<< Feature Articles >> The results of the Megalithic Portal Map Dowsing Experiment and Competition
Submitted by Andy B on Tuesday, 22 November 2011 Page Views: 8292
Mysteries I have been busy analysing the entries for our dowsing competition from last week and here are the results. Various people sent descriptions of how they did the dowsing and it's clear a lot of thought went into these entries.Thanks to a mention by Maev Kennedy in The Guardian and various dowsing groups passing the word around we had 42 dowsers of all levels of experience (and scepticism) taking part.
We also had 16 people who sent in guesses - perhaps a lot of thought went into these as well - who knows! The numbers guessing were a little disappointing, clearly map dowsing is more fun. Not surprisingly all those in the guessing group were either sceptical or open minded / impartial about map dowsing. So on to the actual results:
For Target X, George's Rock Art, 7 people out of 42 correctly dowsed the rock art as being in square B3. This is 17% of entries correct. I am not a statistician but I unearthed my 25 year old A-level Maths and an online calculator of cumulative binomial probability to calculate that the chances of getting 7 or more correct answers by chance alone are around 0.057, which is one in 17. This is (I think) almost at a 5% significance level. Hopefully a stats expert can help us out here. So the results are on the verge of what you would call 'statistically significant'. Just 'by eye' a result of 7 appears to be slightly above chance to me - by pure chance you would typically expect 3 or 4 correct 'hits' but to get 2 or 5 by chance is almost as likely.
Tuesday evening update: Note that only squares B4 and A3 got 6 and 5 dowses respectively and although 'wrong', these are adjacent to the correct square. You could take this as evidence towards or against the case for dowsing.
However looking at both maps there seems to me to be a bias in the dowsing results generally towards the right hand half of the maps which is not present with the guessed results. This is even though Target Y is well to the left. There also appears to be a small but measurable bias away from the visible 'valley' on Target X. See the comments below for more analysis.
The dowsers who hit the correct square for target X are
Lynne Kelly (sceptic)
Jill Maunder (open minded / impartial)
Ina Sigerist (experienced dowser)
John L Hanscum Jr (experienced dowser)
Brian Hummerston (experienced dowser)
Roger Albin (sceptic)
and Michael Rinaldi (occasional dowser)
For Target Y, Runemage's quartzy boulder, 2 out of 42 identified the correct square as C1. This result is a little low but within the limits you would expect to get by chance (Using the probability calculator again says the chances of getting 2 or fewer correct hits is one in three) so this is within what you would expect from a small sample.
The dowsers who hit the correct square for target Y are Sue Brown (occasional dowser) and Brian Hummerston (experienced dowser).
50% of the dowsing group who declared their attitude were experienced or occasional dowsers who believed that map dowsing works. However it is interesting (but probably not significant) that 5 of the 8 correct entries came from the group of dowsing believers whereas only 2 out of the 8 correct dowsers came from the sceptical group.
A final surprise - Brian Hummerston, who declares himself an experienced map dowser managed to correctly dowse both squares which contained our ancient sites! Initially I got really excited by this but it then dawned on me that someone has a 1/144 chance of getting both correct by chance, and we have 42 people entering. So purely by chance there is a roughly one in three chance that of one of our 42 dowsers would get both correct. It still seems a good result, and I'm sure Brian will think that too!
Brian is one of our calendar winners through dowsing, the other winner chosen at random being Michael Rinaldi.
For the guessing group, one out of 16 hit upon the correct square in each case which seems to be a reasonable chance level, and is good on the calendar giveaway front as well. The correct guessers are Hamish Fenton (yes he of the super kite photographs) and second Simon Charlesworth, our Welsh stone alignments guru.
What is interesting about the guessing entries is that while 69% of the guessing entries said they were very sceptical about map dowsing, both of the two correct guesses came from the 25% of people who declared themselves as 'open minded / impartial'. What we can read into that I don't know...
So a very successful experiment and competition. Thanks to everyone who took part.
The raw data:
Target X Dowsing: A1,B3,C1,B4,A1,A2,B4,B1,B3,B4,B4,B4,A3,C1,A4,A2,C3,C3,B1, B2,C4,A3,A3,A1,B1,B3,A4,A3,A4,B4,C1,B3,B1,B3,C4,C4,C2,B2,C4,B3,B3,A3
Target X Guessing: A3,B2,C1,A4,C1,A2,C1,B2,C2,A3,B4,A4,A2,B3,A4,A4
Target Y Dowsing: B3,C2,C3,B2,B3,B2,A4,B2,A3,B4,C2,A3,B3,B4,B2,B4,A3,A3,C4,B4,A2,B2,B2,C4,B1, C4,B3,B4,C2,C4,C2,B4,C4,C1,C3,C2,C1,B3,C2,C3,C3,B2
Target Y Guessing: B1,C1,B4,B4,B3,B3,B3,B3,B4,B2,B2,A3,C3,B1,C4,C2
As I said above, if any statistics experts would like to help me analyse the data, please get in touch. I also have the results in an anonymised spreadsheet, if you would like a copy please email me via the 'Contact Editor' link to the left. Or please feel free to leave comments below on what you think of the results.
<< The Megalithic Portal Mysteries Forum Grand Map Dowsing Competition