FACIAL GEOMETRY AND REINCARNATION Pilot Study Analysis (2/27/09)
Assumption to be Tested: Similar facial geometry in robust cases is consistently stronger than in the public at large.
While focused on childhood memories and unique body markings, Ian Stevenson also noted certain physical characteristics in the subject (S) often matched those in the alleged previous personality (PP). Other researchers (Semkiw, Finkelstein, Keene, Ravel, and others) who evaluated apparent past-life matches initiated by a variety of clues similarly noted that in their strongest cases the S and PP looked alike. In most of these cases, the past-life hypothesis was predicated on verifiable memories, validated dream or hypnotic material, or similar personality features.
In the Reincarnation Experiment (REXP), we reviewed well-documented, independently researched cases in both the public domain and private stories and also found strikingly similar facial features. Our psychophysical model uses correspondences in personality factors (mental, emotional, interpersonal), career interests, and evidence of a common set of specific memory traces, as well as physical factors. We did not include cases based solely on subjective look-alikes. (See further discussion at this link.)
Nevertheless, publication of findings by the above researchers and ourselves stimulated an increase in reincarnation enthusiasts matching individuals with well-known personalities whose images are available on the Internet. Most of these cases are based only on someone's intuitive "hit" or personal projections.
Our review of many of these alleged matches revealed wide variations in the degrees of correspondence between their actual facial dimensions. We decided to establish a more objective procedure to evaluate how much the look-a-likes were actually alike. While using photo overlays and grid measurements of photographs, portraits, sculpture, etc., we discovered a study by Martin Braun that dealt with the use of biometrics with 18th century portraits. (See The Soul Genome, p. 87.)
We developed a similar process to compare the facial images (where available) in order to establish a reliable measure of the variances between images of the S and their alleged PP. We theorized that if strong facial similarities were part of the mechanism (psychoplasm) that appears to account for the reported legacy of memories, habits, and personality traits, there should be a strong correlation between close biometric measures and corresponding nonphysical factors.
Evaluating the Assumption: Using Biometrics and Statistical Analysis.
Based on Braun's success of using only seven dimensions, six measures were selected from the biometrics used in photo identification cameras and computer software. (See the attached illustrations -- from biometrics and our simpler model.) Millimeter measures were taken from the available images of both the S and the posited PP in a set of 13 robust cases collected by REXP.
The six measures were converted into 3 ratios for both the S and PP. (See the attached form.) When the ratio variances were compared, a degree of difference was established for each pair. Similar measurements and ratios were obtained on 132 random photo pairings taken from the Internet by an independent collector.
(Four people were involved in measuring different individual sets of images. Multiple images were used in cases where good frontal photos were available. We assumed slight differences in pinpointing the points to measure would average out for both the robust sample and random sample.)
Statistical Test Performed: The null hypothesis was that no significant difference would occur, i.e., that the so-called reincarnation group would be no different from the random group. The two-tailed unpaired "t" test was used to develop the statistical significance of the different levels of variance associated with the S/PP measures and those associated with the random sample. Since our assumption was that the robust group's mean and standard deviation would at the extreme close side of the bell curve, the data was converted to a one-tail test. A summary of the findings follow:
P value and statistical significance: The two-tailed P value is less than 0.0001 By conventional criteria, this difference is considered to be extremely statistically significant. Converting t = 4.1299/df = 433 to one-tailed test makes P = 0.000022, i.e., the odds are about 20 in 100,000 the difference is due to chance.
In other words, the odds that the extremely close facial geometry of the S/PP pairs happening by chance is 2 out of 10,000. This extremely statistically significant finding suggests that either the we in the REXP completely fabricated the data for the robust group or that some other reasonable explanation is called for. The "reincarnation hypothesis" suggests that that mechanism may be something like the psychplasm (which may be conceived of as a bioenergetic, information-rich hologram that determines the epigenome's activation of relevant DNA patterns).