Social Icons

Saturday, 11 May 2013

TobaccoTactics Wiki Stats Debunked - UPDATED with Awesomeness

[This post was originally published at 12:08 p.m. on Friday, 10 May 2013. It has been updated with further awesomeness and republished accordingly. See Updates 1 and 2 below.]

With all of the half-truths, distortions of fact, misdirection and flat-out lies that the tobacco control industry ejaculates on a hourly basis, sometimes the lies are so fantastical that the mind boggles.  I suppose we're used to it, or at least we expect it. Even so, you would think they would at least be clever enough to lie about something that no-one could ever prove false.  This is not the case it seems.

On the main page of the University of Bath's TobaccoTactics wiki, under the heading "Is TobaccoTactics.org meant for me?" and in sub-section "Visitor Statistics" is one such falsehood that can be proven false.  It reads:
Since it was launched in June 2012, the website has received 1.4 million hits, with over 800,000 pages served with a daily average of nearly 6,300 hits! These April 2013 statistics, along with positive feedback from users of the site, reflect its success and popularity.
Here's a screen-capped image, just to confirm the above:
Over 1 bazillion served! We would we lie to you? (Don't answer that.)

Really?  C'mon, pull the other one, ladies. 1.4 million hits from web viewers all over the world in 10 months' time is a huge number of hits for a brand new web site that doesn't have stolen images of nude celebs or Lolcats.  But the statistics the editors of TobaccoTactics have put on their main page do not seem to be true or accurate.  I know this because Tyler knows this the site's own statistics show an entirely different story.  Have a look:

Click to enlargify - see highlighted text

Those statistics shown above are screen-capped from the TobaccoTactics wiki statistics page on 9 May 2013. You can view the current stats here via this link.

The truth according to the wiki software, which is designed to keep track of this kind of stuff, is that the wiki has had only 409,826 views as of this writing.  That's a far cry from "1.4 million hits and 800,000 pages served."  And of those views, how many were generated by the wiki's editors and didn't come from outside of the University of Bath?

But it's worse than that.  Because apparently Anna Gilmore (the Queen of Junk Science) and Eveline Lubbers not only suck at lying, they might also suck at maths, too. The Visitor Statistics statement on the main page says they get "a daily average of nearly 6,300 hits."  OK, 6,300 is an average, a median, so some days the wiki would get less and some days more, but it's impossible for me to know what the highs are and lows are without having access to daily stat information.  We'll just have to go with their bullshit figure of 6,300.

So let's completely waste our time by doing some basic calculations.*

Let's put the date of the launch of the wiki at 1 June 2012 (it was actually around the 5th, 6th or 7th of June, I can't recall which day, but I'll give them a few extra days).

The statistics they gave were taken from 1 June 2012 to 30 April 2013, or eleven months, or 334 days.

1,400,000 divided by 334 =

4,192

But Eveline Lubbers wrote that the daily average was 6,300.  OK, let's calculate for that:

Let's multiply 334 by 6300.

We get  2,104,200 hits.

So the figure they gave is 1,400,000 hits, which is 704,200 hits fewer than the result calculated for an average of 6,300 hits daily.  Something isn't right.  I mean, sure it's possible that they they could have 20,000 hits in single day to come up with such a high average, but that possibility is extremely unlikely.

None of that matters, though.  Because the statistics page's View Statistics says they've only had 409,826 proper hits.  That's a difference of 990,174 hits, even after nine extra days have passed from the end of April.

The statistics page also says that "Views to non-existing pages and special pages are not included."  Is it possible that the special pages of the wiki, an area that few people would bother to look at or even knew existed, received a whopping 990,174 hits, more than twice the number of actual content page views?  Anything is possible in the fairy tale land of tobacco control, as we all know, but in this case probably not. The only possible explanations I can think of are spammers inundating the site with referral link spam all day long, or a few hundred web-crawlers (such as Google or Bing) indexing the entire site per day. Even those are beyond the realm of probability.  Another explanation is that the extra views are generated by the editors as they edit and upload files to the wiki.  If the latter is the case, it's incredibly disingenuous to count those as page hits on your main page. One last explanation I can think of is that maybe, just maybe, the statistics page is broken or was reset, but I think that is also unlikely.

So what do we know? We have 409,826 "proper" page views or all-time content page hits by 9 May 2013.  What's the real average?  Let's calculate from 1 June 2012 to 9 May 2013. That is 343 days.

409,826 divided by 343 =

1,195 (avg)

You know, 1200 hits average per day is not shabby. It's a decent figure. It's not huge; it's not tiny. It's a fair amount of hits for an anti-smoker hate site set up to attack bloggers and anybody that disagrees with the tobacco control industry. Sure, the wiki is not overwhelmingly popular by any stretch, but what anti-smoker site is? Regardless, why do they feel the need to state an average figure that is five times higher?  Why do they claim they have received almost 3.5 times more hits overall?

Well, maybe the answers to those last two questions is the tobacco control industry cannot help but make shit up to give the appearance of massive support. This is a public relations confidence trick. Tell people you have a huge fan base, and you hope that people will believe you are a force to be reckoned with. The truth, however, is that their support is marginal.

The truth is that we can never trust any person or organisation in the tobacco control industry. We certainly never trusted Anna Gilmore in the past, so no reason now to start trusting her or sidekick, Eveline Lubbers, either.

*If my maths are incorrect or if I missed something, please let me know in the comments so I can correct -- I had "indulged"with a few drinks last night whilst writing this post.

UPDATE (10 May 2013, 21:20) : I have just received a message that the TobaccoTactics wiki is updated with evidence of from their server logs or something. I haven't checked myself because I'm in the middle of something right now, but in the in interest of fairness I wanted to put a quick note on this post that they have responded. I will check later and certainly I will update this post further after I have looked at what they've posted. -- Jay

UPDATE 2 (11 May 2013, 00:20):  I have now had a chance to review Eveline's web statistics.  I thank Eveline Lubbers for having the courage to post them up.  I am pleased to say that my post above remains a more accurate picture of the traffic the TobaccoTactics wiki receives. I will explain below.

But first this very important message to Eveline Lubbers:

Look, Eveline. I don't know you, and I don't mean to be cruel, but you know absolutely fuck all about the Internet (using Google, Twitter and editing a wiki does not make you an expert) and you know even less about correctly interpreting web statistics. There must be hundreds of capable IT persons at the University of Bath.  I would strongly advise you to consult with any of them, and after that kindly update your wiki's main page accordingly. Or you could hire me.  My fee is twice the amount they are paying you. The upside to hiring me is that at least you know I would be honest and fair (if not a bit foul-mouthed at at times). The downside for me is that I'd have to spend time with the Tobacco Control Research Group and I cannot think of anything more horrible at present (excluding Karen -- she's a cutie). And if you think for one second that I enjoy writing about this shit, you're wrong. I'd rather be doing anything else, but there's no way in hell I'm going to sit back and let the tobacco control industry deceive everybody.  I certainly don't need to teach you how to analyse data. But hey, if you want a propaganda war, I will beat you at it every time -- I will "pwn" you.  I think you should stick to being an author or whatever you do best, and this gig with Anna, no matter how well paid it is for you, it is not for you. I hope we are very clear, Ms Lubbers. I'm sorry I have to tell you these things. You're ruining your reputation every day you work for the tobacco control industry. 

OK, so let's begin analysing the wiki traffic stats.  This is the image that Eveline posted up on the wiki to show their traffic:

Click to enlargify Source: TobaccoTactics
So I know that looks impressive -- 1,527,557 "hits" -- but hits are misleading until you understand what they really mean.  For now, let me highlight the only two important columns in your charts you need to concern yourself with to understand your "true traffic" stats.

The only data you should look at is "Visits"
See there?  The visits column is what you want.  And over the last eleven months you got 227,649 visits.  This is, unfortunately for the TobaccoTactics team, even worse that what the wiki statistics show and what I wrote above.  But here we have real data to look at and to do some maths on.  So I will.

The first thing to do in complete fairness, however, is to toss out the June stats, because I think the TobaccoTactics team didn't start their web traffic analysis until the last week or so of June, based on the stats.  June's stats appear to be very, very low.  So, as you do with statistics, we're eliminating them from the calculations (I'll add them back in later by assuming that the 2,345 is one week's worth of data). Likewise, I will exclude May's statistics, since we're only 10 or 11 days in, but will not add that in.

Here is a screen cap of my calculations based on the above data:

Jay's super awesome calculations in Excel

So from July 2012 to April 2013, you have had a total of 216,756 visits, with (and this is naughty, because averaging averages is absolutely rubbish maths, but the hell with it, the results are mostly identical for figures this small) an average of 713 visits per day during this time period.

That is respectable traffic. That is nothing to be ashamed about.

But I want to be fair. So I'm going to include June's data, and I'm going to have to adjust June's stats because the data provided above looks like an aberration. I will assume that the June data of 78 views and 2345 hits is only one week's worth of traffic. It could more, it could be less.  But I don't know, and neither do you, it seems.  So, I will multiply the Monthly visits by four, or 2,345 x 4 = 9,380.  I admit this seems a bit low, but... give me more and I'll fix it.  Here's an adjusted spreadsheet:

We should have left June out of the calculations

Right. The TobaccoTactics traffic is actually better off by excluding June altogether as you can see (and see how fair I am to you guys? I'm Mr Fairness. Definitely fair. So fair that ... eh, never mind).

If we count June, then we have an overall average of 677 visits daily and a total of 226,136 visits during those eleven months.  This isn't shabby either.  It's better than what I get on my blog -- granted, I haven't been posting much lately, but I digress.

So you're probably wondering why the crackers "hits" don't matter.  That's because one page view can generate 4, 5, 6, a dozen or more hits. Why? Because it does. It's a quirk of how the Internet works. But I don't want you to take my word as fact, so I've found this guy, whose job it is to analyse this stuff, to explain it to you:

Source: Elbel Consulting Services, LLC

See? I'm not making it up.  You might also want to read this page from that guy which explains how to interpret web stats much better than I ever could.

So, there you have it.  And now you know why we will never trust any research or data that comes out of the Tobacco Control Research Group at the University of Bath.  Because ... well, my friend Bucko The Moose put it most succinctly in his tweet this afternoon:

I favourited this tweet for its awesomeness.

Can you update your fucking wiki with the proper visitor stats now, please?  Thanks.

UPDATE 3 (11 May 2013 17:50): Grandad weighs in here with his post 9.2 million hits. Nice one.

UPDATE 4 (12 May 2013 18:12): Simon Clark of FOREST also wrote this piece called "Stats life – welcome to the fictional world of tobacco control."