Statistical Evaluations for Public Experiment
#6
Here are three test procedures that evaluate the remote-viewing session
with the target data. All of these tests utilize Farsight's Session
Analysis Machine (SAM). Click on the test names to get an explanation
of each test.
The analyses below are for:
Viewer: Courtney Brown
Session: Session #2
TEST #1: Comparing
the Remote-viewing session Data with the Target Attributes
(Click for explanation)
The session data are: |
Session/Target Matches: |
surface: surface |
match |
surface: level topology |
match |
land: land |
match |
land: manmade |
match |
land: level topology |
match |
atmospherics: manmade smells |
match |
atmospherics: smoke or burning (natural or manmade) |
match |
surface structure(s): surface structure(s) |
match |
surface structure(s): one |
match |
surface structure(s): multiple |
match |
surface structure(s): subjects inside |
match |
surface structure(s): subjects on base surface outside |
match |
structure(s) materials: manmade materials |
match |
structure(s) general location: on land |
match |
structure(s) general location: on a flat surface |
match |
subject(s): subject(s) |
match |
subject(s): male |
match |
light: bright |
match |
energetics: explosive, swirling, or multi-directional movement |
match |
energetics: kinetic (fast or slow, one direction) |
|
energetics: fire or heat |
|
activity: activity or movement by subject(s) |
|
activity: activity or movement by object(s) |
match |
sounds: booming or roaring |
match |
sounds: wind-type sounds |
|
sounds: loud |
match |
temperatures: hot |
|
dominant session elements: structure(s) on a surface |
match |
sketches: structure(s) |
match |
sketches: structure(s) on a surface |
match |
sketches: subject(s) |
match |
sketches: horizontal base surface |
match |
sketches: radiating or explosive energetics |
match |
The target attributes not perceived are:
Missed Target Attributes: |
surface structure(s): city |
subject(s): female |
subject(s): one/few |
subject(s): many/crowd |
subject(s): focused gathering |
environment: urban |
sounds: talking, shouting, voices |
temperatures: moderate |
sketches: subject(s) in a structure |
sketches: subject(s) on an outside base surface |
The total matches between the session and the target are: 28
The total number of target attributes not perceived: 10
The total number of session entries is: 33
The total number of target entries is: 38
A. The total matches between the session and the target as a proportion
of the total number of target attributes are: 0.737
B. The total matches between the session and the target as a proportion
of the total number of session entries are: 0.848
General session/target correspondence (the average of A and B above): 0.7926634768740032
The normal chi-square value with 1 degree of freedom testing the fit of
the session to the target based on the table below is: 40.958
The alternative chi-square value with 1 degree of freedom based on only
the distribution of chosen session attributes (the top row of the table
below) is: 26.425
The correlation between this session's data and the target attributes is:
POSITIVE
NOTE: The chi-square value does not take into account the direction of the
relationship between the session data and target attributes. The chi-square
value is a useful measure ONLY if there is a positive correlation between
the target's attributes and the session's SAM entries. (That is, there needs
to be a reasonably high number of target and session matches.)
|
Target 0: |
Target 1: |
Session 1: |
5 |
28 |
Session 0: |
50 |
10 |
Chi-square Values: |
Significance Level: |
3.84 |
0.05 |
5.02 |
0.025 |
6.63 |
0.010 |
7.88 |
0.005 |
10.8 |
0.001 |
INTERPRETATION OF THE CHI-SQUARE STATISTIC
1. If the value of the chi-square statistic is equal to or greater than
the chi-square value for a desired significance level in the table above,
and if the correlation between the session data and the target attributes
is positive, then the session's data are statistically significant descriptors
of the target.
2. If the value of the chi-square statistic is less than the chi-square
value for a desired significance level, then the remote-viewing data for
the session are not statistically significant. This normally means that
there are decoding errors in the data.
3. If the value of the chi-square statistic is equal to or greater than
the chi-square value for a desired significance level but the correlation
between the session data and target attributes is negative, then the session
either has major decoding errors, or there may be conscious-mind intervention
and/or invention in the data gathering process.
HEURISTIC COMPARISON: Comparing the Session with a Target with Randomly
Chosen Attributes
The total matches between the session and a target with randomly chosen
attributes are: 17
The total number of session data entries is: 33
The total number of target attribute entries is: 38
The total matches between the session and the target as a proportion of
the total number of target entries are: 0.447
The total matches between the session and the target as a proportion of
the total number of session entries are: 0.515
The normal chi-square value with 1 degree of freedom testing the fit of
the session to the target based on the table below is: 2.403
The alternative chi-square value with 1 degree of freedom based on only
the distribution of chosen session attributes is: 1.550
TEST #2: THE
RUSSELL PROCEDURE (Click for explanation)
Part I.
The expected mean number of chance matches for this session is: 13.484
The standard deviation (hypergeometric distribution) for this mean is: 2.280
The 90% confidence interval for this is: [9.732, 17.235]
The 95% confidence interval for this is: [9.014, 17.954]
The 98% confidence interval for this is: [8.182, 18.786]
The unweighted (actual) number of matches between the session and the target
are: 28
The weighted number of matches between the session and the target are: 34.678
INTERPRETATION: If the unweighted and/or weighted number of matches between
the session and the target are outside of (that is, greater than) the desired
confidence interval, then the number of matches obtained in the session
was not by chance.
Part II.
IF THE SESSION DATA WERE RANDOM, HOW MANY SAM ENTRIES WOULD BE NEEDED
TO DESCRIBE THE TARGET AS COMPLETELY AS IS DONE BY THE ACTUAL SESSION?
From 1000 Monte Carlo samples:
The mean number of random session pseudo SAM entries that are needed to
achieve 28 matches with the target is: 67.271
The standard deviation is: 5.311
Lowest number of pseudo attributes from sample = 48
Highest number of pseudo attributes from sample = 81
The 90% confidence interval for this is: [58.535, 76.007]
The 95% confidence interval for this is: [56.862, 77.680]
The 98% confidence interval for this is: [54.923, 79.619]
Compare these intervals with the actual number of session entries: 33
INTERPRETATION: If the actual number of session SAM entries is outside
of (that is, less than) the desired confidence interval, then the number of entries utilized by the remote viewer to obtain the number
of matches between the session and the target was not by chance.
TEST #3:
CORRESPONDENCE and CORRELATION (Click for explanation)
PART I.
The correspondence data in the table immediately below are computed using
the targets from the public demonstration only. The "Session/Target"
correspondence numbers are calculated between the remote-viewing session
for this experiment and all of the targets used in the public demonstration.
The "Target/Target" correspondence numbers are calculated between
the real target for this experiment and all of the other targets in the
public demonstration pool.
Experiment
Number: |
Session/Target
Correspondence: |
Target/Target
Correspondence: |
Experiment #1 |
0.559 |
0.520 |
Experiment #3 |
0.667 |
0.906 |
Experiment #4 |
0.758 |
0.802 |
Experiment #5 |
0.731 |
0.827 |
Experiment #6 |
0.793 |
1.0 |
Experiment #7 |
0.793 |
0.938 |
Experiment #8 |
0.688 |
0.720 |
Experiment #9 |
0.236 |
0.285 |
Experiment #10 |
0.7 |
0.865 |
Experiment #11 |
0.581 |
0.641 |
Experiment #12 |
0.298 |
0.326 |
Experiment #14 |
0.679 |
0.734 |
Experiment #15 |
0.750 |
0.917 |
The correlation coefficient is: 0.877 with an N of 13
INTERPRETATION: All targets have a variety of descriptive characteristics.
When comparing one target with another, both similarities and differences
will be found between the two. The correspondence numbers are one measure
of the degree of similarity between any two sets of SAM data, and these
numbers can be used to compare one target with another target, or a remote-viewing session with a target. The closer a remote-viewing session is to
describing all of a target's characteristics, the higher will be the correspondence
number between the session and the target. Since a pool of targets normally
contains targets with a great variety of descriptive characteristics, comparing
correspondence numbers for the remote-viewing session and its target across
a variety of other targets tests how closely the session describes all of
the essential characteristics of its real target. When compared with other
targets with many different characteristics, both the remote-viewing session
and its real target should have correspondence numbers that vary similarly.
The correlation coefficient summarizes this relationship. The correlation
coefficient can vary between -1 and 1. The closer its value is to 1, the
more closely the remote-viewing session describes its real target's various
characteristics.
PART II.
The correlation coefficient is computed as in Part I above, but now using
a large (240) pool of SAM targets.
The correlation coefficient is: 0.968 with an N of 240
The lowest correspondence number for the session and pool is: 0.108
The highest correspondence number for the session and pool is: 0.833
The lowest correspondence number for the target and pool is: 0.108
The highest correspondence number for the target and pool is: 0.921
INTERPRETATION: Similarly as with Part I above. The closer the value of
the correlation coefficient is to 1, the more closely the remote-viewing session describes its real target's various characteristics.
|