Because the universe is beautiful enough without having to lie about it

More Fraud Ratings

January 15th, 2011 Posted in Internet

So I was just browsing through some old files the other day and I came across the script that I wrote to gather fraud ratings. We’ve seen this a few times before in this blog, but briefly to recap – I wondered how the number of web hits for a selection of famous “psychics” would change over time, and how that would compare to the number of pages with that psychic’s name plus the word “fraud”. Of course, none of this means that any of these people actually are frauds – maybe they really are able to perform the physically impossible feats that they claim to achieve. For legal reasons, I prefer to remain ambiguous on my own position regarding these performers, though I suspect most people will be able to guess 😉

The thing was, the script I wrote was just running on a weekly cron job that meant it had kept running over the last few years without my input, and hence I have several years worth of data now. So I though I’d throw up some graphs, and they make for interesting reading.

As the scales for these values are different (for some reason, John Edward seems to inspire six times more web pages than the other three combined), I’ve scaled them all to their level as of July 2008, so I could display them all on one single graph. Having said that, the graph for Edward was a bit strange, so I’ve plotted him separately. So here’s the Google hit count for each of these three subjects (minus John Edward) over a two-and-a-half year period, scaled to July 2008 = 1:

Hit Count for Sylvia Browne, James van Praagh and Uri Geller

Combined Hit Count

Note that there are occasional spurious high counts. I have no explanation for those, except that occasionally you do get hit count artefacts from Google that may be something to do with the state the servers were in when you queried them. These anomalies are clearly not ‘real’, as they represent a difference in counts of a factor of at least ten, sometimes as much as one hundred times! What I should probably have done was take a measurement every day for a week, then take the average each week. That would smooth it out a bit.

Next, let’s take a look at the hit count for John Edward. This is particularly bizarre for two reasons. Firstly, it’s very much higher than the other three. This isn’t entirely surprising as John and Edward are both very common English names. However, what is most bizarre is that the count changes very substantially for a while in the middle of 2008. My theory for this is that it might be related to the American Politician John Edwards, who was very much in the news at that time because of revelations of an extra-marital affair.

Hit count for John Edward

Hit count for John Edward

So I don’t really know what to think of the John Edward graph. Probably we can’t derive much from it – it’s far too noisy. Probably, as I’ve said, a consequence of both his first and last names being very frequently used.

The first graph shows something very interesting – it shows that the number of web pages written about all three of these subjects is steadily declining over time. If this were due to changes in Google’s algorithms, I would expect a far less smooth descent, and I would also expect it to be less well coordinated between the three subjects. The decline is definitely significant, and is of the order of roughly a factor of two. What I don’t know is whether or not this is tied to some larger trend. The oracle for all things Internet-statics-related, Netcraft, showed in their January 2011 survey that the Internet growth is definitely still continuing, but interestingly that there was a bump approximately coinciding with the year 2009. This doesn’t quite match the bump I see in the Edward figures, which is about 9 months earlier. The bump spotted by Netcraft is linked to blogs in the domain, which were public for a while in 2009, and were then taken private again. That’s 30M pages. Due to the dates involved, it seems not to be related to the bump I saw and I prefer my earlier explanation.

The bump at the beginning of the graph for Uri Geller, around the end of 2008, coincides with a number of his more public television appearances across Europe, and perhaps the decline leading into this could be related to his high profile US TV series, “Phenonenon”, which finished in April 2008.

So what happens when we add in the word “fraud” into our search terms? Well we do get a different picture, but perhaps not what we were expecting.

Adding in the term "fraud" into the search

So the noise level here is a bit worse than before – again, a weekly smoothing would sort this out to some degree. However, what we do see is, again, a definite downwards trend, with a few exceptions. Take a look at the graph for Uri Geller, for example. There is a marked increase in the number of web pages mentioning his name with the word “fraud” in the period between the end of 2009 and summer 2010. I’m not sure what happened here to cause this – Uri had a string of high profile television performances between 2007 and 2009, as I mentioned earlier, but nothing much in this time. James van Praagh also has a “fraud bump” in the first half of 2010. I wonder if these correspond to some kind of skeptical action? I couldn’t think of anything in particular, but feel free to comment if you have any ideas.

The final graph is the fraud graph for John Edward. Again, I don’t know what to make of this.

John Edward Fraud Graph

John Edward Fraud Graph

There’s really not much to say, except for the bump corresponding with the bump in the total number of pages, in early 2009. But then, after May 2010, the count goes crazy. As I’ve said before, I don’t really trust any increase of that magnitude, so it looks like Google is doing something mad with the page hit counting. I wish I knew more, but Google is a bit of a black box, I’m afraid.

So there you go. Nothing much to take from the “fraud rating” side of things, except for a potential flag for increased anti-psychic action in mid 2010. However, the overwhelming message to learn from this study is that psychics are losing ground on the Internet. There are now less than half as many pages about leading psychics than there were at the end of 2008. And in that time, the number of pages on the Internet, according to Netcraft, has increased by roughly 50%. And that can only be a good thing. Maybe we, as a society, are finally growing out of this ridiculous notion that certain privileged people have the ability to do impossible things with their minds.

Be Sociable, Share!

Post a Comment

To protect against spammers, please enter the letters you see below

Please don’t bother posting "you’re wrong, you jerk" comments, unless you can back them up with valid scientific research papers.