HomePsychic Powers

December, 2002
John Edward is a supposed psychic who has written best selling books and has a syndicated TV show of his own: Crossing Over with John Edward. Whilst I do not live in a country that shows Crossing Over, I have been able to read many transcripts from the show thanks to some diligent work from people on a John Edward discussion board.

These transcripts are poured over in great detail by both believers and sceptics on the board looking for signs of cold reading or genuine mediumship. But often they're only looked at one at a time. It occurred to me that you'd learn more by examining a large number of transcripts to see if any patterns emerge.

So, I've taken a little time to go over these transcripts (a total of 21 from Crossing Over) to try and quantify John Edward's talents. I also had access to (again, thanks to posters on the discussion board) extensive notes from two live seminars and (thanks to CNN) three transcripts on Larry King Live.

The majority of this article will compare the live, unedited transcripts from "Larry King Live" and the two seminars with the edited Crossing Over. Although I have only five unedited examples, compared to twenty one edited, those five combined are longer than all the others put together.

You may feel that it is a pointless quest and that you're already secure in your belief that Crossing Over edits in John Edwards favour. After all, a TV show that doesn't show its star in a good light is perhaps more implausible than talking to the dead. Nevertheless, I wanted to pick over the pieces of the various transcripts to make sure. I've often read that Crossing Over is only edited for time, and that out of a three hour session, they get six shows. There's trimming for adverts, but that's about it. This "edited for time" is a claim I've seen used several times, so I thought it worthwhile to try and settle it once and for all.

In the article I look at the hit rate, then John Edward's success with getting names and finally a brief look at those amazing guesses that come out of nowhere which (according to many believers) "a cold reader would never get".

The hit rate

In order to judge this, I had to define hit rate to make things simple and clear. To this end, I called a statement a hit if "the sitter responded in an affirmative manner or could relate to the information given." This threw up some problems. For a start, John Edward often gives out many guesses at once, so quite a lot of his stuff is never acknowledged. Many months ago, when I began this survey (using Crossing Over scripts only), I was looking for clues for cold-reading, and this certainly has all the hallmarks of a mentalist tactic: All the audience will remember was that you gave a lot of information, and the sitter kept agreeing with you. Thus I counted these as a miss. That may seem harsh, but simplicity was paramount.

The other two problems, though, worked in John Edward's favour. First was the sitter's willingness to fit parts of their life to anything John had just said. These I counted as hits, unless it was completely obscure and a real stretch. The next problem was John's habit of saying "do you understand?" after some questions. The sitter would always respond "yes" and I was never sure if they were saying that they understood the words John had just used, or if they could relate to the things he had said. Either way, it went down as a hit.

So with all that, the hit rate for all the Crossing Over with John Edwards transcripts I looked at was a healthy 74%.

Looking at the "Larry King Live" and seminar transcripts, the hit rate weighed in at slightly more pallid 49% (Combine the two, and you get 58% for all transcripts).

So straight away, there's some big discrepancies here, and we have our first sure sign that Crossing Over is not just edited for time.

Names

I looked at the success rate for getting names in a few ways. One was the basic hit rate, one was a study of use of initials or full names and the last was according to the relationship of the person guessed.

I should point out that the vagueness of my definition of 'hit' works in John's favour again. If he offers a guess, and the sitter comes back with something similar, it goes in as a hit. Only if it was completely off the wall did I put it as a miss. Towards the end, however, in the Larry King Live transcripts I admit I felt a little sorry for John and let a few go in as hits (Betty for Billy or Keith for Kathleen) that I perhaps shouldn't have.

First the hit rate for Crossing Over with names is 90%. This is an astonishing success rate, and raises a few doubts even before you look at the other hit rate, which is 51%, even with me being generous.

Then came the type of guesses he used. I split them into three categories, (i) initials, (ii) choice of full names, or (iii) one full name. So that "Glen, Greg, G name" would count as a guess using an initial, "Glen, Greg" is a choice of full names and "Greg" is one full name.

Looking at the amount of usage for each type, there's little evidence of tampering since they seem broadly similar. In Crossing Over he relies on initials 54% of the time, and in the unedited scripts 52%. He uses a choice of full names 16% in his own show and 24% at other times. Lastly he uses single full names 30% in Crossing Over and 24% on other shows.

So let's look at the success of each of these methods. First, John's favourites, the initials. In Crossing Over he gets a hit using initials 85% of the time. Elsewhere it's 40%

For a choice of names, in the twenty one transcripts I saw his hit rate was a stunning one hundred per cent! In the unedited scripts it was 64%.

Using a single name gets John Edward a 93% hit rate on his own show, but 60% on others.

This could be construed as evidence for mediumship, since you would expect the initials to get higher hit rates than the others. Mind you, it could be that initial guesses are simply too broad. When asked "Who's the L name?" you'd be forgiven for not knowing where to even start looking. Friend, cousin, in-law? When a name is offered, it narrows the field, and people are able to concentrate on who was called "Lynda or Lindsey", and this could be useful in remembering names from distant family or friends.

So lets look as the distribution of relations that come up. Again, for simplicities sake, I've narrowed the field to five categories. Immediate family (including in-laws and cousins), distant family, friends/others and direct hits. That last one needs a little explaining. A direct hit is when John gets the name of the sitter or the deceased after the deceased has been established. For example, "I'm getting a P name," "That's my Aunt Patty" wouldn't be a direct hit. However, "I'm seeing an older woman," "My aunt," "A P name," "She was called Patty" would be a direct hit. And John didn't have to guess that name immediately after the deceased was established, by the way. Any time later in the sitting would do. So, the results…

If you're wondering why those numbers don't add up to 100%, it's because the rest were misses.

The spread seems similar here, too. Immediate family getting the lion's share of attention, which you'd expect. The higher percentage of direct hits in the edited material is no surprise either. I feel that it's clear misses are edited out, but there's no deliberate policy of favouring one type of guess over any other, for example they don't appear to try to edit out guesses with initials to make John look better.

Guesses from out of the blue

A big mainstay of pro-John Edward arguments is his ability to get amazing hits that "cold readers wouldn't get". Often people's memories of these amazing hits are inaccurate, and they'll make a claim that John Edward pointed to one person and said something amazing and exact. While this happens, I've only seen a transcript for it happen once. More usually, it turns out that the sitter volunteered the majority of the information, which John builds on or that his statement was to an area of the audience not one person.

But I think it's worthwhile to take a little time to look at these types of hits, and they've been sorted according to their sources. First I give the hit a title (in bold), then the text (in different formats according to the style of the transcriptee) and lastly sometimes I add notes in italics. It's worth noting the repeated guesses, such as people with missing legs and father's not using their real names.

I'll make my conclusions now, and leave the excepts as a coda to this whole thing. I'd say that relying on these types of hits is a dangerous tactic, because they are far outweighed by the occasions on which John Edward went out on a limb only to face a big "no". This is especially clear on the unedited examples where John Edward strikes out in some style. Judging by the skewed nature of hits to misses in this category, especially when you compare the Crossing Over stuff with the other unedited texts, it's plain something's certainly going on in the editing room.

For the record, when I began this by looking at Crossing Over transcripts, I was impressed. While he made some gaffs, he seemed to be getting some good hits and the miss rate was pretty low. Now I've seen more transcripts of him working live, I'm quite confident that the claim "Crossing Over is edited for time, not for content" is at best euphemistic.

The reason, I guess, the editing isn't more strict (ie, why don't they remove all inaccurate guesses) is because John Edward does a lot of live performances, and if people arrived there expecting a flawless performance like they'd seen on TV, they'd soon ask for their money back. As long as John Edward keeps pitching it at the right level and keeps getting at least a few amazing hits, he can rely on over eager sitters and viewer's erroneous memories of past successes to keep him going.

Read the Greatest Hits And Misses of Psychic John Edward.