Q. How do you score CRV session results?

A. We score sessions in a manner which is far different from the way it is done in the research laboratory. We are more "real world" oriented, and use the scores for much more than just bulk research data. We use the scoring to develop VIEWER PROFILES: that is, profiles of a viewer's strengths and weaknesses. That gives us a DEPENDABILITY rating for each viewer for each category of information. We then use those profiles and dependability ratings to help us in viewer selection and ongoing viewer training.

The individual session scores provide information for a viewer profile database that P>S>I keeps on the work of all its associated viewers. Instead of trying for an overall "accuracy" score, we track that database to find out what each viewer's strengths and weaknesses are. Therefore, when we need a viewer to work on, say, the color of the getaway car, we can look in the database and see which viewers have the highest percentage of correct perceptions of colors. This gives us that viewer's "DEPENDABILITY RATING" as far as colors are concerned. Some other viewer might be great at perceiving sizes and shapes correctly, but be very weak at perceiving colors. We can use him/her to answer another question. Right now, we need someone who is very dependable at reporting the correct color. And so it goes for each question tasked to us by police departments, etc. "What kind of ship is carrying the drugs?" For this one, we would select the viewer who is best at shapes and sizes. "What kind of drug is being smuggled in?" For this, we would get the viewer who is best at smells and tastes.

It became obvious many years back that different viewers have different strengths and weaknesses in the remote viewing arena. It seems logical, then, to look at the tasking which comes in and assign to that task the most proficient viewer for that task. (In actual practice, tasking comes in and I break it up into its component questions, then task each question to the viewer whose track record shows that he/she is most dependable at answering that type of question.)

However, to do this requires a lot more than just having a monitor or analyst think back on a viewer's past results and say, "You know, Joe Smith is really good at that - let's give him the task." That is nothing more than a personal value judgement. In order to know exactly what a viewer's strengths and weaknesses are, you have to have to collect a LOT of data, organize and keep it properly, and then do a LOT of analytic work on it. What develops is then no longer a personal value judgement, but an exact VIEWER PROFILE.

There are certain requirements:

First, you must have feedback in order to judge each perception correctly. I agree with Ingo Swann that, if you don't have feedback, you may be doing a lot of amazing stuff, but you aren't doing Controlled Remote Viewing.

Second, you must have a "non-waffled" scoring system. If, for example, the viewer says:

"There is a red moving vehicle against an unmarked, green background."

...and the feedback shows that Object #1 is a green vehicle against a plain red background, and you cannot tell whether it is moving or not. You have a problem in scoring. A debunker would say, "He got the red in the wrong place! See? Remote viewing doesn't work at all!" A preson who is desparate to believe anything and everything psychic would say, "Well, he got the red right, just in the wrong place, and most vehicles move, so let's give him credit for those two perceptions". Each of these is as unscientific, illogical, and undependable as the other. (BTW: the second example is what usually happens in scoring a viewer's session by most people's methods). In order to facilitate a "non-waffled" scoring environment, I devised the following "outlined summary" method for restructuring the viewer's perceptions into a more judgable format for evaluation against feedback. The perceptions are placed one per line, so each can be judged individually, and placed in outline order so you can easily see the context in which each perception was reported. The viewer's statement is changed from:

"There is a red moving vehicle against an unmarked, green background."

to:

There is a vehicle
...which is red
...and moving
...against a background
.......which is unmarked
.......and green


Within the confines of this structure, there can be no "waffling". It is the vehicle and only the vehicle which is perceived (and judged) either red or not red, moving or not moving. It is the background and only the background which can be judged as green.

There is a vehicle __Y__
...which is red __N__
...and moving __?__
...against a background __Y__
.......which is unmarked __Y__
.......and green __N__


Third, you must decide which strengths and weaknesses you want to track. This is generally decided according to what strengths will be required by the viewer's work situation. In other words, what problems will the viewer be tasked against when he/she works in the real world? I track all perceptions by categories and sub-categories:

Sensories:

Touch(Textures, Temperatures, Weight, Density) Smell Taste Sound Visual(Color, Luminance, Acuity) Emotions(Feelings, Ambience)

Tangibles:

Dimensionals(Sizes, Shapes, Patterns, Relationships, Age) Directionals(Alignment, Direction)

Intangibles:

Conceptual(Purpose, Topic, Name) States of being(Live/dead, Gender, Condition, Changes in states) Causation(Actions done, Actions received, Motions, etc.)

And so on... Once you have decided which categories you would like to track, you can mark each perception as to the category it goes in:

There is a vehicle __Y__ Objects
...which is red __N__ Colors
...and moving __?__ Motions
It is against a background __Y__ Objects
...which is unmarked __Y__ Patterns
...and green __N__ Colors


...and a viewer profile begins to emerge:

Colors 0.00%
Motions ?.??%
Objects 100%
Patterns 100%

(Of course, an accurate viewer profile can't be derived from one session. Accuracy of viewer profiling is only achieved after hundreds of sessions.) By analyzing every session a viewer does against feedback, I am able to develop a "profile" concerning each viewer, in order to know certain facts about him/her. Those facts include the following:

PRODUCTIVITY: The average quantity of perceptions he/she normally produces.

Let's say that in 100 sessions, Viewer A has produced 1000 perceptions about colors. We can call those "Perceptions(color)" to distinguish them from, say, perceptions of size, which would be Perceptions(size). In such a case, he averages 10 color perceptions per session:

Productivity(color) = (Perceptions(color)) /
(Sessions) Productivity(color) = 1000/ 100

Productivity(color) = 10 color perceptions per session

SCORABILITY: The percentage of a viewer's perceptions for which feedback is usually available.

Let's further say that of those 1000 color perceptions, we could only get feedback on 900. That is, 900 perceptions were scored either "Yes" or "No" and 100 were scored "?". In such a case, your percentage of scorable perceptions is 90%:

Scorability(color) = (Scorables(color)) / (Perceptions(color)) * 100 Scorability(color) = (900/1000) * 100 Scorability(color) = 90%

PURITY: The percentage of scorable perceptions a viewer gets which are normally correct.

Since we cannot judge Viewer A on perceptions without feedback, we drop out those perceptions for which there is no feedback and judge PURITY OF PERCEPTIONS only by those which could be scored "Yes" or "No". This is the one which people normally take to be the measure of "how good" or "how accurate" a remote viewer or psychic is. Let's say that of the 900 color perceptions which could be scored, he got 540 correct.

Purity(color) = (Yes(color) / Scorables(color)) * 100 Purity(color) = (540 / 900) * 100 Purity(color) = 60%

RELIABILITY: The predictable percentage of all impressions which can be expected to be valid for a target WHEN NO FEEDBACK IS AVAILABLE.

The above three measurements make up the most important aspects of the viewer's profile. However, those are all figured on the basis of information which has been received as feedback. All the beans are counted, and the bad beans (those perceptions which could not be scored) have been thrown out. What if you want to predict a viewer's performance for a session when it is too early to have feedback? In that case, you have to consider how many of ALL the perceptions will be correct. You will have to judge the number correct against the TOTAL number of perceptions, not just those for which you will probably have feedback. This gives you a viewer's RELIABILITY.

Reliability(color) = (Yes(color) / Perceptions(color)) * 100 Reliability(color) = (540 / 1000) * 100 Reliability(color) = 54%

PROFICIENCY: The percent variance from chance. Here is the tricky one. How much better or worse is this viewer's results than if we just threw darts at a dartboard to get the answers? Using that analogy, let's say that the dartboard were sectioned into two sides, one for "the person is dead" and the other for "the person is alive". The chance of getting the correct answer is 50/50, or 50% chance. If we were judging a viewer in the category "dead or alive", and the viewer got the right answer 500 times out of 1000, that viewer would be said to be "AT CHANCE", and his/her PROFICIENCY(dead or alive) would be no better than using the darts. However, if he got the right answer 900 times out of 1000, his PROFICIENCY(dead or alive) would be "ABOVE CHANCE", and you would be much better off using him to find the answer. A score of 200 out of 1000 would be "BELOW CHANCE", and you would be better off with the dartboard. However, if you want to find out a specific color, you would have to divide the dartboard into many more sections; one for red, one for heliotrope, etc. How many sections? That depends on your own personal color vocabulary. You would only have sections for those colors the viewer can name, simply because he/she will probably not give as a perception any color for which he/she doesn't have a name. Therefore, in order to know how many sections this dartboard has, you have to know the viewer's own personal vocabulary of color words.

Going back to our first example, let's say that the viewer has done 100 sessions in which he/she has given 1000 color perceptions, 900 of which can be judged because there is feedback, and 540 of which are correct. In order to know his/her Proficiency(color), we must know how many color words are in that viewer's vocabulary. There are ways to test for this, and those will be taken up below. For this example, though, let us say that this viewer can recognize and name 50 different colors and shades. In such a case, the dartboard would have to be divided into 50 sections, and the chance of hitting any one section would be 1 in 50, or a 2% chance. The PROFICIENCY score is only concerned with VARIANCE FROM CHANCE. Therefore, if we take the person's PURITY(color) score and subtract the CHANCE(color) score from it, we come up with the person's PROFICIENCY(color):

Proficiency(color) = Purity(color) - Chance(color) Proficiency(color) = 60% - 2% Proficiency(color) = 58% Above chance

There are MUCH better ways to do this, such as using chi-square tables to get a probability, or p-score. However, for the lay person (which most viewers are), this is usually the clearest and most understandable way. Since it is the viewing student's understanding I am after, I use this for the students. If you are a statistician, rolllll your eyes back into your head and utter a sigh of disgust, then be patient and forgiving. The bottom line is that this works.

RELIABLE PROFICIENCY: The predictable percentage variance which can be expected for a target WHEN NO FEEDBACK IS AVAILABLE:

A viewer's proficiency(color), as shown above, is based on proven feedback, with the unknowns thrown out. Again, if you want to know the viewer's proficiency(color) for any new target, where no feedback is available, you have to throw the unknowns back in:

Reliable Proficiency(color) = Reliability(color) - Chance(color) Reliable Proficiency(color) = 54% - 2% Reliable Proficiency(color) = 52% Above Chance

--------------------------------------------------------------------------------
Now, back to the question of how to get a viewer's actual vocabulary. There are basically two ways to do it. First, you can give a viewer a vocabulary test. The trouble with this is that generally, the test, itself, is only an indicator of the vocabulary of the person making up the test. The second way is to keep a list of all the words a viewer uses - again, for example, color words. As the viewer performs more sessions, each color word in the new session is compared to the growing list. If it is not on the list, it is added - and all the viewer's profile numbers change. The ideal way is to use a combination of the two methods. Unfortunately, that is also a very cumbersome, time consuming feat. What I generally do is to take all the basic colors: red, yellow, green, etc., and assume that the viewer will know those words. I then add to that list as the viewer turns up with new color-word vocabulary not already on his/her list. But remember, a separate list of each category of words must be kept for each viewer.
--------------------------------------------------------------------------------

Therefore, in this example, the viewer would have the following VIEWER PROFILE entries for dealing with color perceptions:

Productivity(color) = 10 perceptions(color)/session Scorability(color) = 90% Purity(color) = 60%. Reliability(color) = 54% Proficiency(color) = 58% Above chance Reliable Proficiency(color) = 52% Above Chance

But that's for color perceptions only. Now, in order to get the viewer's OVERALL profile, all you have to do is figure the same calculations for tastes, smells, shapes, sizes, concepts, emotionals, textures, motions, ... And then all you have to do to get the average profile for American viewers is to do the same for hundreds of viewers from America. Then, for viewers above age 35... Then, for left handed viewers... Then... Then add to that the fact that the VIEWER PROFILE contains other aspects not shown here: such things as charts showing a viewer's track record of each of these aspects for sessions worked in the morning hours vs. sessions worked in the afternoon or evening. The result is that there is no one simple number which can be given for any viewer or group of viewers to say that their "accuracy" is such and such. Parapsychology may be hell to prove with logic or evidence, but for bean-counting statisticians, it is a true unending paradise. This, then, is P>S>I's way of measuring the viewers' results. It is totally APPLICATIONS oriented. It sees "accuracy" as only a minor part of the equation. The result is, that when answering incoming tasking, the resultant "scoring" for accuracy, precision, reliability, etc. for the group effort can be significantly higher than the "scoring" would be for any one remote viewer alone. And therein lies the falacy of what you are hearing in the newscasts today: the military and CIA are interested in the numbers which result from CRV APPLICATIONS in the field, and the newscasts are reporting the results of CRV RESEARCH in the lab.
--------------------------------------------------------------------------------
I feel strongly that this type of scoring system would work for all types of parapsychological efforts: dowsing, PK studies, etc. The thing to remember is that it is strictly applications-oriented in nature, not research-oriented.

Lyn Buchanan