Friday, 20 November 2015

Where is the love? How IFLS is failing Science.

Whoo-whoo! Science is cool now! Big Bang Theory! Brian Cox! Neil DeGrasse Tyson! I Fuckin' Love Science Facebook page!

As of 18, November 2015, the page had almost 23,000,000 likes. That's over twenty million people receiving daily science stories and updates. Surely as a lifelong geek/nerd/square (do people still say square?) I'm over the moon about this? I would if what was on offer was "science". Unfortunately much of what IFLS offers isn't science, its pseudo-science, surveys conducted without adequate controls and blatant sensationalism.

Let's take a look at the some of the stories from 18/11/15, one day in the history of the page, to see how what was shared represents science.

Surveys and sensationalism

Follow the links and what you will find is a report about a study conducted by the Journal of Family and Marriage, what IFLS don't point out is that in no way could this research constitute anything conclusive and certainly its not a metric to judge the potential success of a relationship. See if you can spot why in this description of the study from IFLS site.
"For the study, 376 dating couples in their mid-20's were queried each month for nine months "
It should be immediately obvious that this is an extremely small sample size, not only this but there is little variation within that sample. All the couples are in their mid twenties, any data extracted from them can't be extrapolated to couples outside this extremely limited age bracket.  It simply isn't representative of the population as a whole! To be honest I could stop here: this is a killer for ANY research of this nature. Also the survey is extremely short-term. Reading on further flaws present themselves:

"Their comments – such as “we were fighting a lot” and “we have more in common than I thought” – and types of bonding (passionate versus friendship-based love, for example) were converted into numerical values and graphed."

 Whoa! These are QUALITATIVE statements, the assignment of a numerical value to them is completely arbitrary. It isn't even ranked data such as "how happy are you in your relationship" scaled between 1-5 perhaps. Therefore any visual representation of this data isn't meaningful!

This isn't science. It is pretty pictures.

Again this is a link that essentially hinges on a survey. The article actually cites the Daily Mail  as a source, which for anyone with a passing interest in science in the media, should ring immediate alarm bells. To say the mail's science reporting is lack-lustre would be an understatement.

The research does come from a considerably larger sample size at 2000, over a larger age range. But it was conducted by an agency that specializes in finding drivers parking spaces, its a commercial survey based on a browser game. Its a  marketing tool by a new company to get people to their site.

 Plus apparently I have the reaction speed of an 85-year-old, possibly because I was watching the TV as I took part in the test. I wasn't asked to input any other data such as my actual age or how much wine I'd imbibed. Were any of the 2000 "subjects?" Who knows because there isn't an actual research paper one can read to assess the methodology used.

This isn't science!

This article links to a press release rather than the actual paper which is a bit annoying. Reading the actual article brings to mind some of my complaints about the previous two "study based" pieces we've had... the final two paragraphs solidifies my suspicion:
"The experiment, unpublished but under review, does have its obvious limitations..."
So this isn't even published! I wonder why? Couldn't be because isn't not methodologically sound perhaps? One of the lead researchers concedes exactly that...
"That said, the findings aren’t necessarily generalizable, Hugh-Jones admits to IFLScience, because the participants were members of online marketing panels, rather than a random sample."
I'd go further. The findings are pretty much meaningless. These online panels gather data from individuals who are often incentivised to take part. in-game currency, for example, being one of the most common rewards. Do you really think that Candy Crush addicts are taking this online research seriously? No, they rush to completion in order to collect their coins.

Therefore, the headline of the article is pure sensationalism.

The worst thing about the reporting of these studies is that they are presented as legitimate research and as such they give a very misleading impression of what scientists do and how data is gathered and presented.

In terms of science reporting, they put IFLS on par with the British tabloids, and that is not a flattering comparison. In fact its a damn scary one when you consider that IFLS may well be the most viewed science news source in the world.

If you want to read more about these kinds of surveys and market research I highly recommend Michael Marsh's BadPr site.

If you had any doubt what IFLS prime motivation is, take a glance at the next two stories from 18/11/15

Clickbait ahoy!

The link leads to pretty much nothing. Neil DeGrasse Tyson speculates on the question posed above. The overwhelming question for us is "who fuckin' cares?" This is click bait fluff and nothing more, it is not science. There's no research behind this. Let's face it sex sells. It also generates a lot of website views. So does Tyson, who is the darling of IFLS memes and posts.

You know what else sells... violence.... and Ronda Rousey.

There's no doubt that Rousey's defeat to Holly Holm was big news and guaranteed hits for any website that featured a story regarding it, and the linked article from Wired is OK. Rhett Allain does a decent job of estimating the force in the kick. His calculations are, as the author concedes, based on fairly massive assumptions, actually wild guesses, though.

My problem with IFLS piece about the article is that it adds nothing to the original and gives the distinct impression that the IFLS author may not of actually understood it and may be taking credit for something that Allain himself did:
"Allain has managed to calculate some pretty interesting stats. For example, he estimated that her average foot speed was 6.3 meters per second (14 miles per hour). From the rest of his work, we were able to add that the recoil of her foot was around 1.6 meters per second (3.6 miles per hour)."
 Wait "we were able to add that the recoil of her foot was around 1.6 meters per second (3.6 miles per hour)" Is that, we as in IFLS? Its hardly difficult! Speed is just the magnitude of velocity, its very simple to work out:

Taking Allain's values for the velocity vectors (-1.26,-1.04,0) m/s

The speed |v| is simple to find:
That's it. IFLS science can't take credit for that, most high school students could perform the same calculation with ease. They'd probably also notice that IFLS don't even give their answer to the appropriate number of significant figures.

There's no escaping reality here. This is an article about an article. And the main article isn't cutting edge, it is certainly not research, its an interesting application of some physics principles, sure. But it's not news.

Of course these weren't the only stories published by IFLS that day, they were probably the worst too. Some of the others were definitely of more merit, but why the filler? The authors of IFLS must know that the above articles are crap, or nonsense or wholly derivative of other articles. 

This isn't an off day for IFLS either, unfortunately. Similar sloppiness infects their history.

Plagiarism accusations

In 2013 Alex Wild, a Scientific American blogger and entomologist took umbrage when IFLS used one of his photos both without credit and permission. Wild had this to say in response, about IFLS and its creator Elise Andrew:
"The people who actually made the content that drives Andrew’s ascendant business- the professional scientific illustrators, the photographers, the cartoonists, the graphic designers- aren’t given anything in return. They aren’t paid. They aren’t acknowledged. They aren’t even asked."
In an article for Scientific, American Wild examined IFLS content and found that 59 of the page's last 100 posts were uncredited and that only 26 of the credited posts correctly cited the author. Further, to that, a Facebook staffer told another creator who raised a similar complaint, that as of 2013 there were over 6000 copyright complaints pending against IFLS.

Andrew herself denies this allegation.

This leads me back to a point I made earlier. Why did IFLS publish an article about a Wired article? Why didn't they just link to the Wired piece. One can only conclude that it is in order to garner hits for their own site. They used some else's work to gain visits to their site.

One could also make the point about the Mail article from earlier, but that would require me actually suggesting that someone link to the Mail for a purpose other than mockery and derision.

You're Getting warmer.... Warmer... Colder.... 

On July 13th this year IFLS published an article with the following title: Thanks To Reduced Solar Activity, We Could Be Heading For A Mini Ice Age In 2030, which was met with a slew of criticism. The main failing of the piece stemmed from the fact that author, Caroline Read, fundamentally misunderstood the study she was reporting on and failed to do further fact checking.
"...A recent announcement from solar scientists suggests that the Sun may soon enter a period of significant reduced activity, possibly causing a mini ice age by 2030 – just 15 years from now. These predictions were announced at the National Astronomy Meeting in Llandudno, Wales, so it hasn't been possible to evaluate the research yet. However, Professor Valentina Zharkova from the University of Northumbria who made this announcement claims that the findings come from a computer model of sunspots that has made "unprecedentedly accurate predictions," as reported in The Telegraph."
 The actual report centered on a drop in solar magnetic activity, this reduced magnetic activity having a barely negligible effect on the energy output, and thus Earth's temperature. The majority of climate scientists and solar scientists rejected this claim of causality out right.

Despite heavy criticism IFLS dug itself a deeper hole the following day with a piece, again by Read, entitled There Probably Won't Be A “Mini Ice Age” In 15 Years in which Zharkova states
“We didn't mention anything about the weather change, but I would have to agree that possibly you can expect it... It will be cold, but it will not be this ice age when everything is freezing like in the Hollywood films,” 
Nowhere in this second article does Read attempt to say "we were wrong, sorry about that" nor does she link to any of the legitimate criticism of the first article. A bit chilly does not equate to an ice-age, the first article was clearly sensationalism unchecked.

Astronomer  Phil Plait offered one of the most solid rebuttals of the claims made by IFLS and Zharkova in his Slate column also from the 14th July. Hitting on most of the points made by Zharkova in the second piece and dismissing them he concludes:
"Some climate change deniers have been claiming for a while that the lower number of sunspots can lead to a repeat of the brutal cold snap that gripped Europe in the 17th and 18th centuries, but that’s silly. There were several factors that came into play there (including huge volcanic eruptions that magnified cooling; volcanic gases reflect sunlight and cool the Earth a wee bit), and even then the effects were localized to Europe. And even then, summers were normal; it was just that winters were extra cold.And again, all of this is a drop in the bucket. Any cooling effects by the Sun would be on top of a much larger heating trend due to global warming."
 A further article on the subject on IFLS takes the absolute piss frankly. Jim Wild writes on 15th July:
"Wouldn’t it be great if scientists could make their minds up? One minute they’re telling us our planet is warming up due to human activity and we run the risk of potentially devastating environmental change. Next, they’re warning that the Earth is heading for a mini ice age in the next 15 years."
Except Jim "scientists" didn't warn that "Earth is heading for a mini ice age" IFLS did! The scientist in question corrected them, albeit half-heartedly. But wait a darn second... this article wasn't written for IFLS! It was actually wrote for the Conversation! Again IFLS have duplicated an article rather than just linking to the source. And in perhaps the most bare-faced way imaginable, shrugging their shoulders and walking away from a mess they generated!

Its clear that IFLS has lost its way. One would be hard pressed to deny that the page at least started as an admirable attempt to take real science to the masses. somewhere along the way the page seems to have lost its focus however. Articles are pushed through with little to no fact checking, and with an emphasis on hits rather than legitimate science correspondence. And yes we've avoided it thus far... money. IFLS is a business now with advertisers to consider and tee-shirts to sell. This has muddied the waters of what the page set out to do.

The contributors to IFLS and its creator must quickly remember that being one of the largest providers of science communication comes with a massive responsibility, one they are not fore-filling with click-bait, poor research, and sensationalism.