Another peer-reviewed paper for you. “Tickled to Death: Analysing Public Perceptions of ‘Cute’ Videos of Threatened Species (Slow Lorises — Nycticebus spp.) on Web 2.0 Sites” by K. Anne-Isola Nekaris and others.
Seems the authors surfed over to YouTube and watched a video of some guy tickling a slow loris. Like this one:
“Cute”, said many in comments. “I want one,” said others, overwhelmed by the cuteness but, as is usual in the English language, saying this as a figure of speech. Meaning they didn’t really intend to acquire the wee beastie, but used the phrase to indicate they thought it was adorable.
This perplexed and shocked Nekaris who knew the slow loris was “endangered.” What if everybody wanted one. So she and her fellows set to analyzing “12,411 comments” over “33 months” to this “Web 2.0” video to quantify people’s “perceptions” of the big-eyed poisonous creature.
Because “[a]nalyses of webometric data posted on the internet allow us quickly to gauge societal sentiments.” And because—this is science, now—“Web 2.0 sites provide online platforms for people to interact and collaborate to share interests, activities, backgrounds, or real-life connections by allowing users to share their personal details, information, ideas and imagery.”
Plus “Web 2.0 resources such as YouTube are amongst the most powerful media for increasing awareness”.
Since we are all postmodernists now, there is nothing more scientific than raising awareness. This is when you tell somebody about some political cause and the act of telling them convinces them to believe instantly that which is required of them.
Anyway, “There is an urgent need to quantify the impacts the role the internet has on wildlife trade by recording user behaviour and related attitudes that may lead to such behaviours [emphasis added].”
Turns out the cutie video was posted in lots of places. Our scientists tracked where and when to “give an indication of virility of the video.” Comments from all the sites where gathered and classified. Most were of the awwww variety. Some commenters knew the filthy animal was poisonous. But—recall we’re doing science—“Comments containing more than one category (e.g. ‘it’s so cute — where can I get one?’) were scored twice.”
And then the unexpected happened. On a dark and stormy night in March 2011 a shocking new video was posted, upsetting the scientific structure of the experiment. This video:
Isn’t that the cutest thing you’ve ever seen? Then, somehow (I didn’t read that closely). celebrities got involved.
Grade-A names like Deidre Funk, Tom Kaulitz, Ricky Gervais and, yes, Betty White. Stephen Fry was caught saying “Those of you who champion the slow loris do have a point, of course.” Well, you can guess the rest.
Enter statistics: “non-parametric statistical tests”, “chi-square tests”, “outliers”, and, even a “Kolmogorov-Smirnov two-sample test.” Pure science.
Results? “The three most common types of comments referred to the animal being cute; the viewer commenting on what the animal was doing; or the viewer wanting one as a pet.”
Two spikes were discovered in viewing patterns. One, confirmed with a wee p-value, coincided with the release of the March 2011 umbrella video. “The second spike in January 2012 corresponded with the airing of Jungle Gremlins of Java,” a television program featuring various bizarre animals (recipes were not given).
Finding number two: The drop in “I-want-one” comments “was not significant after the March 2011 spike, where the proportion of commentators who stated they ‘wanted’ a loris did not significantly differ from those who stated they wanted one during the previous six months (χ2 = 1.48, df = 1, p = 0.22).”
Finding number three: Celebrities matter. “Most commentators coming to the site as a result of celebrity endorsements wrote neutral responses (75%), referring just to the celebrity.”
All this led to the discussion of tobacco. Yes, tobacco. “Since many countries have now banned the advertising of potentially harmful products in the interest of public health, the internet has become the new mechanism of choice for advertising products such as tobacco.” Perhaps there was a video of the slow loris smoking a cigar that I missed.
Anyway, more importantly, “Celebrity endorsements are meant to have a powerful psychological effect, as those viewing the endorsement are believed to follow a typical pattern, whereby they hope to identify with the celebrity, wanting to adopt their image.”
This had something to do with the depressing conclusion “In our study, despite the proportion of people wanting [a slow loris] as a pet statistically dropping, the number of commentators that wanted one remained high.”
—————————————————————-
Thanks to the Neuroskeptic where I learned of this “study.”
Dr. Briggs, you spend a lot of time and effort detailing misuse of statistics. How about an occasional example of what would be, in your opinion, a valid and USEFUL application of the discipline in which you received your Doctorate?
Rob,
Ask and ye shall receive.
Start with, for instance, this series (basically a justification of logical probability). Or this book review.
But your point is well taken. More positivity! (I feel its lack myself.)
And then, on the other hand…there is so much flotsam and jetsam that poses for scientific knowledge these days, and it’s increasing, it’s some kind of service to show how silly it all is.
Thanks, but I had something in mind that was an example of how, in the real world, statistics was properly (in your view) applied to assist in a decision. Or even how it might have been appropriately used to understand the confidence to be had in data from a study. I can’t imagine that you think EVERY published scientific paper is garbage.
Briggs;
I hope to read your assessment of two papers addressing similar problems, each with a strength/failing. Student’s Guinness work offers examples of important findings.
Your meditations and observations instruct; your blog roll is a treasure chest: keep up the good work.
‘I want one” does not equal “i will have one”.