hooray, more galaxies ? Everyone loves galaxies, right ?

Well, sort-of. Galaxies are definitely, objectively awesome. But cosmological models predict far more galaxies than we actually observe. The new discoveries are spectacular in that they uncover a huge population of what were unusually large, faint objects, which probably need a lot of dark matter just to survive being torn apart inside the clusters. On the other hand, the numbers of smaller galaxies being found is just nowhere near enough to save the models.

Could it be that those models are just plain wrong ? Sure. But right now, there are so many difficulties in the details of establishing exactly what the models predict, it's probably premature to throw them out just yet. The only safe conclusion is that for now, we don't really understand what's going on.

... with guest appearances by the Doctor, Godzilla, and Gimli the dwarf...

Placeholder post intended to be replaced with a better summary.

Mopra data in FRELLED

For those of you supporting / interested in supporting the Mopra telescope on Kickstarter, here's a quick flythrough of one of their data cubes. About 10% of the full survey data if I'm not mistaken.

Pluto is a planet

An attempt at a more rational justification for calling Pluto a planet...

The IAU's definition of a planet being distinct from a dwarf planet makes no sense linguistically. Call them instead "giant" or "major" planets and dwarf planets. Then "planet" just refers to any large round object that doesn't shine by nuclear fusion. Within that broad category you can have many more sub-categories of what may be very different objects. Just as "gas giant planets" are now understood to be totally different from "terrestrial planets", so "icy planets" like Pluto could be recognized as distinctly different objects.

Yes, Ceres and other objects would also become planets, but they'd still be dwarf planets. The Solar System would still only have eight major planets. Everybody wins !

Apparently I'm a journalist

Via Michael J. Coffey .

Surprisingly, I'm a journalist.

https://www.guidedtrack.com/programs/1q59zh4/run?normalpoints=18&sunkcost=0&planning=1&explanationfreeze=2&probabilistic=1&rhetorical=3&analyzer=3&timemoney=2&intuition=7&future=6&numbers=9&evidence=9&csr=6&enjoyment=0

You are Intuitive: You tend to trust your intuitions — you size up situations quickly and stick with your judgments once you’ve made them. This tendency can be useful when you need to think on your feet, or when you’re using a skill that you’ve already honed to perfection.

You are Subjective: People and stories interest you more than facts and figures do; you focus on the essence of ideas over the details. Your mind is more qualitative than quantitative. This trait lets you focus on the big picture over the nitty-gritty.

You are Carefree: You tend to live in the moment. You don’t waste a lot of emotional energy fretting about the future. Instead, you focus on getting the most out of life right now.

You are Skeptical: You treat new information and ideas with caution and skepticism. Spurious arguments rarely fool or confuse you, and your beliefs are based on foundations of hard logic. You possess a fine-tuned BS detector.

Personally I would have said almost the exact opposite about the first three.

Apparently I don't have any especially strong skills and may be vulnerable to the "sunk cost fallacy".

The Sunk Cost Fallacy is a cognitive bias that can distort your decisions about which pursuits are worth continuing and which aren't. (Like whether to finish eating an unappetizing dish that you've already paid for, for example.)

No, no, no, no, NO. That's not a fallacy, that's called bloody-mindedness, and used properly it is very far indeed from a weakness. You just have to accept that sometimes time gets wasted. That's part of the process.
I disagreed with the premise of the question that it was possible to know for certain that a project would be unsuccessful. That's a fallacy in itself.

"It appears that you have a fairly weak understanding of the way that evidence should affect your confidence in a theory."

Umm.... reaaaally ?

Interesting though.

http://www.sciencedump.com/content/explorer-attorney-or-inventor-take-rationality-test

3D multi-wavelength data in FRELLED

Proof-of-concept 3D multi-wavelength data rendering in FRELLED. So far as I know, no other FITS viewer can do this.

The blue shows neutral atomic hydrogen clouds in the Milky Way while the red shows CO (a proxy for molecular hydrogen). The HI is from the GALFA-HI survey while the CO data is from the Extended Outer Galaxy Survey. A certain Kevin Douglas was kind enough to provide data sets which had already been correctly aligned.
https://www.youtube.com/watch?v=WQojSoWnpzs

Failing to start a fight

A recent paper by Davide Punzo et al. describes the importance of 3D software in analysing astronomical data sets, particularly neutral hydrogen. They do a good job of this. However, it also contains the statement :

"A recent development is the use of the open source software Blender for visualization of astronomical data (Kent, 2013; Taylor et al., 2014), but this application is more suitable for data presentation rather than interactive data analysis."

Let's just say I vented my, err, displeasure at this rather misleading statement quite loudly and privately to many and various colleagues. It was also the only mention of my work (which has taken 3 years and runs into 11,000 lines of code) in the whole paper.

So in the astro-ph article below, I redress the balance and give a full account of FRELLED. It can already do some of things Punzo et al. want HI viewers to be able to do. And so it should : unlike the software they reviewed, FRELLED was designed to view HI data. That's what it's for.

Anyway, this is already getting a very positive response, with invites to post this on Wolfang Burg's 3D Astrophysics blog which I wasn't aware of before, but looks very nice, and to the peer-reviewed journal Astronomy and Computing which Punzo et al. published in. Even Punzo admits :
"Thank you very much for the preview of your submission to astro-ph. We think it is a fair representation of FRELLED's capabilities, and advantages and shortcomings in the context of the criteria we mentioned in our paper. "

So unlike most "comment on" astro-ph articles I've been utterly unsuccessful in provoking a fight, but maybe that's for the best. :)
http://arxiv.org/abs/1506.04621

"I'm not a scientist, but..."

Why Philosophy Matters for Science : A Worked Example

"Fox News host Chris Wallace pushed Republican presidential candidate to expand on his criticism of Pope Francis for talking about climate change.... “if he’s not a scientist, and, in fact, he does have a degree in chemistry, neither are you …So, I guess the question would be, if he shouldn’t talk about it, should you?” "

It's not often I agree with Fox "News" about anything, but I've been saying this for a while. The "I'm not a scientist" defence is fine provided you don't then express an opinion about scientific matters. You don't somehow magically become more qualified to have a scientific opinion by not being qualified. That's not how it works.

"To that Santorum essentially said that politicians have to talk about things they’re not experts in all the time so anything is fair game. ...  And Santorum pushed back that fighting action on climate change is about defending American jobs."

Yes, politicians have to talk about things they're not experts in. But you wouldn't formulate a financial strategy without consulting the bankers. Rick, you're either saying that a) you're more qualified than the experts but non-experts shouldn't talk about science, which is self-contradictory, or b) you understand the scientific consensus but just don't care about it. Which is like saying that if a team of engineers have told you a dam is about to burst and flood a town, you don't need to evacuate that town.

At this point, Rick, I see no way to avoid labelling you as an idiot.

"At one point, Wallace notes that “somewhere between 80 percent and 90 percent of scientists” who have studied the issue agree. But Santorum is having none of it, calling it a “speculative science” and saying that he doesn’t believe anyone who is so sure of their facts. “Any time you hear a scientist say the science is settled, that’s political science, not real science, because no scientists in their right mind would say ever the science is settled.”

Yes Rick, I agree you shouldn't believe anyone who says an issue is settled. But perhaps you should believe everyone if they say an issue is settled. If a single engineer says the damn will burst, then perhaps you've got a problem or maybe you've just hired an incompetent engineer. If, however, 45 out of a team of 50 engineers say the dam will burst, treating that opinion as mere speculation is a recipe for disaster.

See also : http://astrorhysy.blogspot.cz/2014/09/quack-quack.html
http://www.slate.com/blogs/the_slatest/2015/06/07/fox_host_to_santorum_if_only_scientists_can_talk_climate_change_shouldn.html