Remote facilitation of attention focusing with psi-supportive versus psi-unsupportive experimenter suggestions.
In an attempt to control for many of these factors, Watt and Brady (2002) conducted two studies using remote facilitation of focusing of attention as the psi task. This psi task had only been reported twice before (Brady & Morris, 1997; Braud, Shafer, McNeill, & Guerra, 1995), and both studies had obtained positive evidence for a psi effect. In Watt and Brady's study, the experimenter Caroline Watt (CW) was the remote helper throughout. Prior to the session, half of the participants read an article intended to give the impression that CW had a track record of previous positive psi results, whereas the other half read a similar article that suggested CW had a history of null psi results. Thus, the experimenter's sex, appearance, personality, and psi belief and ability were controlled in these studies. The principal manipulation was of the participants' beliefs about their experimenter. During the session CW attempted to treat the participants identically; however, CW was not blind to the suggestion condition. Participants in the positive suggestion condition consistently gave CW more positive ratings for warmth, professionalism, and ability to instil confidence in the task; however, no difference in psi scoring was found between the positive suggestion and negative suggestion groups. (2) These studies suggest that, when experimenter variables are controlled and limited to varying the participants' beliefs about the experimenter's previous psi research track record, this has little effect on participants' psi performance. However, the Watt and Brady studies showed that participants' perceptions of the experimenter could be affected by even a small difference in how the experimenter is portrayed prior to them meeting the experimenter.
Watt and Brady (2002) investigated the effect of varying participants' prior knowledge about their experimenter. To build on our findings, we decided to conduct a study that systematically varied the psi supportiveness of suggestions made by the experimenter to the participants during the session. Hall of the participants were randomly allocated to the positive suggestion condition and half to the negative suggestion condition. In the positive condition, during the presession chat, the experimenter actively and explicitly referred to his positive psi belief and attempted to encourage the participants to consider and discuss possible psi experiences that they might have had. In addition, he referred to previous successful research using the same psi task and made positive suggestions for success in the session. In the negative condition, the experimenter portrayed himself as a disbeliever in psi and attempted to encourage the participants to consider and discuss possible "normal" interpretations of their possi ble psi experiences. He suggested that our own research had not confirmed previous positive results with this psi task and also suggested that this might be due to our greater methodological rigour. The experimenter attempted to maintain a warm and professional approach to all participants. This contrasts with the approach of Honorton, Ramsey, and Cabibbo (1975), whose experimenter in one condition was "friendly ... casual ... supportive" and in the other condition was "abrupt ... formal ... unfriendly." Thus in the present experiment we attempted to systematically vary the experimenter's professed psi belief and the associated psi-supportive versus psi-unsupportive context of the experimenter--participant interaction, while again holding constant experimenter personality, appearance, and so on.
Watt and Brady (2002) obtained no evidence of remote facilitation of focusing using the experimenter as remote helper. The present study therefore implemented a procedural change to the psi task, to investigate whether this would improve performance on the task. In the present study, participants were asked to bring in a friend to act as remote helper. This procedure had successfully been used by Brady and Morris (1997). Therefore, in the present study participants were working in pairs with a friend rather than a stranger. This procedural change also had the benefit of (as far as it is possible) eliminating the experimenter's role in the psi task. Therefore, it would be more likely that any difference in psi scoring might be attributed to how the experimenter portrayed himself and interacted with participants during the presession chat.
In both of their studies, Wiseman and Schlitz (1997, 1999) administered a psi belief questionnaire following the presession chat. Wiseman's participants indicated lower psi belief than Schlitz's participants. Watt and Brady (2002) administered a psi belief questionnaire after participants had read the priming article but prior to participants' meeting and chat with their experimenter. No difference in reported psi belief was found. This suggests that participants' expressed psi belief is unaffected by their prior knowledge of the experimenter's psi research track record. Rather, we hypothesize that during the presession chat participants may detect and respond to the demand characteristics arising out of their perception of the experimenter's attitude towards psi (Wiseman being a psi counter-advocate, Schlitz being a psi advocate). In the present study, we attempted to confirm this hypothesis by administering the psi belief questionnaire following the presession chat with the experimenter. We therefore predic ted that those participants in the negative suggestion condition would have lower psi belief scores than those in the positive suggestion condition.
The volunteer participants were recruited principally through the Department of Psychology volunteer panel and were asked to bring a friend for the testing session. Most of the participants were members of the general public, but a few were postgraduate students and colleagues in the Department of Psychology. They were selected on the basis of their interest and willingness to participate in the study. Participants were informed in advance of the procedure, and they knew that each would play the role of helper and helpee. All participants were paid [pounds sterling]5 for participating.
Belief in Psi Questionnaire. This was a 12-item questionnaire with 1 item on participants' belief that they could demonstrate psi in a controlled experiment and 11 items on belief in and experience of telepathy, clairvoyance, precognition, and psychokinesis. These terms were all defined for the participants. High scores indicated higher belief. Scores could potentially range from 12 to 84.
Eye-Contact Questionnaire. This was a 67-item questionnaire devised by Ian Baker (IB) for his own research purposes. We will not report details of responses to this questionnaire in the present article (further details are reported in Baker, 2001).
Helper's Session Questionnaire. This was intended to obtain qualitative information about the helper's response to the session and consisted of two presession questions:
1. How comfortable are you with the idea of remotely helping your friend to focus? (5-point scale ranging from not at all comfortable to very comfortable).
2. How confident are you of being able to help your friend to focus during the help periods? (5-point scale ranging from not at all confident to very confident).
There were two postsession questions:
3. To what extent do you feel that you were helping your friend to focus during the help periods? (5-point scale ranging from not at all to completely).
4. How would you rate the quality of your own focusing during the help periods? (5-point scale ranging from completely focused to completely distracted, reverse scored).
This questionnaire was scored so that higher scores represented more positive responses.
Helpee's Session Questionnaire. This was intended to obtain qualitative information about the helpee's response to the session and also to validate the condition manipulation. There were two presession questions:
1. How comfortable are you with the idea of your friend remotely helping you to focus? (5-point scale ranging from not at all comfortable to very comfortable).
2. How confident are you of having fewer distractions when your friend is attempting to help you to focus? (5-point scale ranging from not at all confident to very confident).
There were four postsession questions:
3. To what extent do you feel that you were being helped to focus by your friend? (5-point scale ranging from not at all to completely).
4-7. How would you rate your experimenter? (Warm-Cold; Professional-Unprofessional; Instilling confidence for task-Not instilling confidence for task; Believing in psi-Not believing in psi; each on a 5-point scale, reverse scored).
The present study investigated whether the psi task would be more psi-conducive than that in Watt and Brady (2002) because the present study had friends as helpers rather than the experimenter as helper. To obtain information about whether participants would prefer to do the psi task with their friend as opposed to an experimenter, the final two questions were prefaced: "Imagine instead of your friend you had the experimenter being your remote helper":
8. How comfortable would you be with the idea of your focusing being remotely helped by the experimenter? (5-point scale ranging from not at all comfortable to very comfortable).
9. How confident would you be of having fewer distractions when the experimenter was attempting to help you to focus? (5-point scale ranging from not at all confident to very confident).
This questionnaire was scored so that higher scores represented more positive responses.
Video Recording. For the purposes of recording the presession interaction between the experimenter and the participants, a video camera was used. A mirror was positioned so that the experimenter's face could be filmed simultaneously with the participants' faces. This would enable future examination of nonverbal interactions between the experimenter and participants. No further details will be given on this measure in the present article (for more details, see Baker, 2001).
The psi task was the same as that used in all previous remote facilitation of focusing studies. One participant, the "helpee," sat in a comfortable chair in a darkened, sound-attenuated room and focussed his or her attention on a lit candle in a blue translucent glass holder. When helpees became aware that their attention had drifted from the candle, they were asked to press a hand-held response button. They were then to gently bring their attention back to the candle, until the next distraction, and so on. At the same time, in a separate room located about 20 m away and up a short flight of stairs, the second participant, the "helper," also sat in a comfortable chair with a lit candle in a blue glass holder. The helper's room contained a computer monitor that gave instructions either to "HELP"--in which case the helper was to focus his or her attention on the candle and maintain the mental intention that this would help his or her friend to focus also--or "CONTROL"--in which case the helper did not attempt to focus on the candle or to help the friend. The helper wore headphones that played a tone every time the instructions on the monitor changed, so that the helper would know to look at the monitor for new instructions. The experimenter, helper, and helpee were all blind to the influence schedule prior to its start. After the schedule commenced, only the helper had information about the schedule. The helpee had no information about the moment at which the schedule commenced.
The help--control schedule was arranged in a counterbalanced fashion in eight pairs: four help--control pairs and four control--help pairs. Therefore, there was a total of 16 one-minute influence periods. A computer program randomly allocated the order of these eight pairs. This was done for each individual session once the program was initiated. The computer running the program was contained in a third room, accessed only by the experimenter once the participants were sequestered in their respective rooms. The controlling program recorded the time of each button press from the helpee and whether the press occurred in a Help or Control epoch. The program automatically wrote this information both to hard disk and floppy disk at the end of the session, together with summary statistics about the sum of Help presses, the sum of Control presses, and a percentage influence score (PIS) calculated from the ratio of these two sums (Braud & Schlitz, 1991). A score of .5 represented exact chance performance and a score of greater than .5 represented fewer distractions in the Help periods compared with Control periods, whereas a score of less than .5 represented the opposite trend. The controlling computer monitor did not display the session statistics at the end of the session, so the experimenter and participants were blind to the session results until all session measures had been completed.
Design and Analysis
Participants worked in pairs and swapped roles during the session so that each was the helper once and each was the helpee once; there were therefore two psi trials per session. Participant pairs were randomly allocated to session condition (20 positive suggestion pairs and 20 negative suggestion pairs, giving a total of 40 participants in each condition). This random allocation was prepared in advance of the entire study by CW with random number tables. The condition designation for each session was sealed in separate envelopes so that IB was blind to the condition designation at the time of recruitment and scheduling. This represented a methodological improvement on the nonblind and nonrandom scheduling methods of Watt and Brady (2002) and Wiseman and Schlitz (1997, 1999). The decision as to which participant would be helper first in each session was made by a coin toss by IB during the session. The primary outcome measure for the psi task was each participant's sum of Help button presses compared with their sum of Control button presses. Each participant therefore contributed a pair of data-points for a related t test. It was predicted that participants would have fewer Help than Control button presses, indicating a remote influence of attention-focusing effect. A single outcome measure for each psi session was needed for the second hypothesis. It was felt that PIS was appropriate for the current data, but in general there are limitations with PIS (see Schmidt, Schneider, Binder, Burkle, & Walach [20011 and discussion of this question in Watt & Brady ). It was predicted that participants in the positive suggestion condition would have higher PIS scores than those in the negative suggestion condition (unrelated t test comparing positive suggestion PIS score with negative suggestion PIS score). Third, it was predicted that the psi belief questionnaire responses for those in the positive suggestion condition would show higher psi belief than for those in the negative suggestion condition (Mann-Whitney test). As these three predictions are directional, one-tailed p values are used. Two-tailed p values are used for all other analyses. As it is possible that participants' own psi belief might be a moderating variable in their responsiveness to the experimente r's psi-supportive or psi-unsupportive suggestions (e.g., believing participants might be more responsive to the positive suggestion condition, disbelieving participants might be more responsive to the negative suggestion condition), as an exploratory measure we also divided participants according to belief and conducted a 2 x 2 analysis of variance (ANOVA) of psi scores in positive and negative suggestion conditions, for high-belief and low-belief participants. The other session questionnaire measures were used for exploratory purposes, and no specific predictions were made.
IB recruited and scheduled participants. Each participant was asked to bring a friend to the session. On their arrival at the lab, IB offered them refreshments and seated them in the reception suite and commenced the videotaping. IB then performed a controlled measure of nonverbal communication (staring) without participants' knowledge (as described in more detail in Baker, 2001). IB then described the session procedure. In the positive suggestion condition, he said that he himself had had psi experiences and was a psi believer. He also said that we had had success with the remote focusing psi task in the past and that we expected that success to continue. He then asked participants if they had had any psi experiences and chatted about these in a supportive way. In the negative suggestion condition, IB said he had not personally had any psi experiences and did not believe in psi. He also said that although he had had some ostensibly paranormal experiences, he had always been able to find a normal explanation for them. He asked participants if they had had any possibly psychic experiences, and when chatting about these he focused on possible normal interpretations of these experiences. He noted that although some parapsychologists had had positive results with the remote focusing task, we had not, and that this might be due to our adoption of stricter methodological precautions. In short, the aim was for IB to make psi-supportive suggestions in the positive suggestion condition and for him to make psi-unsupportive suggestions in the negative suggestion condition.
Following this initial chat, IB switched off the video recorder and asked participants to complete the belief questionnaire. He then tossed a coin to determine who would first be helper and helpee, and participants then completed the appropriate presession helper and helpee questionnaire. Accompanied by the helper, he seated the helpee in the sound-attenuated room and asked them to press the response button to become familiar with its operation. He lit the helpee's candle and positioned it on a table about 6 ft (1.8 m) in front of the helpee at eye level. He indicated the desktop lamp that the participant could switch on if needed during the session (e.g., if the candle stopped burning). He answered any final questions that the helpee might have and noted that the helpee would not know when the influence period began, so the helpee was to act as if the influence period was commencing directly when the room doors were closed. He then wished the helpee good luck, extinguished the light, closed the two doors, a nd exited the experimental suite with the helper, also closing the outside door of the suite. He took the helper to their room, seated them in a comfortable chair, and lit their candle. He gave the helper headphones and answered any final questions the helper might have. He wished the helper good luck and then exited the room, closing the door. IB then went to the control room, which was adjacent to the helpee's room and which contained a sensor that would flash a red light if the helper opened the door of the room. TB was therefore able to verify that both participants stayed in their designated rooms during the influence period. IB then entered the session details on the control computer and initiated the influence schedule, which was preceded by a 15-s time delay. It is estimated that it took IB up to 3 min from closing the helpee's door to starting the influence schedule. During the schedule, IB's computer monitor displayed the words "session in progress." At the end of the influence period, IB retrieved the helper from the helper's room, and both then retrieved the helpee and returned to the reception room. Both helper and helpee completed their respective postsession questionnaires and then swapped roles for the second part of the session.
Helper and helpee again completed their respective presession questionnaires and were shown to their respective rooms in the same way as before. Following completion of the second psi task, they were again retrieved from their rooms as before, returned to the reception room, and completed their respective postsession questionnaires. They were then asked to complete the staring questionnaire. Finally, IB debriefed participants, explaining how he had attempted to make psi-supportive or psi-unsupportive suggestions. Participants each then received their [pounds sterling]5 payment. IB then took the participants to the control room and showed them their session psi results. Participants were then thanked, were asked not to discuss the study with other possible participants, and left the building. Following completion of the entire study, all participants were sent individual feedback letters giving their psi belief and staring questionnaire scores, as well as information about the study results as a whole.
The study ended when, as planned, a total of 80 participants had taken part (40 in the positive suggestion condition and 40 in the negative suggestion condition).
There were 53 female and 27 male participants. Ages ranged from 14 to 81 years, with a mean age of 37.9 years (SD = 14.5).
Although all of the participants successfully completed the psi task, 1 participant (in the negative suggestion condition) did not press the response button during any of the Help epochs, therefore it was not possible to calculate a PIS score for that participant (one cannot divide a zero score). This participant's data were therefore included in the overall psi measure (a comparison of Help presses to Control presses) but were excluded from the psi and suggestion measure (a comparison of positive suggestion P15 scores with negative suggestion PIS scores).
Suggestion Manipulation Check
All of the participants had been asked to rate their experimenter's degree of belief in psi to check whether the experimenter had successfully communicated this to participants. The mean experimenter psi belief rating for those in the positive suggestion condition was 4.3 (SD = 0.8), which was significantly higher than the mean score of 2.7 (SD = 0.9) for the negative suggestion participants (Mann-Whitney Z = 6.281, p<.0001, two-tailed). This suggests the experimental manipulation was fairly successful. However, with a scale midpoint of 3, we can see that those in the negative suggestion condition rated IB as having only moderate disbelief about psi. This may be because of the implicit psi-supportive context overall, given that the study was taking place in a parapsychology unit.
The mean number of distractions in the Help and Control epochs was 10.35 (SD = 8.3) and 10.76 (SD = 8.8), respectively. This represented a slight tendency for fewer distractions in the Help epochs, as predicted. However, the difference was not statistically significant, related t(79) = 1.040, p = .151, one-tailed. The associated effect size (Cohen's r) = .12. This does not replicate the magnitude of effect or the significant psi results found in the studies by Brady and Morris (1997) and Braud et al. (1995). It does replicate the null psi results of Watt and Brady (2002).
Psi and Suggestion
The mean PIS score for those in the positive suggestion condition was 0.49 (SD = 0.13, n = 40), compared with 0.53 (SD = 0.11, n = 39) for negative suggestion participants. Therefore, there is a nonsignificant trend in the direction opposite to that predicted, for negative suggestion participants to have higher PIS scores than positive suggestion participants, unrelated t(77) = -1.386, p = .170, two-tailed; ES = .16.
Suggestion and Psi Belief
Mean psi belief score was 47.2 (range = 12 to 84, SD = 18.15). As predicted, the positive suggestion participants had a higher mean psi belief score (M = 50.2, SD = 18.3) than negative suggestion participants (M = 44.1, SD = 17.7). This trend is marginally significant, unrelated t(78) = 1.510, p = .067, one-tailed. There was no significant correlation between psi belief score and PIS scores, [r.sub.s](79) = .135.
Psi Belief as a Moderating Variable
We wished to investigate whether psi believers and disbelievers might respond differently to the different suggestion conditions. For instance, the believers might show highest PIS scores in the positive suggestion condition, whereas the disbelievers might show lowest PIS scores in the negative suggestion condition. The median belief score was 47. Participants were split into high psi belief and low psi belief groups. The former group, designated psi believers, consisted of 40 individuals whose belief scores ranged from 48 to 84 (mean score = 62.4, SD = 10.0). The latter group, designated psi disbelievers, consisted of 40 individuals whose belief scores ranged from 12 to 46 (mean score = 32.0, SD = 9.6). A 2 x 2 ANOVA was then calculated on PIS scores for these two groups in each suggestion condition (see Table 1 for descriptive data). No significant main effects were found for condition, F(1, 75) = 2.131, p = .149, or for psi belief, F(1, 75) = 0.370, p = .545, and no interaction was found between the two, F (1, 75) = 0.046, p = .831.
Experimenter Evaluation Ratings
Table 2 shows participants' mean questionnaire ratings of the experimenter. As in the previous two remote facilitation of focusing studies (Watt & Brady, 2002), the positive suggestion group tended to give higher ratings of the experimenter's warmth and ability to instil confidence in the task than the negative suggestion group. However, these trends were not statistically significant (see Table 2).
Prior Confidence, and Perceived Success at Psi Task
All of the participants indicated only moderate prior confidence and perceived success at the psi task (see Table 3). Helpers indicated significantly higher perceived success in the positive suggestion condition compared with the negative suggestion condition (Mann-Whitney z = 1.965, p = .05, two-tailed). This indicates that the experimenter's psi-supportive or psi-unsupportive suggestions can affect participants' reports of their performance at the psi task, suggesting that participants may feel their experimenter is an important component of the experimental system.
Friend Versus Experimenter as Helper
The present study introduced a methodological change in which participants had a friend as remote helper rather than the experimenter. The session questionnaires asked helpees to indicate their comfort and confidence of success having their friend as helper and also to evaluate the hypothetical situation of having the experimenter as helper. Table 4 shows that all of the participants indicated that they were quite comfortable in the experimental setting, and there was little difference between conditions. There was a tendency for participants' confidence to be higher in the positive suggestion condition compared with the negative suggestion condition, and this trend was significant for the question about the experimenter as helper (Mann-Whitney z = 2.093, p = .04, two-tailed). Overall, participants indicated significantly higher comfort with their friend helping than with the hypothetical scenario of the experimenter helping (mean comfort with friend helper 4.62, SD = 0.8; mean comfort with experimenter helpe r = 4.10, SD = 1.1; Wilcoxon z = 3.837, p = .0001, two-tailed). Likewise, participants overall indicated significantly higher confidence with their friend helping than with the experimenter hypothetically helping (mean confidence with friend helper = 2.90, SD = 1.2; mean confidence with experimenter helper = 2.62, SD = 1.1; Wilcoxon z = 2.094, p = .04, two-tailed).
Quality of Helpers' Focusing
To gain more information about the helpers' role in the psi task, we asked them to rate their quality of focusing during the Help epochs. Overall mean focusing quality rating was 3.41 (SD= 1.0), and positive suggestion participants indicated greater quality of focusing (M= 3.6, SD= 1.0) than negative suggestion participants (M = 3.2, SD = 0.9). A Mann-Whitney comparison gave z = 1.778, p .075, two-tailed. This gives some indication that the experimenter's psi-supportive or psi-unsupportive suggestions may affect participants' behavior during the psi task. Possibly, the helpers in the positive suggestion condition tried harder than those in the negative suggestion condition: The former group had a tendency to give higher quality of focusing ratings than the negative suggestion helpers (see below for details). There was no correlation between quality of helper's focusing and PIS score, [r.sub.s](79) = -.008. However, it is interesting to note that there was a statistically significant negative correlation betwe en quality of focusing and total number of button presses (distractions) during the session, [r.sub.s](80) =-.240, p<.05, two-tailed. Therefore, the better the helper's focusing, the fewer the number of distractions reported by the helpee. This pattern could be due to chance, but it is worth investigating in a follow-up study as it may be indicative of some sort of "diffuse" psi effect, that is, an effect in both Help and Control epochs rather than an effect occurring during the Help epochs.
The present result repeats the nonsignificant overall psi results of the remote facilitation of focusing study by Watt and Brady (2002), in which CW acted as remote helper. It was thought that by introducing a procedural change to the present study, whereby participants had a friend acting as remote helper, this might enhance performance on the psi task. There is some slight suggestion that this may be the case as Watt and Brady (2002) found a negative psi effect, related t(59) = -0.823, whereas the present study found a positive psi effect, related t(79) 1.040. However, the difference is not statistically significant ([z.sub.diff] = 1.305, p = .10, one-tailed).
There are a number of possible reasons why the present study did not find an overall psi effect whereas two previous remote facilitation of focusing studies did. We lay these out for consideration, but their status is of necessity post hoc and speculative. The first possibility is that the psi effect in the two previous successful remote facilitation of focusing studies was spurious and that the present study, like the Watt and Brady (2002) study, reflects the true situation that psi does not exist. The second possibility is that the numerous procedural differences between our study and the two previous successful studies may have obscured or constrained the psi effect. The most obvious procedural difference is that the present study, like the Watt and Brady study, asked questions about experimenter effects. Therefore, there was an additional layer of complexity to the present study that may have had some effect on the study's overall psi results. For instance, the experimenter in the present study was role-p laying, and therefore his behavior was perhaps not natural compared with a situation in which a genuine psi proponent or a genuine skeptic was interacting with participants. However, IB considers himself to be neutral on the question of the existence of psi. Therefore, it was perhaps relatively easy for him to shade more towards the negative or the positive without feeling this was unduly unnatural or deceptive. It is probably true to say, however, that if he held strong beliefs about psi, his interactions with participants may have been qualitatively different.
Another procedural difference is that in the previous two successful remote focusing studies, there was some flexibility in the choice of role for participants (Brady & Morris, 1997) and in the timing of the start of the session (Braud et al., 1995). Both of the unsuccessful remote focusing studies presented participants with less choice, and this may have affected their motivation. The two unsuccessful studies may also have had a different participant population than the successful studies. However, there is little similarity between the participant population of the two successful studies. Braud et al. selected participants with a particular interest in attention and remote mental influence, whereas Brady and Morris seemed to have a broader sample more similar to that for the two unsuccessful remote focusing studies. It is therefore difficult to see any systematic differences between successful and unsuccessful studies in terms of participant population.
It is also possible that the present study had a different kind of influencer than in the two previous successful remote focusing studies.
However, the present study adopted the same procedure as the successful Brady and Morris (1997) study, in asking participants to bring in friends to act as helpers. Watt and Brady's (2002) study had the experimenter act as helper, whereas the successful Braud et al. (1995) study had three laboratory staff members act as experimenter and helper. Again, it is difficult to see systematic differences between the successful and the unsuccessful studies in terms of who was the influencer.
It is worth noting that participants showed relatively few distractions in the present study compared with the three previous remote focusing studies (in the present study, the mean number of distractions in the Help and Control epochs was 10.35 and 10.76, respectively). Watt and Brady (2002) had a mean of 12.58 Help and 12.20 Control button presses; Braud et al. (1995) had a mean of 12.43 Help and 13.60 Control presses; and Brady and Morris (1997) had a mean of 18.45 Help and 19.60 Control presses. All studies used the same duration of testing schedule and the same number of Help and Control epochs. We do not know why participants in the present study had relatively few distractions. However, one important consequence of this is that there is less lability in participants' behavior and therefore, according to some parapsychologists, less of an opportunity for a psi effect to manifest (Braud, 1980). In support of this line of reasoning, the study with the greatest mean number of distractions (Brady & Morris, 1997) also had the greatest effect size.
Finally, in this discussion of possible reasons why the present study obtained null psi results, we can consider the possibility of experimenter psi. There is no information about the psi research track record of the three experimenters/helpers in Braud et al.'s (1995) study. Braud appeared to have no interaction with participants, though he was the overall investigator and he does have a reputation as a psi-conducive researcher. Brady, the experimenter in the other successful remote focusing study, had no previous experience of psi research and therefore no track record. Watt, the experimenter and helper in Watt and Brady's (2002) study, has conducted many psi studies, a few of which have obtained positive psi results, but more have obtained null psi results. Her track record as experimenter is mixed. However, Watt was successful as a sender in a ganzfeld study by Morris, Dalton, Delanoy, and Watt (1995). Overall, her attitude towards psi is on the positive shade of neutral. Baker, the experimenter in the pr esent study, had never before been an experimenter in a parapsychological study, so he had no track record, and he describes himself as neutral towards psi. Watt had no interaction with participants in the present study. There appears to be insufficient information for us to make any useful suggestions at this stage as to how experimenter psi may contribute to the varying outcome of the four remote focusing studies that have now been published.
The present study varied the experimenter's interaction with the participants, specifically whether the experimenter made psi-supportive or psi-unsupportive suggestions. We found that positive suggestion participants showed little difference in PIS performance compared with negative suggestion participants, so in terms of experimenter effects on psi performance, this study has provided no evidence of such an effect.
However, our study does reveal many interesting psychological factors associated with experimenter suggestion. First, we have confirmed the findings of Wiseman and Schlitz (1997, 1999) of a difference in psi belief scores dependent on the suggestion condition. Those in the positive suggestion condition had higher psi belief scores than those in the negative suggestion condition. Unlike Wiseman and Schlitz, our study controls for many potentially confounding experimenter differences, such as sex, personality, physical characteristics, and experimenter psi, because we had the same experimenter in both conditions. In other words, Wiseman and Schlitz found systematic differences in their participants' belief scores without controlling for many experimenter characteristics. We found similar differences in belief scores while controlling for these characteristics and varying just the experimenter's psi-supportive and psi-unsupportive suggestions. Therefore, we can suggest that Wiseman and Schlitz's participants' di fferential psi belief scores could have been influenced by their experimenters' explicit and implicit communication of their own beliefs about psi and the psi-supportive and psi-unsupportive context that may arise out of this.
Also, we found that participants indicated higher confidence, higher expectancy of success, higher perceived helping, and better quality focusing in the positive suggestion condition compared with the negative suggestion condition. This points to a complex set of ways in which the experimenter's psi-supportive and psi-unsupportive suggestions may affect how the participant approaches the experimental task.
We also found that, as with the previous Watt and Brady (2002) studies, participants in the positive suggestion condition tended to give their experimenter a more positive evaluation than those in the negative suggestion condition, although the experimenter was the same in all cases. Again, this further reveals the potential extent of the impact of the experimenter's behavior on the participant. However, all participants may have been susceptible to response bias in this situation, as they were being asked to rate the experimenter in the presence of the experimenter.
In some respects, the present study raises more questions than answers. We have revealed some of the complex psychological impact on participants of the psi supportiveness of the experimenter, but we have not revealed a relationship between this and actual psi performance. Possibly our manipulations are too far removed from how experimenter suggestion operates in everyday parapsychology studies. Our experimenter was simply role-playing, and this may have made the psi-supportive/unsupportive suggestions seem artificial or contrived. This may have limited the demand characteristics and reduced the impact of the experimenter's suggestions. The strength of our approach, using a single role-playing experimenter, is to control for many experimenter characteristics, some of which may contribute to an experimenter effect and others of which may be irrelevant. We have isolated and examined one factor: the experimenter's psi-supportive suggestions. This is a simplified setting. In reality, a wide variety of varying exp erimenter characteristics may be involved in parapsychology's experimenter effect. As in all psychological research, there is a trade-off between control and ecological validity.
TABLE 1 MEAN PIS SCORES FOR PSI BELIEVERS AND PSI DISBELIEVERS, BY SUGGESTION CONDITION Condition Psi believers Psi disbelievers Positive suggestion 0.488 (0.16, 16) 0.498 (0.10, 24) Negative suggestion 0.522 (0.12, 23) 0.544 (0.10, 16) Note. SD, N in parentheses. TABLE 2 MEAN EXPERIMENTER (E) RATINGS Positive Negative Measure suggestion suggestion z P E Warmth 4.75 (0.4) 4.55 (0.7) 1.086 .28 E Professionalism 4.70 (0.6) 4.75 (0.4) -0.128 .90 E Instilling confidence 4.63 (0.6) 4.37 (0.9) 0.967 .33 Note, SD in parentheses. z scores are based on Mann-Whitney comparisons; p values are two-tailed. TABLE 3 HELPEES' AND HELPERS' RATINGS OF PRIOR CONFIDENCE PF SUCCESS AND OF PERCEIVED SUCCESS AT THE PSI TASK Positive Negative Measure suggestion suggestion z p Helpee's prior Confidence 3.02 (1.2) 2.78 (1.2) 0.787 .43 Helper's prior Confidence 2.80 (1.3) 2.80 (1.2) 0.128 .90 Helpee's perceived success 2.95 (1.3) 2.41 (1.3) 1.083 .28 Helper's perceived success 3.30 (0.08) 2.72 (1.4) 1.965 .05 Note. Mean scores, SD in parentheses, z scores based on Mann-Whitney comparisons; p values are two-tailed. TABLE 4 HELPEE'S MEAN RATINGS OF COMFORT AND CONFIDENCE WITH FRIEND (F) HELPING AND EXPERIMENTER (E) HELPING Positive Negative Measure suggestion suggestion z p Comfort with F 4.60 (0.9) 4.65 (0.7) -0.356 .72 as helper Confidence with F 3.02 (1.2) 2.78 (1.2) 0.787 .43 as helper Comfort if E 4.10 (1.1) 4.10 (1.2) 0.063 .95 had been helper Confidence if E 2.88 (1.0) 2.38 (1.1) 2.093 .04 had been helper Note. SD in parentheses. z scores based on Mann-Whitney comparisons p values are two-tailed.
We are grateful to the Society for Psychical Research for funding this study. We thank our anonymous referees for their helpful comments.
(1.) EDA stands for electrodermal activity, DMILS stands for direct mental interaction with living systems.
(2.) No psi results were reported for Study 1 because of a computer error.
BAKER, I. S. (2001). The relationship of gaze-avoidance to shyness and belief in the power of gaze, and an analysis of vocabulary for eye-fixation research. Unpublished master's thesis, University of Edinburgh, Edinburgh, Scotland, United Kingdom.
BRADY, C., & MORRIS, R. (1997). Attention focusing facilitated through remote mental interaction: A replication and exploration of parameters. Proceedings of Presented Papers: The Parapsychological Association 40th Annual Convention, 73-91.
BRAUD, W. G. (1980). Lability and inertia in conformance behavior. Journal of the American Society for Psychical Research, 74, 297-318.
BRAUD, W. G., & SCHLITZ, M. J. (1991). Conscious interactions with remote biological systems: Anomalous intentionality effects. Subtle Energies, 2, 1-46.
BRAUD, W., SHAFER, D., McNEILL, K., & GUERRA, V. (1995). Attention focusing facilitated through remote mental interaction. Journal of the American Society for Psychical Research, 89, 103-115.
HONORTON, C., RAMSEY, M., & CABIBBO, C. (1975). Experimenter effects in extrasensory perception. Journal of the American Society far Psychical Research, 69, 135-139.
KENNEDY, J. E., & TADDONIO, J. L. (1976). Experimenter effects in parapsychological research. Journal of Parapsychology, 40, 1-33.
MORRIS, R. L., DALTON, K., DELANOY, D. L., & WATT, C. A. (1995). Comparison of the sender/no sender condition in the ganzfeld. Proceedings of Presented Papers: The Parapsychological Association 38th Annual Convention, 244-259.
SCHMIDT, S., SCHNEIDER, R., BINDER, M., BURKLE, D., & WALACH, H. (2001). Investigating methodological issues in EDA-DMILS: Results from a pilot study. Journal of Parapsychology, 65, 59-82.
SCHNEIDER, R., BINDER, M., & WALACH, H. (2000). Examining the role of neutral versus personal experimenter-participant interactions: An EDA-DMILS experiment. Journal of Parapsychology, 64, 181-194.
WATT, C., & BRADY, C. (2002). Experimenter effects and the remote facilitation of attention focusing: Two studies and the discovery of an artifact. Journal of Parapsychology, 66, 49-71.
WISEMAN, R., & SCHLITZ, M. (1997). Experimenter effects and the remote detection of staring. Journal of Parapsychology, 61, 197-207.
WISEMAN, R., & SCHLITZ, M. (1999). Experimenter effects and the remote detection of staring: An attempted replication. Proceedings of Presented Papers: The Parapsychological Association 42nd Annual Convention, 471-479.
Department of Psychology
University of Edinburgh
7 George Square
Edinburgh EH8 9JZ
|Printer friendly Cite/link Email Feedback|
|Author:||Watt, Caroline; Baker, Ian S.|
|Publication:||The Journal of Parapsychology|
|Date:||Jun 1, 2002|
|Previous Article:||Relationship between the hypnagogic/hypnopompic states and reports of anomalous experiences.|
|Next Article:||Psi on the Web: two replication studies.|