An attempted replication of the PRL ganzfeld research - Psychophysical Research Laboratories - Autoganzfeld, part 2

Richard S. Broughton

The work of Charles Honorton and his collaborators at the Psychophysical Research Laboratories (PRL) has established the ganzfeld technique as one of the most reliable methods of eliciting evidence of ESP in the laboratory (Bern & Honorton, 1994; Honorton et al., 1990). This has, in turn, led to new, independent ganzfeld experiments that attempt to replicate the original findings and to extend our understanding of the underlying processes. Among the principal efforts underway are those at the University of Utrecht (Bierman, Bosga, Gerding, & Wezelman, 1993; Van Kampen, Bierman, & Wezelman, 1994) and at the University of Edinburgh (Dalton, 1997; Dalton et al., 1994).

In the early 1980s the Institute for Parapsychology began a similar program of ganzfeld research rising a nonautomated procedure. The overall success rate in the Institute's manual series has been lower than that obtained by PRL researchers, but essential features of the PRL work have been replicated in our program (Kanthamani & Broughton, 1994).

Following the closure of the PRL, Honorton donated the equipment that had been used in the automated ganzfeld series to the Institute for Parapsychology where it was reassembled and installed by Kathy Dalton. Dubbed Autoganzfeld II, this installation essentially reproduces the testing system used at PRL, though the physical arrangement of the laboratory is very different. The soundproof room used at PRL was not moved to the Institute.

From October 1993 through March 1996, we conducted formal collection of ESP-ganzfeld data using Autoganzfeld II. The principal aims of this project were as follows:

1. An independent replication of the PRL work with a completely different team of researchers in a different location. (We recognize that it is not a 100% independent replication since we are using the same hardware and software.)

2. An extension of PRL's and our research into personality variables and ganzfeld success using a more detailed personality assessment, the NEO-PI-R (Costa & McCrae, 1992), which has been adopted by other ganzfeld researchers.

3. To examine a subset of first-time participants consisting of emotionally close sender-receiver pairs.

4. To develop content-analysis methods that will identify patterns in the participant's mentation report that may predict adjustment to the ganzfeld test situation and its success or failure as an ESP test. This aspect of the project is continuing and will be reported in the future.

Following the approach of Honorton, Autoganzfeld II was planned as an ongoing project composed of fixed-length series of 50 sessions each. As a first phase of the project, we decided to conduct a formal replication of the PRL work. This would consist of three series (150 trials) with first-time participants ("novices," in PRL terminology). Within the replication, one series would be confined to "emotionally close" sender-receiver pairs. When it became apparent early in the program that there were a number of individual volunteers for whom no sender was readily available, we created a clairvoyant (no sender) series. Finally, to accommodate participants who wished to be tested additional times, we created a general series, for which the stopping point was to be when the formal replication series were completed. Neither the Clairvoyant series nor the General series was considered part of the PRL replication.

Autoganzfeld II was terminated in 1996 with three replication series, one clairvoyant series, and 8 sessions in the general series. Those data are reported here.(1)

METHOD

Automated Testing System

A detailed description of the autoganzfeld testing system is given in Honorton et al. (1990, pp. 100-110). Here we will present only a summary of our version of that system, focusing on the principal differences between our system and the original.

The original automated ganzfeld testing system used an Apple II computer to control a VCR that showed the target material to the sender and later presented the target pool to the receiver for judging. The computer programs paced the experiment by timing the various parts and prompting the experimenter when actions were needed. The programs also handled data storage, which was backed up with paper records.

Following a period of storage, the components were reassembled at the Institute for Parapsychology by Kathy Dalton. After confirming that the equipment was functioning, Dalton installed the system in the rooms that were to be used at the Institute. R. S. B. reviewed the computer programs and made minor changes to accommodate differences between the PRL setup and the present one.

Autoganzfeld II uses essentially the same hardware and software to control the ganzfeld experiment as outlined in Honorton et al. (1990), with certain exceptions. The RNG was relocated from an expansion chassis to the Apple II itself, and the expansion chassis along with a silent printer was removed from the system. Appropriate changes were made to the software.

The target pool remained the same as that used at PRL, consisting of 80 static targets (still photos and art prints on video) and 76 dynamic targets (film clips). One target pack has always been excluded due to a fault (see Honorton et al., 1990, p. 119n).

The audio equipment (cassette decks, amplifiers, etc.) for presenting the relaxation instructions and pink noise, as well as for recording the receiver's mentation, are either the same pieces of equipment or functional equivalents.

Layout of Testing Rooms

The layout of the experimental rooms is very different from PRL's. We do not use an acoustically isolated chamber for the receiver. Instead, we isolate the sender from both the receiver and the experimenter by placing the sender in a room on the opposite side of the building. The sender's room is approximately 12m from the experimenter's room and 15m from the receiver's room. A minimum of four doors and several passages with right angle turns separate the sender from the experimenter, and the receiver is separated by one additional door and room. All doors are kept shut during the experiment. Also, the sender's room is adjacent to the office of R. S. B. who was always aware of sessions and monitored the room for any unusual activity or noise.(2) The rooms for the experimenter and receiver are in a wing that had been added to the original structure and do not share a common floor structure even though they are on the same level of the building. Figure 1 illustrates the layout of the research floor of the Institute.

None of the rooms are specially prepared for acoustic isolation, nor were any formal sound transmission measurements taken. Distance and the active monitoring of both sender and receiver locations provide an adequate degree of isolation between the sender and both the experimenter and receiver. Unauthorized egress by the sender or any sounds emanating from the sender's' room during the ganzfeld session would cause the trial to be discarded.(3)

A closed-circuit television enables the experimenter to monitor the room adjacent to the receiver's room which provides the only access to the receiver's room.

The experimenter's room contains all of the equipment that was located on the experimenter's console in the PRL setup, with the exception of the VCR that projects the target pictures. In our setup, the VCR is located in a closet in a separate office so that none of the noises associated with the tape movement can be heard in the experimenter's room.

The receiver's room is 2.5m x 9m, located on the second floor in the quietest corner of the building. The receiver relaxes in a comfortable recliner and the red light for the ganzfeld is provided by two 100-watt red flood lamps mounted high on the walls about 2m from the receiver's face. The television that presents the judging pictures is located in the corner faced by the seated receiver. During the actual ganzfeld period, the room is illuminated by the red flood lamps alone.

The sender's room is about 4m x 2m, located in the opposite corner of the building as described above. It contains a reclining chair and the television used for showing the target picture. Video and audio lines are routed from the experimenter's room via the attic which is immediately above.

Subjects

Over all series there were 89 males and 110 females ranging in age from 15 to 78 years (mean = 33.9, SD = 13.3). Participants were recruited primarily from the local community through advertisements, local publicity for the Institute, and word-of-mouth. Since one series was devoted to emotionally close sender-receiver pairs, we specifically encouraged emotionally close pairs to participate, both in advertising and by word-of-mouth.

Participant Orientation

After initial contact (typically by telephone or visit), participants are sent a packet containing a welcoming letter, some information about ganzfeld experiments in general, a participant information form (PIF) which is based on that used at PRL, and Form S of the NEO-PI-R. When participants indicate they are coming with a partner, both are sent the complete packet for information and completion. Most participants complete the questionnaires at home and bring them along to the test session. Occasionally a participant scheduled at short notice will complete the questionnaires in the lab before the test session. Only a very few participants did not complete the questionnaires beforehand, usually because of time constraints. One participant mailed the questionnaires in later and two other participants never returned their questionnaires.

For the test session, participants are greeted by the experimenter and made to feel welcome. Further introduction to the ganzfeld approach to ESP testing is provided as appropriate. Participants are shown the testing rooms with an explanation of the roles of the sender and receiver. Participants who arrive in pairs are encouraged to decide for themselves who will be the sender and who will be receiver.

Experimental Procedure

The conducting of the experiment closely follows the procedure described in Honorton, et al. (1990, pp. 105-110). The role of the Target Selector (TS) is usually taken by R. S. B. or another member of the staff, and is only slightly different in that the VCR and video tapes are located in a closet in a different room from the controlling computer and there is no need to place a cover over the VCR.

Another minor difference is in the last stage of the experiment. Since we have no printer attached to the Apple II that controls the experiment, after the participants depart, the experimenter takes the data disk to another Apple II and prints out the paper record of the session.

Experimenters

C. H. A. conducted all but a few of the experimental sessions. R. S. B. conducted the remaining sessions.

Experimental Series

Three series were planned as a replication of the PRL novice ganzfeld program. These were two ordinary "First-Timers" and one "Emotionally Close" first-timers. A fourth series of Clairvoyance (no sender) sessions was created to accommodate participants for whom no sender was readily available. A general series was started for repeat participants. Although this had a nominal total of 50 sessions, we considered this only a pilot series and did not actively recruit participants.

First-Timers (ordinary): This series included all participants who had not had prior experience in the autoganzfeld but who did not fit the category of "emotionally close." Senders in this series were usually friends or lab friends (i.e., a friend who is on the staff), though occasionally a member of the staff previously unacquainted with the receiver served as sender. The planned two series, FT1 and FT2, were completed.

Emotionally Close First-Timers: This series was for sessions in which both the sender and the receiver are emotionally close. This includes happily married couples, parent-child pairs, and siblings. After consultation with a clinical psychologist, we decided not to include newlyweds or courting couples as this is often an anxiety-prone period in a relationship and possibly not conducive to ESP.

The EC series was not designed to be a formal comparison with non-emotionally close sender/receiver pairs. Rather, it was an exploratory investigation using a subset of participants that might provide important clues both to the nature of the ESP interaction and to what classes of participants are likely to succeed in the ganzfeld. A formal comparison would have required, at the least, that both the receiver and the experimenter be kept blind as to whether the sender was the emotionally close partner who came to the session or a complete stranger. We felt that since a primary goal was to replicate the PRL findings, the necessary manipulations - either deception or an elaborate explanation to avoid deception - plus the additional manipulations, would make such a testing situation radically different from the PRL sessions.

The idea that emotionally close people may be especially prone to experiencing ESP between them has always been a feature of spontaneous case investigations (Schouten, 1982); a few experimental investigations have been reported as well. In a free-response experiment that explored agent-percipient relationship and GESP performance (Smart, 1946), significantly positive results were obtained by closely related pairs (twins, married couples, and engaged couples), whereas unrelated pairs gave significantly negative results. Inspired by Stuart's work, Rice & Townsend (1962) designed an experiment to test the hypothesis that emotionally close pairs would score higher in a GESP test than those pairs that were not closely related. A highly significant scoring difference between the two groups was found. The related couples scored positively to a significant degree, whereas the unrelated pairs scored in a significantly negative direction.

The immediate impetus for focusing on emotionally close participants in the ganzfeld grew out of discussions with a supporter of ganzfeld research who felt that the anecdotal evidence for emotional closeness as a facilitator of ESP was very compelling. He also thought that the ganzfeld would be a good vehicle for investigating that relationship.

Near that same time, we were involved in a complete review of the PRL records in collaboration with Daryl Bem. The PRL investigators classified sender-receiver pairs into only three categories: Lab sender, lab friend, and friend. The "friend" category subsumed all closer relationships. By examining the paper records, however, it was possible to identify with a reasonable degree of accuracy a subset of PRL "friends" who conformed to our definition of emotionally close. This subset amounted to 36 sessions with 15 direct hits for a scoring rate of 42 percent. Thus, at least informally, it appeared that emotionally close participants were among the most successful in the PRL series.

A total of 51 EC sessions were run, with the extra session being held in case any of the earlier data sets were unusable. All sessions proved to be usable, so we are reporting the extra session for completeness.(4)

Clairvoyance series: This series was created for participants who wished to participate but could not bring their own partner and for whom it was not possible to supply a lab sender. This was not confined to first-time participants; nonetheless, all but four participants were first-timers. This series was seen as a pilot series, examining whether results could be obtained in the absence of an sender. No formal comparison was intended, and an effort was made to present the clairvoyant session as the "normal" way of doing the procedure, with no reference made to a sender's role. One series of 50 sessions was completed.

General Series: The general series was for participants who had already experienced an autoganzfeld test session in our lab. Virtually all in this series were participants who had enjoyed their first autoganzfeld session so much that they requested additional sessions. Since we were focusing on first-timers, we did not actively recruit for repeat sessions, but we accommodated them as they arose. It had been decided at the outset that the termination of the replication series would dictate the close of the General series. When the replication series terminated, we had 8 sessions in the General series.

Planned Analyses

Overall evidence of ESP was to be assessed by direct hit analysis with the two FT and the one EC series constituting a formal replication attempt. Direct hits were to be used for an examination by sender-receiver pairs, as well as for all other categorical comparisons with target type, practice of mental disciplines, and personal psi experiences identified beforehand as being of interest. Standardized ratings were to be used for correlational analyses using information from the PIF and domain and facet scores from the NEO-PI-R. All but the Belief-in-psi PIF variable were exploratory. Standardized ratings were to be used for a geomagnetic analysis with a negative correlation predicted.

RESULTS

Randomness Checks

Prior to embarking. on this project, a global certification test was performed on the target-generating system. This consisted of extracting the target-generating instructions from the controlling program and imbedding them in a program that generated a large number of autoganzfeld targets in the range of 1-160, excluding the four targets in the faulty pack. In the pre-series test, 62,400 trials were generated and Chi-square tests revealed no consistent departures from the expected uniform distribution.

Ongoing randomness checks of 15,600 trials were made at irregular intervals and there was no evidence of consistent departures from chance expectation.

Direct Hit Results

As shown in Table 1, the replication series produced a total of 40 hits in 151 sessions for a hit rate of 26.5% (z = 0.34, ns). The associated effect size [Pi] = .52 falls below Bem and Honorton's (1994) estimated 95% confidence-interval lower limit of .53 for the general population. The scoring rate falls within the 95% confidence interval (25.5% to 49.5%) for novices in the PRL series (Honorton et al., 1990). Most of the hits were obtained in the emotionally close series (EC), where there were 19 hits in 51 sessions for a hit rate of 37.3% (z = 1.81, p = .035, one-tailed). The second ordinary first-timers series (FT2) produced an unexpectedly low hit rate of 18% (z = -1.33, ns).

[TABULAR DATA FOR TABLE 1 OMITTED]

Despite pervasive folklore among scientists that one should not report interim results, we presented an interim report on this work at the 1995 convention of the Parapsychological Association. In that context, it should be noted that 59 of the replication sessions (45 FT2 and 14 EC) were conducted following the interim report. These contributed only 10 hits to the remainder of the series for a post-interim-report hit rate of 16.9% (z = -1.63). The difference between the 92 replication sessions done before the interim report and the 59 done after it is significant, Fisher's exact p = .04, two-tailed.

TABLE 2

DIRECT HIT RESULTS BY SENDER-RECEIVER RELATIONSHIP, REPLICATION
SERIES

                N Trials   N Hits   % Hits   ES [Pi]   ES h    z

Lab assigned       36        11       30.6     .57       12    0.60
Friend/lab         64        10       15.6     .36     -.23   -1.96
friend
Spouse             21         4       19.0     .41     -.14   -0.87
Parent/child       23        10       53.5     .70      .39    1.72
Siblings            7         5       71.4     .88      .97    2.23

Table 2 presents the direct hit results by sender-receiver relationship for first-time participants. As can be seen from that table, parent/child and sibling sender-receiver pairs produced exceptionally high hit rates, 43.5% and 71.4% respectively. By contrast, another category that prior research indicates should do well, friends, produced an unusually low hit rate of 15.6%. Of the 34 friends pairs that were tested after the interim report, only 4 produced hits (11.8%, z = -2.126).

The replication data base was examined for several other simple effects that have been reported in prior research. Table 3 presents the direct hit results for breakdowns by type of target (static or dynamic), sex of receiver, mental discipline practice by receiver, and handedness of receiver. Contrary to earlier findings, the dynamic targets did not produce a better hit rate than the static targets. In fact, the static targets yielded a promising hit rate of 31.3% while the dynamic targets gave only 23%. The hit rates for the target types were not significantly different, Fisher's exact p = .27. Also in contrast to previous findings, receivers who practiced mental disciplines (yoga, TM, etc.) did not produce a better hit rate. In conformity with prior research, persons who reported personal psi experiences did produce a better hit rate than those who did not report psi experiences, but the difference was not significant. Left-handed receivers yielded a noticeably better hit rate (45.5%) than either right-handed or ambidextrous receivers, but this was not significant, [[Chi].sup.2] = 2.19, df = 2, ns.

TABLE 3

DIRECT HIT RESULTS FOR OTHER BREAKDOWNS, REPLICATION SERIES

                N Trials   N Hits   % Hits   ES [Pi]    ES h     z

Target type

Static              64       20      31.1       .58      .14    1.01
Dynamic             87       20      23.0       .47     -.05   -0.05

Sex of receiver

Male                65       18      27.7       .54      .06    0.38
Female              86       22      25.6       .51      .01    0.02

Mental disciplines

Yes                112       29      25.9       .51      .02    0.13
No                  39       11      28.2       .54      .07    0.31

Psi experiences

Yes                138       37      26.8       .52      .04    0.41
No                  13        3      23.1       .47     -.05   -0.43

Handedness

Right              128       32      25.0       .50      .00    -.09
Left                11        5      45.5       .71      .43    1.20
Ambidextrous        12        3      25.0       .50      .00   -0.28

Results for all series taken together are not reported here, but it can be noted that the addition of the Clairvoyant and General series do not substantially alter the findings for the Replication series.

Correlational Analyses

Using standardized ratings of target and decoys (z scores) in the replication series, we examined the effects of various personality and belief factors on ganzfeld scoring. None of the principal items from the Participant Information Form (PIF) yielded any correlation with z scores. Belief-in-psi, for example, produced a correlation of -.03. Examining the data from series EC1, which appeared to have some evidence of psi scoring, did not produce an appreciably different picture of correlations.

Personality data from the NEO-PI-R were available for 150 of the 151 participants in the replication series. None of the five NEO-PI-R domains (Neuroticism, Extroversion, Openness, Agreeableness, Conscientiousness) showed any significant correlation with z scores. Of the 30 facet scores (six for each domain), only facet E4 (Activity), r = .14, p = .08, and 05 (Openness to ideas), r = .14, p = .08, showed any suggestion of a relationship.

In the emotionally close series, with data from all 51 participants, the NEO-PI-R domain of Conscientiousness was negatively correlated with z scores, r = -.29, p = .04. The positive correlation with facet 05 (ideas) was stronger, r = .35, p = .01. Consciousness facets C3 (Dutifulness), r = -.30, p = .03, and C5 (Self-discipline), r = -.34, p = .02, seem the primary contributors to the negative relationship.

Physical Variables

Numerous studies have demonstrated a weak negative correlation between ESP results and variation in the earth's magnetic field (see, e.g., Persinger, 1993; Persinger & Krippner, 1989; Spottiswoode, 1990). based on this research, we planned to examine ESP (as measured by the z scores) and geomagnetism with the expectation that a similar weak negative correlation would be found. Spottiswoode (1997b) recently demonstrated that the strength of this negative correlation depends on the local sidereal time (LST) where and when the subject is taking the ESP test. (LST is a longitude-like indicator of which particular "slice" of the celestial sphere is overhead. For a more complete explanation, see Spottiswoode [1997a; 1997b].)

For geomagnetic and LST calculations, the time of the trial was set as the start of the subject's mentation period, which was 15 minutes after the recorded start time of the experiment. In most cases, the start time was not recorded by the experimenter since the computer program was known to record the time. However, on auditing the data and inspecting the program code, we found that the computer start time was recorded when the target was randomized, which was a variable time before the actual start of the trial. We compared computer time with hand-recorded time where that was available and determined that the modal number of minutes before the actual start was 15. Thus, where hand-recorded time was available, the designated trial time was 15 minutes after that, and where only computer time was available the designated trial time was 30 minutes after the computer time. Geomagnetic data were the three-hour ap indices available from the National Geophysical Data Center as the "Lenhart" files.

All 209 trials were used in the geomagnetic analysis. Spearman's [Rho] (to allow for the statistical properties of ap) yielded a correlation of [Rho] = -.10, (p = .072, one-tailed).

Spottiswoode (1997a) has observed a nearly fourfold increase in ESP effect size in a two-hour window of LST surrounding 13.5 hours. Subsequently, he discovered that the negative correlation of ap with ESP was greatly increased in roughly the same window of LST (Spottiswoode, 1997b). In the two-hour LST window for maximum effect size, 12.5h to 14.5h, we had 15 trials with [Rho] = -.57 (p = .030, one-tailed). Using the similar two-hour window for maximum strength of the correlation that Spottiswoode derived from his geomagnetic analysis, 11.9h to 13.9h LST, we find 16 trials with [Rho] = -.61 (p = .019, one-tailed).

Figure 2 shows the mean z scores from all 209 trials of this experiment plotted in LST space in the same manner as Spottiswoode (1997a), except that our two-hour sliding window average moved in .25h steps rather than his .10h steps. Note that the data reported in this report are included in Spottiswoode's set but constitute only about 7 percent of his total. Figure 3 shows Spearman [Rho] correlation plotted in LST space in the manner of Spottiswoode (1997b) but with .25h steps as well.

DISCUSSION

The principal aim of this project was to provide a straightforward replication of the PRL ganzfeld data from novice participants. In that aim, this project was not successful. Although the results were in the predicted direction, the obtained effect size was far too low to achieve significance even with the relatively large number of participants employed in this study.

The emotionally close subgroup in the replication that was expected to produce above-chance hitting on the basis of a review of the PRL data, did succeed in replicating this finding with a respectable hit rate of 37.5%. It is immediately apparent from Table 2 that within the EC group, the spouses were not in the same class as the parent/child and sibling pairs. This leads to the speculation that perhaps consanguinity may be important.(6) Combining the blood relatives yields 15 hits in 30 trials for a hit rate of 50% (ES[Pi] = .75, h = .52, z = 9.78, p = .006, two-tailed). While it is tempting to invoke genetics or some other aspect of the familial relationship as the key factor, observations of the actual experiment suggest that the picture may be more complex. Of all the participants in these experiments, the parent/child and the sibling pairs seemed to bring the greatest amount of enthusiasm to the experiment and derived the greatest level of satisfaction when hits were scored. These results certainly confirm the commonsense notion derived from spontaneous case reports that ESP is especially "good" amongst family members, but it will take much additional research to address the usual nature vs. nurture issues. Dalton (1997, p. 130) has recently reported an independent confirmation of the unusually high scoring amongst biologically related sender-receiver pairs in ganzfeld experiments.

Offsetting the good performance of the emotionally close series, however, was unexpectedly low scoring in the second of the two regular first-timer series. Apart from paying homage to the lab lore concerning what happens when one reports data part way through an experiment, there is little to speculate about concerning the decline in scoring during the final year of the experiment. One could invoke a decline in experimenter motivation or simple "burnout," or heightened anxiety to keep up the promising results already reported, and such elements were probably present to a degree. In that context it should be noted that C. H. A. served as the experimenter for virtually all of the 909 sessions over two and one-half years. In the PRL program, multiple experimenters were available and different experimenters took the lead at different times over the years. There is little objective evidence, however, to move any of these suggestions beyond the level of speculation. Bierman (Wezelman & Bierman, 1997) has reported similar declines in ganzfeld results following "exposure" of interim findings and suggests that these may be interpreted as being predicted by yon Lucadou's (1995) model of pragmatic information. We also found no evidence that the post-report decline could be explained by physical variables such as geomagnetism or LST.

Few of the previously reported correlates of high scoring received support in these data. The superiority of dynamic targets over static targets observed by Honorton et al. (1990) clearly received no support in these experiments, with the static targets producing insignificantly better scores. Practicing a mental discipline did not enhance scoring as in earlier work (Kanthamani & Broughton, 1994).

Receivers who reported that they had prior spontaneous psi experiences did do better than those who did not report psi experiences, which conforms to previous findings, but the difference was relatively small and nonsignificant. The notably good scoring by left-handed receivers suggests a possible link with brain hemisphere specialization, but further speculation must await a larger database since there were only 11 left-handers in the replication series. Apparently the PRL data did not include handedness, and we are not aware of any precursors to this finding in the ganzfeld literature.

In the overall replication data base, there was no evidence that the NEO-PI-R domains and facets bore any systematic relationship to psi scoring, an observation also made by Morris et al. (1995). The weak positive relationship between facet E4 (Activity) and psi scores is in line with the study of Van Kampen, et al. (1994) who found that this facet discriminated between hitters and missers. Van Kampen et al. found that Openness facet 03 (openness to feelings) was the best discriminator for hitting and missing, but that facet bore a near-zero correlation with psi scoring in our data. Openness facet 05 (openness to ideas), which was suggestively related to psi scoring in our data, is positively correlated with MBTI Intuition (Costa & McCrae, 1992), an MBTI trait that has been shown to correlate with scoring in a manual ganzfeld series at this laboratory (Broughton, Kanthamani, & Khilji, 1989).

Focusing on the EC group as representing perhaps a more "psi-conducive" population, we find that the domain of Conscientiousness is negatively related to psi scoring. Conscientiousness in the NEO-PI-R is negatively correlated with the MBTI Perceiving factor. This means that the EC1 series is in accord with the finding that MBTI Perceiving is associated with better psi scoring (Kanthamani & Broughton, 1994).

As expected, our data displayed a weak negative correlation with geomagnetic variation, though this fell short of significance. This conforms to the general pattern of better ESP being found during periods when the earth's geomagnetic field is relatively quiet. Our data confirm (though not independently, since they are included in Spottiswoode's much larger data set) that the relationship between ESP and geomagnetic variation is much stronger during a roughly two-hour window of local sidereal time surrounding 13.5 hours LST. In fact, it is instructive to compare our Figure 3 with Spottiswoode's (1997b, p. 6) Figure I and observe the similarities, notably, the strengthening of the negative correlation at 6-8h and 12-14h LST, as well as the inverse at 9-11h and 16-18h LST. With the generally weak overall scoring in these data and only 15 trials in the optimum LST window, we did not find evidence of higher scoring in LST period. Nonetheless, Figure 2 shows both the peak scoring between 12.5h and 14.5h LST and the pronounced decline in scoring around 18h LST observed by Spottiswoode (1997a). This persistent finding of a weak geomagnetic influence on ESP, augmented by Spottiswoode's discovery of an interaction with the subject's orientation in the celestial sphere, may provide tantalizing clues to the mechanism by which ESP is "received" or comes into consciousness which only much further research can elucidate.

Although this project did not provide a straightforward replication of the PRL findings, several aspects of the data conform to findings from other studies and provide important indicators as to the many sources of variance that must be accounted for in the continuing quest to bring a higher level of reliability to ESP research.

1 This report subsumes the data in our interim report (Broughton & Alexander, 1995) and includes all of the data in a preliminary report of the completed series (Broughton & Alexander, 1996).

2 R. S. B. was not present for every session, but there are several other offices in the vicinity which were occupied by staff who knew when a test session was underway and knew to report any suspicious noise or activity.

3 No sender was ever observed to leave the room during a session and on only one occasion was sound heard outside the room. That instance involved the exclamation "Thank God!" when the sender was told via the earphones that he could join the receiver for feedback.

4 The extra session did not result in a hit.

5 Since various reports of ganzfeld research have used for effect size either Cohen's h or Rosenthal's [Pi] (PI or Proportional Index), we report both for ease of comparison. All z scores are based on the exact binomial probability with p = .25 and q = .75.

6 This assumes that none of the children were adopted. We had no indications that any of the children were, but we did not directly ask that question.

REFERENCES

BEM, D. J., & HONORTON, C. (1994). Does psi exist? Replicable evidence for an anomalous process of information transfer. Psychological Bulletin, 115, 4-18.

BIERMAN, D. J., BOSGA, D. J, GERDING, H., & WEZELMAN, R. (1993). Anomalous information access in the ganzfeld: Utrecht - novice series I and II. Proceedings of Presented Papers: The Parapsychological Association 36th Annual Convention, 92-203.

BROUGHTON, R. S., & ALEXANDER, C. H. (1995). Autoganzfeld II: The first 100 sessions. Proceedings of Presented Papers: The Parapsychological Association 38th Annual Convention, 53-61.

BROUGHTON, R. S., & ALEXANDER, C. H. (1996). Autoganzfeld II: An attempted replication of the PRL ganzfeld research. Proceedings of Presented Papers: The Parapsychological Association 39th Annual Convention, 45-56.

BROUGHTON, R. S., KANTHAMANI, H., & KHILJI, A. (1989). Assessing the PRL success model on an independent ganzfeld data base. Proceedings of Presented Papers: The Parapsychological Association 32nd Annual Convention, 26-33.

COSTA, P. T., & McCRAE, R. R. (1992). NEO-PI-R Professional Manual. Odessa, FL: Psychological Assessment Resources, Inc.

DALTON, K. (1997). Exploring the links: Creativity and psi in the ganzfeld. Proceedings of Presented Papers: The Parapsychological Association 40th Annual Convention, 119-134.

DALTON, K. S., MORRIS, R. L., DELANOY, D. L., RADIN, D. I., TAYLOR, R., & WISEMAN, R. (1994). Security measures in an automated ganzfeld system. Proceedings of Presented Papers: The Parapsychological Association 37th Annual Convention.

HONORTON, C., BERGER, R. E., VARVOGLIS, M. P., QUANT, M., DERR, P., SCHECTER, E. I., & FERRARI, D. C. (1990). Psi communication in the ganzfeld: Experiments with an automated testing system and a comparison with a meta-analysis of earlier studies. Journal of Parapsychology, 54, 99-139.

KANTHAMANI, H., & BROUGHTON, R. S. (1994). Institute for Parapsychology ganzfeld-esp experiments: The manual series. Proceedings of Presented Papers: The Parapsychological Association 37th Annual Convention, 182-189.

MORRIS, R. L., DALTON, K., DELANOY, D., &: WATT, C. (1995). Comparison of the sender/no sender condition in the ganzfeld. Proceedings of Presented Papers: The Parapsychological Association 38th Annual Convention, 244-259.

PERSINGER, M. A. (1993). Geophysical variables and behavior: LXXI. Differential contribution of geomagnetic activity to paranormal experiences concerning death and crisis: An alternative to the ESP hypothesis. Perceptual and Motor Skills, 76, 555-562.

PERSINGER, M. A., & KRIPPNER, S. (1989). Dream ESP experiments and geomagnetic activity. Journal of the American Society for Psychical Research, 83, 101-116.

RICE, G. E., & TOWNSEND, J. (1962). Agent-percipient relationship and GESP performance. Journal of Parapsychology, 26, 211-217.

SCHOUTEN, S. A. (1982). Analysing spontaneous cases: A replication based on the Rhine collection. European Journal of Parapsychology, 4, 113-158.

SPOTTISWOODE, S. J. P. (1990). Geomagnetic activity and anomalous cognition: A preliminary report of new evidence. Subtle Energies, 1, 91-102.

SPOTTISWOODE, S. J. P. (1997a). Association between effect size in free response anomalous cognition experiments and local sidereal time. Journal of Scientific Exploration, 11, 109-122.

SPOTTISWOODE, S. J. P. (1997b). Geomagnetic fluctuations and free-response anomalous cognition: A new understanding. Journal of Parapsychology, 61, 3-12.

STUART, C. E. (1946). GESP experiments with the free-response method. Journal of Parapsychology, 10, 21-35.

VAN KAMPEN, D., BIERMAN, D. J., & WEZELMAN, R. (1994). Personality and psi: Unravelling relations between extraversion, agreeableness and openness to experience with ganzfeld performance. Proceedings of Presented Papers: The Parapsychological Association 37th Annual Convention, 175-181.

VON LUCADOU, W. (1995). The Model of Pragmatic Information (MPI). European Journal of Parapsychology, 11, 58-75.

WEZELMAN, R., & BIERMAN, D. J. (1997). Process oriented ganzfeld research in Amsterdam. Proceedings of Presented Papers: The Parapsychological Association 40th Annual Convention, 477-492.

Institute for Parapsychology 402 North Buchanan Boulevard Durham, North Carolina 27701 e-mail: richard@rhine.org, cheryl@rhine.org

COPYRIGHT 1997 Parapsychology Press
COPYRIGHT 2004 Gale Group