Exploratory Block Analysis of Field Consciousness Effects on Global RNGs on September 11, 2001

by

Bryan J. Williams
Department of Psychology
University of New Mexico
e-mail: [email protected]
August 12, 2002

Introduction: A metaphorical “Great Disturbance in the Force?”

I felt a great disturbance in the Force...as though millions of voices had cried out in terror, and then were suddenly silenced...I feel that something terrible has happened...

- Obi-Wan Kenobi (portrayed by Sir Alec Guinness), from the film Star Wars (1977)

The September 11th terrorist attacks on America had greatly shocked the world, leaving behind a deeply wounded trail of sadness, anger, confusion, and fear in the hearts of the many individuals engulfed in their devastating wake. The sheer horror of having to witness two hijacked commercial airliners slam into the twin towers of the World Trade Center and destroy them, a third hijacked airliner hit and severely damage the Pentagon, and the fourth hijacked plane, United Airlines Flight 93, crash in a Pennsylvania field had deeply affected us all not only physically and mentally that terrible morning, but perhaps also in ways that we don’t fully realize. Perhaps the events had penetrated our conscious minds even deeper than we ever imagined, their dynamic and emotional impact so powerful that their effects propagated outward through space-time like the ripples of a pebble dropped into a lake to elicit reactions in the fundamental elements of mind and matter. It has been speculated that something to this degree may have occurred because it seems that, according to preliminary analyses, a pattern of randomness deviation anomalies occurred in random physical systems located around the world around the time of the terrible events, and perhaps even before they even happened! Seemingly strange effects such as these are often popularly viewed by the general public as representing some kind of “great disturbance in the Force,” in which it is assumed that the powerful repercussions of the events send something like a tremor shockwave throughout space-time to affect “the natural order of things” in unseen ways. From a different perspective, it is not implausible to think that these strange effects may be nothing more than the psychic “signature” of the approximately 6 billion conscious minds that exist on the earth reacting in unison to the events through a mass subtle, unconscious, emotion-related psychokinetic (PK, or “mind over matter”) effect as they are slowly drawn together into a shared “global consciousness.”

The Global Consciousness Project (GCP) had been initiated in August 1998 in order to more closely study this “global consciousness” hypothesis and the PK effects that are thought to underlie it by setting up the first Internet-based, worldwide network of random number generators (RNGs) and monitoring it for any signs of this PK-based “global consciousness” effect during major world events that tend to draw a wide degree of attention and bring about a large emotional response (Nelson, 2001). The terrorist disaster of September 11, 2001 was just the kind of major world event that the GCP would expect to create a strong emotion-related global consciousness because of the devastating effect it had on people around the world. In the efforts to explore a possible global consciousness, I carried out an independent exploratory analysis of the September 11 events using the data from the GCP’s global RNG network in a combined “block” form of 15-minutes in duration, which might allow for long-spanning trends in the data over time to be seen. In this web paper, I provide a summary of the results of my exploratory analysis, along with some personal speculation and interpretation on what they could mean for an ostensible “global consciousness.” (Note: in order to better understand the results, it might be helpful for readers unfamiliar with PK and field consciousness research to first read an illustrated introductory web paper on the GCP written by the author, which can be found by clicking here.)

Formal Analysis 1: Cumulative Deviation from Expectation

The formal global consciousness test prediction made by the GCP specified a timeperiod over which it was thought that any PK-based global consciousness effects would be seen in the data produced by 37 of the 38 global RNGs (which are also called “EGGs;” the two terms “(global) RNG” and “EGG” are synonymous and will be used interchangeably here) actively running in the network that fateful day. The timeperiod specified began 10 minutes before the first plane crashed into the North Tower of the World Trade Center (WTC) and ended 4 hours after that, so that the timeframe over which the global consciousness hypothesis would be formally tested ran from 8:35 A.M. - 12:45 P.M. Eastern Daylight Time (EDT). This timeperiod covers all the terrorist events and adds about two and a half hours of their aftermath.

For this test prediction analysis, I followed the standard GCP procedure of working the data using Chi-square values cumulatively combined over time to represent a measure of the overall deviation, or departure, from the expected level of random behavior among the RNGs (Nelson, 2001). Normally, based on the inherently unpredictable nature of a truly random number generator, we would expect the behavior of the global RNGs to be completely different from each other, with their deviation from chance expectation at exactly zero. It seems that when world events occur that greatly affect us, the global RNGs tend to move away from randomness and show anomalous patterns in the data that are beyond chance expectation. This is what the Chi-square technique is intended to measure quantitatively, and since pictures speak a thousand words (especially true in the case of 9/11), it is also possible these anomalous patterns graphically through the technique. Thus, Figure 1 shows my result for the formal prediction period in 15-minute block resolution.

Figure 1. Cumulative deviation of Chi-Square graph of the formal prediction period in 15-minute blocks, 8:30 A.M. - 12:45 P.M. EDT, September 11, 2001. The blue curved arc shows the location of statistical significance (odds of 20 to 1 against chance) as time passes.

As can be seen from the graph, there is a rather steady downward trend that starts soon after the South WTC Tower was hit by the second hijacked plane. Throughout this downward trend there seem to be several instances in which the data line slips below the blue statistical significance arc, indicating that the cumulative data during these times were statistically unlikely to be due to chance alone, and may instead be due to the sudden emotional reactions of people around the world as they begin to draw their mass attention towards the live television and radio news broadcasts of the events, and watch or listen in stunned disbelief and confusion as they begin to unfold. The trend then starts to level out into a random walk (the horizontal zigzagging up-and-down behavior we normally expect to see in RNGs, which is also sometimes in slang terms called a “drunkard’s walk,” since it seems to resemble the path of an intoxicated individual when he or she is given the “walk a straight line” DWI test) from about 10:00 A.M. to 12:00 P.M. EDT, after which there is a strong upward inflection back towards the expectation line of zero. The overall statistical result has Chi-square = 620.99, 641 df, p = .71, which is not statistically impressive from a perspective of 15-minute periods, but the fact that the data line had gone below the blue significance arc several times during the course of the events suggests that there were smaller trends present in the data. Taking the period from when the graph begins at 8:30 A.M. up until it “peaks” and reaches its lowest value (throughout which the line went below the blue arc), we see that there are slight post hoc statistical hints of a significant negative deviation (Chi-square = 346.98, 400 df, p = .973, associated z = -1.92, p = .027). Possible signs of a deviation can be more clearly seen in the formal GCP result in a second-by-second resolution, which is shown in Figure 2 below.

Figure 2. Formal GCP result for the formal prediction period in seconds resolution, 8:35 A.M. - 12:45 P.M. EDT, September 11, 2001. The red curved arc shows the location of statistical significance (odds of 20 to 1 against chance) as time passes. (Source: Global Consciousness Project)

It is interesting that in seconds resolution, there is no distinct and immediate “reaction” by the global RNG network during the major terrorist events, with the data showing a random walk. My corresponding 15-minute result in Figure 1 shows a strong negative trend after the South WTC Tower was hit, which suggests that the reactions may have been more “spread out” in time, making them too wide to be visible in seconds resolution. Both graphs also show a strong trend at around 12:00 P.M., suggesting that the EGGs had started to become less “noisy” and more non-random in a way that was not based on expected chance behavior, with this being significantly so in the case of the formal GCP result shown in Figure 2 (Chi-square = 15332, 15000 df, p = .028)(Nelson, 2002). Thus, both graphs together suggest there were some slight indications of an anomalous effect appearing in the RNGs around the immediate time of the tragedy. In order to be more certain of the possibility that there is an anomaly in the data, however, it became useful to explore a much broader range of time surrounding the events.

Context Analysis: Examining Long-Term Effects of the Attacks

Given the sheer magnitude and importance of its events, it becomes rather important and subjectively meaningful to examine the larger context of the disaster, which may also help us to learn more about the kind of impact such events tend to have on the global RNG/EGG network. Figure 3 shows the context of the entire day in Eastern time, with the terrorist attacks noted with small vertical tickmarks along the pink expectation line of zero deviation. The time resolution is again in terms of 15-minute intervals.

Figure 3. Context graph showing the results for the entire day in Eastern time, 15-minute resolution, September 11, 2001. The terrorist attacks are indicated by the small vertical tickmarks along the pink expectation line of zero. The red curved arc shows the location of statistical significance (odds of 20 to 1 against chance) as time passes for the extreme deviation at around 7:30 P.M. EDT.

The trend for the entire day is quite remarkable, particularly the very strong and steep positive trend that begins at around 7:30 P.M. EDT. It is clear that this trend is extremely deviant based on how rapidly it departs from chance expectation (as measured by the area below the red arc) over time. In examining the way the data is distributed statistically over time, I found nothing to indicate that any of the RNGs were malfunctioning and producing “bad data” at this time, so this suggests that the data are being strongly “driven” by something external to the network. What is just as remarkable as this strong trend at 7:30 is the trend that occurs before it, in the early morning hours of September 11. It can be seen that the data from midnight up until around 6:00 A.M. also steeply increase with time, peaking at around 5:30 A.M. This suggests that there might have been sort of a “pre-response” by the RNG network prior to the onset of the events in New York and Washington (this will be looked at further below). Thus, there may have been some especially notable moments throughout the course of the day.
Figure 4 takes a broader contextual view, focusing now on the three days surrounding September 11. The length of the day is measured in Universal Coordinate Time (UTC), the time that the RNGs are synchronized to, which is 4 hours ahead of Eastern time.

Figure 4. Context graph showing the three days surrounding the events in Universal Coordinate Time (UTC), 15-minute resolution, September 10 - 12, 2001. The terrorist attacks are indicated by the small group of vertical lines along the pink expectation line of zero.

It can be seen that while September 10 starts with a consistent downward trend that lasts nearly the entire day, there is an inflection close to the beginning of September 11 that leads into a rather consistent positive trend that lasts well into September 12.
  Moving the contextual view out quite a bit further, Figure 5 shows the full week surrounding September 11 in UTC-measured days.

Figure 5. Context graph showing the full week surrounding the events in UTC time, 15-minute resolution, September 7 - 14, 2001.

It is clear that the image does indeed speak a thousand words, seemingly thrusting before us the powerful, lingering effect of the disaster in visual form. While there are only random fluctuations in the network as expected on September 7, 8, & 9 (indicated by the horizontal zigzagging behavior of the data line around the pink expectation line of zero), there are clear indications that the data on September 11, 12, & 13 are anything but random, appearing to be highly structured, coherent, and linear, all of which are things we would rarely (if all) expect to see in purely random data for long periods of time. The trend is in fact so strong that it is residual, lasting for nearly 2 days after the attacks! The positive sloping trend seen in Figure 5 is consistent with the GCP’s own finding over the same timeperiod in seconds resolution (Nelson, 2002; Nelson, Radin, Shoup, & Bancel, 2002), which is displayed in Figure 6 below.

Figure 6. Graph of the GCP context examination of the entire week surrounding September 11 in UTC, seconds resolution, September 7 - 15, 2001. The small arc shows where the location of statistical significance (odds of 20 to 1 against chance) lies as time passes. (Source: Global Consciousness Project)

Both graphs display very distinct signs that the data following the attacks were clearly removed from chance expectation and were extreme in nature, in much contrast with the random data in the days prior to September 11. It turns out that this powerful and lingering trend, which lasts about 51 hours total, is so statistically unlikely to occur in purely random data that we would only see a trend just like it occurring by chance alone once every 2,300 days (Nelson et al., 2002)! Informally, this extreme two-day deviation following the attacks on September 11 remains to be the largest randomness deviation anomaly ever observed in the now 5-year history of the GCP, and so far there is nothing else in the entire GCP database like it.
Figure 7 shows a closer look at the main part of the trend seen in Figure 5, covering the day of the attack and the two days following.

Figure 7. Context graph showing a close-up of the strong positive trend seen in Figure 5 in UTC, 15-minute resolution, September 11 - 13, 2001. The terrorist attacks are indicated by the small group of vertical lines along the pink expectation line of zero.

The graph again shows that the slope of the trend is positive and nearly linear in its slope, becoming most strong at around 7:30 P.M. EDT (23:30 UTC), and lasting into September 12 and 13. Combining the data from these three days post hoc, I find that the result is significant overall in 15-minute resolution (Chi-square = 10456.12, 10210 df, p = .043), and still remains consistent with the GCP finding (Nelson, 2002; Nelson et al., 2002). Thus, the consistency of this finding across independent studies further adds support to the likelihood that the result is not a fluke or an artifact, and that it might represent some indirect indication of a strong “movement” within the global RNG/EGG network by some external source.

Formal Analysis 2: Standard Variance Among the EGGs

The second formal test prediction made by the GCP called for looking at the degree to which the individual z-scores from the 37 global RNGs/EGGs varied across September 11. Like the Chi-square values, this variability should theoretically expected to be near zero, indicating little change at all from chance expectation among the RNGs. To examine this, I applied the standard measure of variance to the 15-minute data, which is actually pretty closely related to Chi-square (with only slight differences statistically), so the results should be consistent with the Chi-square results. Figure 8 shows this variance measure applied to the early morning data from September 11.

Figure 8. Graph of the standard variance applied to the early morning data in 15-minute resolution, 2:00 - 9:30 A.M. EDT, September 11, 2001.

It can be seen that the result corresponds exactly with the trend found over the same timeperiod in Figure 5, indicating the relatedness of the two measures. This graph emphasizes the early morning “spike” first seen in Figure 5. Taking this early morning rise on its own, it has an estimated probability of .025 (i.e., odds of about 39 to 1 against chance occurrence) based on its equivalent Chi-square result, and may therefore indirectly suggest a sort of “pre-event response” that peaked at around 5:30 A.M. EDT, nearly 3 whole hours before the first plane hit the World Trade Center! This finding also replicates those that were reported by Dean Radin (2002), who found a similar positive trend prior to the attacks, and Richard Shoup (2001), who found a “spike” in the data at around the same time I did, which had odds of about 131,000 to 1 against chance expectation! Thus, there is some reason for thinking that this early morning “premonition” response by the global RNG/EGG network is not a fluke or a statistical artifact, either.

Autocorrelation and Basic Inter-EGG Correlation Assessment

Another part of the analysis explored in a basic manner the degree to which the global RNGs/EGGs were behaving similar to each other in a correlated fashion during the formal prediction period. This was based on an independent finding by Peter Bancel (2001, in Nelson et al., 2002), in which he had looked at how predictable the RNGs were relative to one another over the course of time, in a sense, using a measure called autocorrelation. Figure 9 shows Bancel’s result.

Figure 9. Graph of Peter Bancel’s autocorrelation analysis of the data for September 11, 2001, which is plotted along with 10 days of control data for comparison. (Source: Global Consciousness Project)

The graph of Bancel’s analysis shows that from moment-to-moment up to a period of one to two hours on the morning of 9/11, the devices within the global RNG/EGG network were dependent on each other (whereas they should not be at all in purely random data), which suggests that there was a pattern of order among the random data of the RNGs on September 11 when in fact there should not have been one at all.
            Early on, Bancel indicated that this predictability in the RNGs on September 11 might have been driven in part by a large deviation from chance expectation that collectively occurred in the z-scores between 9:50 and 11:50 A.M. EDT. In attempt to verify that finding in 15-minute resolution, I plotted the cumulative deviation of the individual 15-minute z-scores, which could serve as a basic indicator of intercorrelation among the RNGs if the z-scores cumulatively sum in a positive trend of deviation over a long period. Figure 10 shows the result of this.

Figure 10. Graph of the cumulative deviation of the 15-minute z-scores over the formal prediction period, 8:30 A.M. - 12:45 P.M. EDT, September 11, 2001.

The resulting graph does indeed indicate that, while there is random walk during the major events, there is indeed a strong trend between 9:50 and 11:50 A.M. that appears in the 15-minute data. An estimate of the odds that this 9:50 to 11:50 trend on its own is not due to chance alone (as given by Bancel) is about 400 to 1 (z = 3.71, p = .0025)! The result further stands out as a strong peak within the context of whole day, which is shown in Figure 11 below.

Figure 11. Graph of the cumulative deviation of the 15-minute z-scores for the full day in Eastern time, September 11, 2001. The terrorist attacks are indicated with vertical tickmarks along the expectation line of zero.

           This is consistent with the GCP finding and Bancel’s suggestion, and seems to suggest that there was a high degree of intercorrelation among the EGGs around the time that the World Trade Center towers collapsed. In other words, it appears that the behavior of the global RNGs may have been linked in a strange way, when in fact they should have been expected to act differently from each other.

Odds Ratio Analyses

Another way it is also possible to assess the degree to which the global RNGs/EGGs “reacted” together in unison is to combine them into a single measure of their collective deviation from expectation (represented by a Stouffer’s Z-score), and then determining the probability of how likely this is due to chance alone through an odds ratio. This was done over every second of the entire day of September 11 in Eastern time, and Figure 12 shows the odds against chance for the resulting Z-scores over time.

Figure 12. Graph of the resulting two-tailed odds against chance for the second-by-second Stouffer’s Z-scores over the course of the day in Eastern time, September 11, 2001. The small group of five blue vertical lines that are of equal length in the bottom center of the graph indicate the terrorist attacks.

As can be seen, there were interesting values around 3:00 and 5:00 A.M., which seem to be consistent with the variance change in Figure 8, but clearly the most interesting of all is the very tall spike in the middle of the graph, occurring at 10:12:46 A.M., about 14 minutes after the South Tower of the World Trade Center had collapsed. This was a moment in which the EGGs were so highly correlated with each other that it was associated with a z-score of 4.80, a value so extreme that it has odds against chance of over a million to one (p = 7.93 x 10^-7)!
Although a value like this is incredible, one like it is expected to occur by chance alone every 15 days or so, and analysis by the GCP reveals that one does (May & Spottiswoode, 2001; Nelson, 2002; Shoup, 2001). So it is perhaps not too surprising that one appeared in the data, since it could be just a chance fluctuation, although it is somewhat odd as to why it appeared in the middle of the attacks. Some of the GCP researchers suggested that it was merely a fortuitous event (May & Spottiswoode, 2001), although from a personal viewpoint I’m not so sure about this and remain neutral.
I also examined the degree to which the variance among the global RNGs changed daily over the entire week surrounding September 11 by representing the variance change in the form of a 15-minute block z-score combined across all the RNGs, and determining its associated odds against chance. Figure 13 shows the resulting z-scores over the period from September 7 – 13, 2001, with each of the “0”s on the graph indicating midnight in Universal Coordinate Time. In this kind of graph, any scores that are within the range of +2 and -2 are likely to be due to the noise found within the global RNG/EGG network, while any scores outside of this range are interesting from a statistical approach.

Figure 13. Graph of the z-scores related to the daily change in variance among the RNGs for each 15-minute period of the day, over the course of the week of September 7 – 13, 2001. The zeros along the x-axis indicate midnight UTC of each day. Any scores outside the range of +/- 2 on this graph are statistically interesting.

It can be seen from looking at the graph that the largest daily changes in variance (both positive and negative) among the EGGs over the entire week had both taken place on September 11. The largest negative variance change had occurred at 13:00 UTC (9:00 A.M. EDT), and had a z-score of -2.87 (p = .002). Figure 14 shows the odds against chance (one-tailed) associated with this z-score and other negative ones seen throughout the week.

Figure 14. Associated one-tailed odds against chance for the negative z-scores shown in Figure 12. The zeros along the x-axis indicate midnight UTC of each day.

It should be noted that this block z-score is the collective result for data across the 15-minute period from 9:00 – 9:15 A.M., so it is not necessarily indicative of a “spike” exactly at 9:00 A.M., but rather is a result from combining data associated with the time of the second plane crashing into the World Trade Center towers and the time after it. Its odds are in the range of about 499 to 1 against chance expectation.
The largest positive variance change had occurred later in the day, at 23:30 UTC (7:30 P.M. EDT), and had a z-score of 3.19 (p = .0007). Figure 15 shows the odds against chance (two-tailed) associated with the z-scores throughout the week.

Figure 15. Associated two-tailed odds against chance for the  z-scores shown in Figure 12. The zeros along the x-axis indicate midnight UTC of each day.

It is obvious from the graph that this evening change in variance at 7:30 P.M. is the most extreme of the entire week, and it is also consistent with the very steep and powerful increasing trend in the full day result shown in Figure 3. The result is so extreme that it has odds of about 1,428 to 1 against chance expectation! These results are also very similar to the results of a similar analysis of the daily variance change by Dean Radin (2001), who also found “spikes” like this in his analysis.

Exploratory Analysis: Heroic Sacrifice Aboard Flight 93

Of the four commercial airliners hijacked and taken over by the terrorists that fateful morning, only three of them had actually struck their intended targets. The fourth plane, United Airlines Flight 93 out of Boston, had instead crashed in a field in western Pennsylvania between 10:07 and 10:30 A.M. EDT. As we all know now, based on the accounts from people who had taken cellular phone calls from individuals they knew aboard the flight, the failure of the plane to reach its target was due to a heroic sacrifice by a brave group of passengers, who had apparently attempted to take down the terrorists aboard and foil the hijacking at great risk to themselves. As a way to honor the courage and bravery of these passengers, it was thought that this event deserved an exploratory analysis. It was thought that, since an intense physical struggle was likely to have ensued, as well as the great panic that may have erupted among the other passengers as a result of it, the event would be emotionally-gripping enough to have had a notable effect on the global RNG/EGG network. So to test for this, I used the same Chi-Square method as in the first analysis I did over a six-minute period of second-by-second data.

Figure 16. Graph of cumulative deviation of Chi-square for the six-minute time-period presumed to be associated with the heroic struggle aboard Flight 93, 10:07 – 10:13 A.M. EDT, September 11, 2001. The blue arc indicates the location of statistical significance (odds of 20 to 1 against chance) lies as time passes.

Figure 16 shows the result. As can be seen, the data line fluctuates and zigzags greatly throughout the six-minute period, again possibly reflective of the panic, confusion, and fear that may have erupted aboard Flight 93. There clearly a rather strong deviation in the latter part of the timeperiod, which, in some meaningful way, could be a “signature” of the struggle. Then there is a sharp peak towards the end, where the data are driven by that large “spike” seen above in Figure 11 at 10:12:46 A.M. The rest of the data line began to fall moderately downward beyond the graph up until 10:30 A.M., which shows a negative signature similar to a few other GCP analyses relating to airplane crashes.

Exploratory Analysis: Effect Size for September 11 with Location

Figure 17 shows a graph of a relative measure of the magnitude of the effect of the September 11 deviations as they are distributed across the entire global RNG/EGG network, which is terms of 15-minute block Stouffer’s Z-scores.

Figure 17. Graph of the effect size for September 11 in terms of Stouffer’s Z-scores. The time scale is in 15-minute blocks, Eastern Standard Time.

As we might expect, the largest effect was centered in North America, having a z-score of 2.21. Exploring further this possibility of effect size as a function of geographical location, I found that the effect was indeed focused more in the Western Hemisphere (z = 1.73) as compared to the Eastern Hemisphere (z = - .502), and that with the United States, the effect was centered most in the Midwest (z = 2.52), moderately in the East Coast (z = 1.72), and least in the West Coast (z = - .68). This is similar to the effect size explorations with distance carried out by Dean Radin (2001) and Roger Nelson (2003), who both found similar effect sizes distributed in nearly the same areas.

Discussion and Conclusion: What Could It All Mean?

In all, the results of the exploratory analysis of the 9/11 GCP data done by myself and those done by other independent researchers within the GCP seem to collectively provide some indication of small but statistically measurable randomness deviation anomalies being present within the data recorded by the global network of random number generators on that terrible day. In attempting to make sense of the anomalies, there are a few interpretations we might consider.
The first (and perhaps most important) interpretation is that the anomalies have a rational explanation. This interpretation implies that the anomalies were naturally caused by extreme conditions within the environment around the time of the terrorist attacks, which could include sudden electrical disruptions or power surges around urban areas as a result of the disaster, and increased electromagnetic interference from growing use of televisions, radios, and cellular phones with time. While this interpretation is plausible, it is rather highly unlikely to be true in this case for 4 reasons: 1.) the RNG anomalies are spread out across the network, as basically indicated by the effect size exploration and inter-EGG correlation results, and any extreme condition would probably have only been localized to major metropolitan areas such as New York City, so it is not feasible that a localized extreme condition could have affected the entire worldwide network, 2.) there is evidence to suggest that some of the RNG anomalies occurred several hours before the attacks began in New York and Washington, 3.) the RNG are specially designed to block out electromagnetic interference; and 4.) a basic examination of the natural changes in electromagnetism due to power usage by Dean Radin (2002) indicated no differences between day and night in the power usage around September 11. For these reasons, the idea that the September 11 RNG anomalies have natural causes from the environmental conditions that day does not seem likely.
We must also consider the possibility that results are due to statistical error or chance expectation (Scargle, 2002). The former does not seem very likely either, given that the results have reproduced among a number of independent researchers using the same data, and my own exploratory analysis looked at the possibility that the results could have been produced through the statistical “inflation” of purely random data such that it looks like it is non-random and anomalous, but really isn’t. Given that the same statistically significant trends were found across time ranges from second-by-second to 15-minute block periods, there is little indication that these results are statistical artifact. The latter possibility of chance expectation is still feasible and can never totally be ruled out, but the results do argue against it.
Thus, although we cannot conclude that the results show definitive evidence for it, the most sound explanation is perhaps we detected the psi “signature” of a developing global consciousness being shared by the minds of people around the world as they turned their attention toward the terrorist attacks and responded with similar emotions to them, which had the ability to momentarily disrupt the natural behavior of random physical systems around the world in a very powerful way. It should be noted that the global consciousness effects seen in the global RNG/EGG network possess characteristics that are very similar to those of PK (or “mind over matter”) effects, and suggests that the two might have a common basis. The clearest possible indications of this that come from the 9/11 results include that RNGs all over the world were affected, even far from the source of the main events (Dunne & Jahn, 1992); the trends seemed to be correlated with the severe and widespread emotional response to the attacks, which is consistent with a growing indication of a possible correlation between PK effects and emotional expression (e.g., Bierman, 1996; Blasband, 2000); and perhaps most intriguing of all, the response of the RNGs seemed to begin several hours before the attacks began to unfold, which seems to be consistent with the evidence for retroactive (i.e., “backwards acting in time”) PK effects on RNG data (e.g., Bierman, 1998; Schmidt, 1976, 1993). The effects also seem to be similar to the recent experimental findings on presentiment, in which an anticipatory-related change in electrodermal activity is observed prior to viewing emotionally-charged pictures, especially those that are of a violent and disturbing nature (Bierman & Radin, 1999), suggesting that this early morning anomaly may reflect sort of an anticipatory, emotion-related PK response to the terrible events by a supposed global consciousness. This would also be similar, and may perhaps even be in some way related to the many subjective reports of premonitions relating to the devastating events in the hours and days before September 11 (Carpenter, 2002). Thus, there is still much we do not understand about a supposed global consciousness, but the effects of this far-reaching historical event have given us several hints about its nature.
Almost important as any kind of scientific evidence is the importance of meaning. The bottom line is, if a global consciousness does exist and it formed among all of us that morning, then it appears to have been greatly moved by the events of September 11, 2001. The results seem to reflect the mass waves of depression and fear that seem to have engulfed the entire world and lingered around New York City for weeks following the attacks, as was indicated in a recent survey study of depression and PTSD cases in the Manhattan area around Ground Zero (Galea et al., 2002). From this view it seems that, in a sense, Gaia also appears to have been wounded by the terrorist acts. If global consciousness and PK effects are somehow related, then the results from the global RNG/EGG network on 9/11 may also indicate that the power of a collective, emotion-related PK effect is stronger than we ever imagined. Only further research will shed more light about the mysterious nature of global consciousness.

Acknowledgments

I must thank Dr. Roger Nelson of the Global Consciousness Project for technical advice and comments on my GCP work, Steven Kindsvater-Novak of Fort Hays State University for all the past discussions on psi, consciousness and causality; and my dear friend Billie Pyzel for her support of and personal comments on my GCP work. This web paper is a summarized version of a formal paper describing the results of the exploratory block analysis on the 9/11 data (Williams, 2003).

References (in order of text citation)

Nelson, R. D. (2001). Correlation of global events with REG data: An Internet-based, nonlocal anomalies experiment. Journal of Parapsychology 65(3), September. pp. 247 – 271.
Nelson, R. D. (2002). Coherent consciousness and reduced randomness: Correlations on September 11, 2001. Journal of Scientific Exploration 16(4), Winter. pp. 549 - 570.
Nelson, R. D., Radin, D. I., Shoup, R., & Bancel, P. A. (2002). Correlations of continuous random data with major world events. Foundations of Physics Letters 15(6), December. pp. 537 - 550.
Radin, D. I. (2002). Exploring relationships between random physical events and mass human attention: Asking for whom the bell tolls. Journal of Scientific Exploration 16(4), Winter. pp. 533 - 547.
Shoup, R. (2001). EGG salad - Comments on the GCP EGG data for September 11, 2001. Boundary Institute Technical Report, November 6. Available over the Internet at: http://www.boundary institute.org/articles/EGG_Salad.pdf. Downloaded November 12, 2001.
May, E. C., & Spottiswoode, S. J. P. (2001). Global Consciousness Project: An independent analysis of the 11 September 2001 events. Available over the Internet at: http://noosphere.princeton.edu/papers/Sep1101.pdf. Downloaded November 12, 2001.
Radin, D. I. (2001). Global Consciousness Project analysis for September 11, 2001. Global Consciousness Project draft report, September 21. Available over the Internet at: http://noosphere.princeton.edu/dean/wtc0921.html. Downloaded October 2, 2001.
Nelson, R. D. (2003). What kind of events have effects? Global Consciousness Project exploratory analysis, February 28. Available over the Internet at: http://noosphere.princeton.edu/categories.html. Downloaded May 13, 2003.
Scargle, J. D. (2002). Commentary: Was there evidence of global consciousness on September 11, 2001? Journal of Scientific Exploration 16(4), Winter. pp. 571 - 577.
Dunne, B. J., & Jahn, R. G. (1992). Experiments in remote human/machine interaction. Journal of Scientific Exploration 6(4), Winter. pp. 311 - 322.
Bierman, D. J. (1996). Exploring correlations between local emotional and global emotional events and the behavior of a random number generator. Journal of Scientific Exploration 10(3), Autumn. pp. 363 - 373.
Blasband, R. A. (2000). The ordering of random events by emotional expression. Journal of Scientific Exploration 14(2), Summer. pp. 195 - 216.
Bierman, D. J. (1998). Do psi phenomena suggest radical dualism? In S. R. Hameroff, A. W. Kazniak, & A. C. Scott (Eds.) Toward a Science of Consciousness II: The Second TUSCON Discussions and Debates (pp. 709 - 713). Cambridge, MA: MIT Press/Bradford.
Schmidt, H. (1976). PK effect on pre-recorded targets. Journal of the American Society for Psychical Research 70(3), July. pp. 267 - 291.
Schmidt, H. (1993). Observation of a psychokinetic effect under highly controlled conditions. Journal of Parapsychology 57(4), December. pp. 351 - 372.
Bierman, D. J., & Radin, D. I. (1999). Conscious and anomalous nonconscious emotional processes: A reversal of the arrow of time? In S. R. Hameroff, A. W. Kazniak, & D. J. Chalmers (Eds.) Toward a Science of Consciousness III: The Third TUSCON Discussions and Debates (pp. 367 - 385). Cambridge, MA: MIT Press/Bradford.
Carpenter, J. (2002). Guest editorial: Why parapsychology now? Journal of Parapsychology 66(4), December. pp. 339 - 342.
Galea, S., Ahern, J., Resnick, H., Kilpatrick, D., Bucuvalas, M., Gold, J., & Vlahov, D. (2002). Psychological sequelae of the September 11 terrorist attacks in New York City. New England Journal of Medicine 346(13), March 28. pp. 982 - 987.
Williams, B. J. (2003). Consciousness and global random systems: An independent exploratory block analysis of the correlations on September 11, 2001. Submitted to the Global Consciousness Project June 26, 2003.

Created: August 12, 2002
Revised: September 10, 2003