Psi, perception without awareness, and false recognition

Stuart Wilson

Comparisons between psi and perception without awareness (hence-forth PWA) (1) have been made throughout the history of parapsychology. Frederick Myers (1903) was perhaps the first researcher to suggest a relationship between the paranormal and information below the level of conscious awareness.

Since then, a number of researchers have been impressed by the apparent similarities between psi and PWA (see, e.g., Beloff, 1972; Dixon, 1979; Nash, 1986; Roney-Dougal, 1981, 1986; Schmeidler, 1986). It is important to note, as Beloff (1972) did, that the analogy can begin only at the point at which the information about the target has entered the cognitive system and is awaiting processing. A stimulus that is of such intensity to allow PWA is, nonetheless, a sensory stimulus. It has known physical qualifies (albeit weak), and it is available to the known sensory apparatus of the percipient. A psi input, however, is, by nature, not available to any of the recognised senses, and, in an experimental situation, is actively shielded from the percipient. What the present article is concerned with, therefore, is the way the information (either psi information or information that is perceived outside of 'awareness) is processed once it is in the system.

Assuming that psi processing is unconscious in nature (for justification of this assumption, see, e.g., Broughton, 1988; Stanford, 1990), then it would be reasonable to suggest that it might be processed similarly to other unconscious phenomena. (2) If this were the case, then it would add considerable credibility to the psi hypothesis. It would do this in the general sense that learning that psi has certain properties indicates that it has existence and that it can be studied in similar ways to other psychological phenomena. Comparing psi and PWA may also help shed some light on the way in which weak stimuli are processed (be they psi stimuli or weak sensory stimuli). Nash (1986) contended that many of the factors that have been previously found to be important in psi research were later found also to apply to PWA, and vice versa. For example, Nash identified, among others, that the drawing of responses, unintentional negative scoring (or "missing"), improved performance at the beginning and end of testing (U-shaped scoring distribution), and the effect of believers versus disbelievers were all effects first found in parapsychology that were subsequently found to have a similar effect in PWA research. Likewise, sensory deprivation, relaxation, perceptual defence, galvanic skin response, and right hemisphericity were all effects that were first found to influence PWA scores and were later found to have an effect on psi tasks.

Moreover, understanding PWA is important for parapsychologists because it offers another explanation for certain kinds of anomalous experience that does not imply psi. PWA offers a possible mechanism for many experiences in which people claim to have been influenced by some "unknown" or anomalous means. In many cases, the stimulus identified by the percipient as being anomalous may simply have been outside of awareness. This may have led the percipient to conclude that, as he or she had no conscious awareness of being influenced, then there must have been a "paranormal" force at work.

There have been many comparisons between the two phenomena over the years (see, e.g., Beloff, 1972; Dixon, 1979; Nash, 1986; Roney-Dougal, 1986, 1987; Schmeidler, 1986). One comparison is at the subjective level. Responses in both PWA tasks and psi tasks are often described as guesses, and it is often impossible to distinguish between guesses in the two types of task. In an experiment related to this theme, Miller (1940) presented participants with a PWA test disguised as a test of telepathy. None of the participants were aware that there was a PWA component to the task and were surprised to learn that their guesses were not based on telepathy but had actually been influenced by a weak sensory stimulus. This has implications for spontaneous cases, in which people may have an experience (e.g., an intuition) and attach their own label to the causes (e.g., psi). Subjectively, the effects of psi and PWA may occasionally be indistinguishable, and the interpretation placed on the experience may depend largely on the percipient's prior beliefs. As a hypothetical example, take a person whose route to work each day takes her over a suspension bridge. On one particular day, just as she is approaching the bridge, she decides to take an alternative route. Suppose that this is the first time she has ever done this. She later discovers that the bridge collapsed at around the same time she would have been on it had she taken her normal route. Depending on her prior beliefs, she may attribute her fortunate escape to a number of things. She may conclude that she was guided by means of some paranormal acquisition of information about the bridge collapse. This may be either through some form of unconscious precognition or through clairvoyance of the state of the bridge. Alternatively, it is possible that, on her approach to the bridge, she unconsciously picked up on something about the bridge that was not normal. It may have been a very subtle cue, like the way the bridge was swaying or something similar. Although this was outside of awareness, it might have been enough to influence her decision to take an alternative route, thus avoiding the potentially fatal situation ahead. Stanford (1990) offered a cogent discussion on these issues.

Another comparable effect is that of relaxed passivity. This has previously been shown to enhance both psi and PWA. In parapsychology, the technique known as the ganzfeld involves placing the participant in a state of mild sensory deprivation (see Honorton, 1977). Proponents of this technique maintain that participants will have a better chance of gaining extrasensory information when provided with unpatterned sensory stimulation. In the field of PWA, Dixon (1981) cited various studies suggesting that weak sensory inputs may be enhanced when other senses are suppressed (e.g., Fisher & Paul, 1959, cited in Dixon, 1981).

A third likeness concerns creativity. This has been shown to influence both psi and PWA. Dalton (1997) reported that participants who were found to be creative performed better in a ganzfeld ESP task. Similarly, Dixon (1981) cited Gordon's (1967) post hoc finding that participants drawn from the arts departments of his university tended to perform better in a test of PWA than those drawn from the science and engineering faculty (Dixon, 1981, p. 70), although this may not strictly be a consequence of "creativity" per se.

Finally, Schmeidler (1986) reported striking similarities between her work on personality variables in ESP performance and Eagle's (1962) findings concerning personality and PWA performance. Among these are respondents showing cognitive and affective openness tend to respond better to both kinds of stimuli, as do respondents showing more receptiveness to inner cues, as do respondents found to be oriented toward, and responsive to, people (see Schmeidler, 1986, p. 244).

There have been various studies directly examining the relationship between psi and PWA. Kreitler and Kreitler (1972) investigated the influence of ESP on the degree of PWA. They reported a study in which a sender attempted to influence a participant's performance on a task involving identifying letters presented at a level that was ostensibly, outside of awareness. They found that the presence of a sender did seem to have an effect on the PWA task. Other studies have looked at the relationship between PWA and psi using the ganzfeld technique. For instance, Smith, Tremmel, and Honorton (1976) conducted a ganzfeld study whereby the sender viewed the target either outside of awareness (i.e., through a tachistoscope for 1 ms) or continuously for a period of 10 min. Smith et al. reported an overall significant hit rate, which was found to be primarily a consequence of the PWA-sending condition. A nonsignificant hit rate was found in the condition in which the target was viewed for a 10-min period. Kanthamani and Palmer (1993) conducted a similar study incorporating what they called subliminal sending. In their study, a sender viewed a target slide 10 times, tachistoscopically for 1 ms each time. No significant psi effect was found in this study. Roney-Dougal (1987) conducted a ganzfeld experiment comparing psi and PWA stimuli. She found that using this procedure, both psi and PWA stimuli were perceived at above chance levels.

There was a suggestive positive correlation between the two conditions, and Roney-Dougal did not find any response characteristics that could distinguish psi trials from PWA trials, suggesting that they are processed in very similar ways in a ganzfeld environment. There is another similarity between psi and PWA, and that is in the amount of criticism that they have attracted from researchers sceptical about the claims made. The criticisms of psi research are well documented and need not be outlined here. The biggest criticism of PWA research, and the one that has dogged the field for longest, concerns the question of whether a particular stimulus is truly outside awareness (see Holender, 1986, for the most comprehensive critique of the field). The general objection has been to the need in PWA studies for two measures, one indexing conscious awareness and one measuring unconscious influences. This is known as the dissociation paradigm and has been criticised because of the difficulties of convincingly demonstrating that one particular measure can represent all conscious awareness. The criticism that PWA has attracted is important in the present context, as it applies to most, if not all, studies conducted by parapsychologists on PWA.

Approaches have been developed in the field of PWA that counter the traditional criticism. These new methods rest on the assumption that most measures of perception are probably influenced by both conscious and unconscious factors. If this is the case, then there must be certain tasks that are differentially affected by unconscious as opposed to conscious influences and are amenable to experimental demonstration of the differential effect. If a single task can be found that leads to different consequences whether it is performed with or without awareness of stimuli, then this obviates the need for two separate tasks, one measuring conscious awareness and one measuring unconscious influence. Various tasks have been found that seem to adhere to this requirement.

In 1989, Jacoby and Whitehouse reported a difference in recognition memory that seemed to rely on whether a biasing stimulus was perceived with or without awareness. False recognition is the term given to the resulting effect, and it is defined as an "old" response to what is actually a new word, on an "old/new" recognition test.

Participants in a PWA/false recognition study are first presented with a long list of words and are then presented with a recognition test. This test consists of a context word followed by a test word (e.g., arena followed by tribe). Participants must decide whether the test word was in the original memory list or not (i.e., old or new). Context words are presented in two conditions: matched (when the context word is the same as the test word, e.g., tribe followed by tribe) and nonmatched (when the context word is different from the test word, e.g., photo followed by tribe). The important effect reported by Jacoby and Whitehouse (1989) was that the matched and nonmatched contexts had qualitatively different effects depending on whether they were viewed with or without awareness. When the context words were presented for a short duration (e.g., 50 ms), a new test word was more likely to be judged as being an old word in the matched context than in the nonmatched context. However, this effect was reversed when context words were presented for a longer duration and assumed to allow conscious awareness. In this situation, context words were presented for 200 ms, and it was found that a new test word was less likely to be judged as being old in the matched context than in the nonmatched context (see Table 1).

Jacoby and Whitehouse (1989) explained this effect in terms of perceptual fluency. Perceptual fluency is an heuristic whereby the more easy (fluent) a stimulus is to process, then the more familiar it will appear in a memory task. Thus, when participants are presented with a context word, this increases the perceptual fluency of that word when it is presented as a test word. When the context word is presented outside of awareness (i.e., at 50 ms), the participant is not aware of where the sense of familiarity originates from and assumes that it must be due to having seen the word in the original study list (thus responding "old" to new words). When the context word is presented at a level that allows conscious awareness, participants tend to attribute too much of the fluency of the test word to the previously seen context word. This leads to a reversal of the previous effect, in that participants will be less inclined to respond "old" in the matched condition than in the nonmatched condition (because every time they feel a sense of familiarity, they are attributing it to the context word).

The PWA/false recognition effect implies that a stimulus outside of awareness can influence short-term memory. Based on the assumption that psi may be processed in a similar manner to weak sensory stimuli, and thus may have comparable effects, what evidence exists that psi can influence memory?

Much of the early experimental work on psi and memory focused on finding correlations between psi performance and memory performance (see, e.g., Feather, 1967; Kanthamani & Rao, 1974, 1975; Rao, 1978). Studies attempting to influence memory performance through psi means are rare. Stanford (1970), however, reported a study in which participants' memory for a story was apparently influenced by nonintentional ESP.

There has been some theoretical work that focuses on the role of memory in the psi process. Roll (1966) proposed a theory of ESP in which ESP responses are actually revived memory traces, activated by an external ESP stimulus. In this account, memory images are seen as "the sense data of ESP" (p. 158). Again, however, the focus of experimental research into this model has been on finding correlations between memory performance and psi performance. The difference between this position and the current hypothesis is that the present study is looking to see whether a psi stimulus can subsequently influence performance on a memory task that immediately follows it. The role of preexisting memories in the psi process is not taken into account. Whereas Roll posited a central role for memory in psi processing, the goals of the current study were more modest. The aim was simply to establish whether psi could influence performance on a memory task in a similar way to how a stimulus perceived without awareness might. The bigger questions concerning the role existing memories might play in psi functioning are not considered.

In summary, the purpose of the present research was to reevaluate the hypothesis that the way in which psi information is processed may be similar to the way in which other weak stimuli are processed. Although these issues have been studied before in parapsychology, the field of perception without awareness has moved on considerably since then. The purpose of the following experiments was twofold. First, a replication of the PWA false recognition effect was carried out at the KPU. This was conducted to establish whether an effect could be obtained and to give a reference point for the second study. A second experiment investigated the hypothesis that psi might have the same effect on recognition memory as a weak sensory stimulus.

EXPERIMENT 1A

Experiment 1a was an attempt to replicate the PWA false recognition effect using a different population sample from Merikle and Joordens's (1997) Study 2a, which itself was an updated version of the original study by Jacoby and Whitehouse (1989). It is hypothesised that, when context words are presented for a short duration (e.g., 50 ms), new test words will be significantly more likely to be judged as being an old word in the matched context than in the nonmatched context. Furthermore, when context words are presented for a longer duration (114 ms), new test words are expected to be less likely to be judged as being old in the matched context than in the nonmatched context.

METHOD

Overview

Participants were presented with a list of five-letter nouns. They were then presented with an old/new recognition test. This test consisted of a masked biasing word followed by a test word. Their task was to decide whether the test word had been in the original list ("old") or was a new word ("new"). The biasing word was presented for one of two durations (50 ms or 114 ms).

Participants

Participants were 29 undergraduate and postgraduate students at the University of Edinburgh, Edinburgh, Scotland. Ages ranged from 17 to 39 (mean age = 20 years). Participants were recruited through tutorial groups and through e-mail appeals for participants. All of the participants had normal or corrected-to-normal vision, and English was their first language. They were told they were taking part in a study of memory.

Materials and apparatus

A total of 540 five-letter nouns were compiled from the Kucera and Francis (1967) tables of word frequency. Word frequency ranged from 3 to 60 occurrences per million, which is standard for this type of study. One hundred and twenty-six words were randomly selected from the pool for the study list, and another 126 were randomly selected as the new words on the old/new recognition task. A further 84 words were selected to be used as the context words on the nonmatch trials of the old/new task.

A Sony monitor (Model GDM 200PST) with a vertical refresh rate of 160 Hz was used to display the stimuli. The graphics card driving the monitor was an ATI XPERT XL (Rage Pro chipset) with 32 Mb video RAM. The E-Prime[R] experiment generator program was used to create and run the program. The display settings were adjusted to use 256 colours, and a screen area of 640 x 480 pixels was set.

Procedure (3)

Participants were met by the experimenter and were then taken to the experimental room. This room was essentially a computer lab, consisting of a number of computers. For the duration of the testing period, the lab was empty, with the exception of the participant and the experimenter, who either sat in the lab in a position where the participant could not see him or waited outside the lab. Participants were told to sit down and make themselves comfortable, and they were requested to adjust the height of the seat so that their eye level was at the centre of the computer screen. The experimenter then verbally described the task, giving the participant an opportunity to ask any questions he or she might have had. The experimenter then started the program running. At this point, the experimenter retired, either to a position in the lab where the participant could not see him or, less frequently, outside the lab. The participant was presented with written instructions on the screen describing the task. The main study was in two parts. The first part involved presenting a study list of 126 words, at a rate of 1 per second, with each word presented for 500 ms followed by a 500-ms blank field. Participants were told to silently read each word to themselves and that their memory for these words would be tested in the second part of the experiment.

The second part of the experiment consisted of the old/new recognition task. Each trial had the following structure: (a) masking stimulus ([pounds sterling] [pounds sterling] [pounds sterling] [pounds sterling] [pounds sterling] [pounds sterling] [pounds sterling]) presented for 500 ms, (b) a context word presented for eithe 50 ms or 114 ms, (c) masking stimulus presented again for 500 ms, and (d) a test word presented until a response was made. All stimuli were size 18 point and presented at a location in the centre of the screen.

Participants were asked to read the context word to themselves if possible and then to respond "old" or "new" to the test word with respect to the initial study list. Recognition memory was tested in three contexts: (a) match (context word and test word are identical; e.g., arena--arena); (b) nonmatch (context word and test word are different, and the context word is a new word; e.g., tribe--arena); and (c) baseline (context word was the letter string xoxoxox). For all three of these contexts, the test words were old on half of the trials and new on the remaining trials. This was determined randomly by the computer.

The old/new recognition test consisted of 12 practice trials followed by 240 experimental trials. Practice trials had tour exemplars of each of the three contexts. The six old test words used for the practice trials were the first and last three words of the study list. The experimental trials consisted of 120 trials in which the context word was displayed for 50 ms and 120 trials in which the context word was presented for 114 ms. There were never more than 3 consecutive trials when the same duration was used.

Once testing was complete, the experimenter informed the participant that the experiment was over and answered any questions the participant wished to ask. This often took the form of explaining to the participant the purpose of the experiment. At no point did any participant inform the experimenter that he or she had worked out what the purpose of the experiment was during the testing period.

RESULTS AND DISCUSSION

A 3 x 2 repeated measures analysis of variance (ANOVA) revealed a significant main effect of context, F(2, 58) = 3.50, p < .05, and a significant interaction between duration and context, F(2, 58) = 14.51, p < .01. Further analysis of this interaction revealed that in the 50-ms condition, there were significantly more "old" responses to new words in the matched context than in the nonmatched context, t(28) = 4.797, p < .001.

In the 114-ms condition, this result was reversed, with significantly more "old" responses in the nonmatch condition than in the match condition, t(28) = 1.706, p < .05. Figure 1 demonstrates the nature of the interaction. These results confirm the hypotheses and are in accordance with previous studies (e.g., Jacoby & Whitehouse, 1989; Merikle & Joordens, 1997). They demonstrate a qualitative difference in recognition memory depending on whether the context stimulus was viewed with or without awareness.

[FIGURE 1 OMITTED]

EXPERIMENT 1B

Having confirmed that recognition memory can be influenced when a biasing stimulus is perceived outside of awareness, the aim of Experiment 1b was to obtain a similar false recognition effect using psi as a biasing stimulus. In this study, the biasing stimulus was presented to a sender who was isolated from the participant and who was attempting to influence the participant's responses on selected trials. If information obtained through psi is subsequently processed in a similar way as PWA, then it would be logical to expect it to have similar influences on cognitive processes, in this case recognition memory.

The present experiment is, as far as the author is aware, the first study looking at the effect of psi on recognition memory. The main hypothesis was that there would be more "old" responses to new words when a sender had sent the target word in the 5s immediately prior to the trial. A nondirectional exploratory hypothesis concerning reaction time was also included, stating that a difference would be observed in reaction times between the sending and nonsending conditions. A pilot test conducted prior to the main experiment suggested that there might be a gender effect in this task. This is something that previously has been found in parapsychology, using different kinds of ESP tasks (see, e.g., Dalton & Utts, 1995; Freeman, 1967). In accordance with this pilot data, it was hypothesised that female participants would display more false recognition in the sending than in the nonsending condition.

METHOD

Participants

Forty participants took part, ranging in age from 18 to 35 years old (mean age = 22 years). There were 25 female and 15 male participants. All of the participants had normal or corrected-to-normal vision, were native English speakers, and had not taken part in Experiment 1a. They were recruited in the same manner as Experiment 1a. Again, they were told that they were taking part in a study on memory.

Materials and apparatus

From the initial 540 word pool used in Experiment 1a, 120 words were selected randomly to serve as the study list, and a further 100 were selected to serve as potential new words in the old/new recognition test.

A button box was used to register responses. This was different from the one used before (due to the different software used in the creation of the experiment). This button box consisted of two buttons, labelled "old" and "new," and was small enough to be comfortably held in one hand during the testing procedure.

Procedure

The setup was different from that of Experiment 1a because the experiment was carried out in a different room. As the present study required a sender, the experimental suite at the Koestler Parapsychology Unit was used. These rooms have been previously used in various ganzfeld experiments and consist of a sensorially shielded room where participants are located. The sender's room is approximately 25 m away (see Figure 2) and up a flight of stairs (see Dalton et al., 1996, for a full discussion of the security issues involved in this setup).

[FIGURE 2 OMITTED]

Participants were met by the experimenter, taken into the experimental room, and asked to sit down and make themselves comfortable. The experimenter then described the general procedure. Participants were told that they would initially see a list of words and that their memory for these words would be tested. They were told that they would be given an old/new recognition task, consisting of a black spot followed by a word. Participants were told to watch the black spot and relax for the duration it was on the screen. They were told to respond to the word, indicating whether they thought it was old or new. If they were unsure, participants were instructed to go with their "gut instinct." They were then asked if they had any questions. After this, the participant was given the button box to hold. It was then explained that the experiment took a total of 10 min alter which the experimenter would re-enter the room and debrief the participant.

The experimenter then started the program in the experimenter's room and retreated to the sender's room while the participant viewed the initial word list. The experimenter's room was locked during this period, and the key remained with the experimenter. Thus, the participant did not have any normal access to what was being sent during the experiment. The study list was presented in the centre of the screen, at a rate of one word per second with each word presented for 500 ms with a 500-ms blank field.

The second part of the experiment consisted of the old/new recognition test. On each trial, participants viewed a black fixation spot for 5s, before a test word appeared. The task was to decide whether the test word was in the original list (old) or not (new), and responses were made on a two-button box. The difference between this and the old/new task in Experiment 1a was that no context word was presented in the present task.

There were 80 trials in total, consisting of 40 "old" trials and 40 "new" trials, randomly selected by computer. Half of the new trials were selected by the computer as being psi trials (i.e., 20 psi trials). On these trials, the word and a visual image related to the test word was shown to the sender during the 5s before the participant viewed that word. The sender attempted to "send" this word, and thus create a sense of false recognition in the participant. (The author was always the sender in this study. I have had limited experience as a sender, conducting one previous study in which the results were mixed.)

RESULTS

Table 2 shows mean number of false recognition responses in the sending and nonsending conditions and mean reaction times in the respective conditions. There is a slightly higher mean in the sending condition, but this difference is extremely small. Likewise, reaction times are slightly faster in the sending condition, but again this difference is small.

Table 3 shows the mean number of false recognition responses for male and female participants in each condition. Whereas female participants appear to show slightly more instances of false recognition in the sending condition, the opposite is true of male participants.

Table 4 displays the reaction times in each condition for male and female participants. Female participants appear to be faster in the sending condition, whereas the opposite is true for male participants. Again, however, these differences are extremely small.

Although the hypotheses in this study were directional, all tests were two-tailed. This was to detect any potential psi-missing that may have been present in the data.

A 2 x 2 mixed ANOVA was carried out on the false recognition data, with condition (sending vs. nonsending) and gender as the variables. This revealed no main effect for condition, F(1, 38) = 0.001, p = .98; no main effect for gender, F(1, 38) = 0.00, p = .99; and no interaction, F(1, 38) = 2.65, p = .11. In addition to this, power and effect sizes were also calculated. These results can be seen in Table 5.

To test the reaction time data, a 2 x 2 mixed ANOVA was conducted to test the two-tailed hypothesis that reaction time might be influenced according to the sending/nonsending conditions. Again, the results revealed no main effect for condition, F(1, 38) = 2.36, p = .63; no main effect for gender, F(1, 38) = 1.43, p = .24; and no interaction effects, F(1, 38) = 2.74, p =. 11. Power and effect size calculations were carried out, and these can be seen in Table 6.

All of these results are in accordance with the null hypothesis. It is, however, interesting to note that, had a one-tailed p been used, both interactions would approach significance.

GENERAL DISCUSSION

The results of Experiment 1a were as hypothesised. Participants responded differently when the context stimuli were viewed for short periods as opposed to long periods. The author concurs with the original explanation offered by Jacoby and Whitehouse (1989) that this is likely due to perceptual fluency. The results of Experiment 1b were not as predicted, however. Experiment 1b was based on the premise that, once in the system, psi may work in a similar way to PWA and thus may have an influence on the same processes that PWA has been found to affect. The false recognition effect in PWA is fairly robust, as demonstrated by Experiment 1a. To my knowledge, this was the first time that recognition memory had been used as a possible mediator for psi. The overall results were nonsignificant. The first explanation for this pattern of results is that there was no psi happening at all. Although this may seem like an obvious conclusion, it is important to point out. It is also possible that, if indeed there was a psi effect present, then it was too weak to be detected. As can be seen from Table 5, the effect size related to the psi condition was negligible.

One other possible reason for the nonsignificant results is the small number of false recognition responses elicited during each session. In Experiment 1a, the overall mean number of "old" responses was 16.65, whereas in Experiment 1b the mean was 5.58. This is a considerable difference and most likely reflects the reduced number of trials in Experiment 1b.

Another, related, explanation for the nonsignificant results in Experiment 1b is that the participants' good performance may have served to obscure any psi that may have been happening, through a ceiling effect. Participants made only a few mistakes per session, so the quality of their memory performance may have overridden any other (i.e., psi) factor. Experiment 1b consisted of fewer trials than Experiment 1a (240 trials in Experiment 1a, 80 trials in Experiment 1b--20 of these being psi trials), due to the time required for "sending." This may have been an important factor as it may not have been a sufficient number of trials to allow enough mistakes or uncertainties that might have given rise to the effect. For example, if there are many trials, then it is likely that a participant will be unsure of the correct response more often. When this is the case, then they are more likely to use secondary cues (e.g., a libeling of familiarity) to be able to respond. This may be where any false recognition effect comes to the surface. If, however, there are too few trials, then memory performance may be good, and the need to use secondary cues is substantially reduced, thus reducing any effect. It is, therefore, possible that the test itself was not powerful enough to pick up any psi that may have been present.

An interesting pattern of results emerged when gender effects were analysed. A short pilot version of Experiment 1b seemed to suggest that women displayed more psi than men. While this may seem an odd finding, it has some precedent. Dalton and Utts (1995) reported that in the PRL (Psychophysical Research Laboratories) ganzfeld database, it was found that male/female, sender/receiver pairs performed best, followed by female/female pairs, with male/male pairs performing worst. As the author always took on the role of sender in Experiment 1b, then this may begin to make some sense. Another aspect of Experiment 1b was that this pattern of results held true when the reaction time data was analysed. It may be of note that there was, again, a small (nonsignificant) interaction between male and female participants. Male participants performed faster in the nonsent condition, whereas female participants tended to respond faster to sent items. Again, this is nonsignificant. It is difficult to speculate about why reaction time might differ between conditions and between genders. There was no real a priori reason to expect reaction times to differ in a particular direction. On the one hand, it could be argued that reaction time should be quicker when the sender is sending, due to the "helping hand" given to the decision. This is in accordance to what one would expect if the psi information is unconscious in the same sense that other weak perceptual stimuli are. In these cases, the response to an unconscious stimulus is automatic, and thus faster than any response in which conscious perception mediates the response. On the other hand, it may be argued that the reaction time may be slower in the sent condition, due to the fact that the sender is sending erroneous information (i.e., trying to influence the participant to say "old" to a new word). This may slow down the decision as the participant struggles with conflicting information.

The above is, it must be noted, purely hypothetical as none of the gender interactions were significant. The effect sizes for both interaction effects were small, and the power for both interactions was only .35 and .365, respectively, suggesting that a more powerful test of these interactions is required. However, the fact that gender and psi interactions persisted in two different measures of psi is interesting in itself, as is the fact that the effect sizes of each interaction are so similar.

Any further work looking at the effect of psi on recognition memory should be aware of the difficulty in maximising power. If psi exists, then it is likely to be a weak effect and difficult to detect. This requires that power be maximised in psi studies, by either increasing the number of trials or increasing the number of participants. In the present study, it may not have been productive to increase the number of trials, as this would lead to extremely long, arduous sessions. Many participants in Experiment 1b commented on the difficulty that they experienced performing the task continuously. Having a break during the session was considered but was not implemented as it was felt that this may have an adverse effect on the memory task. Another issue to be considered is the time interval between trials, which was 5s in Experiment 1b. The reason for this was to give the sender time to send. Reducing the amount of time could increase the number of trials, and thus increase the power. However, if the sending time is reduced, then it may have a detrimental effect on the sender, who is attempting to influence the participant during this period. Indeed, an argument could be made for increasing this time for sending purposes. This is an issue that deserves further investigation.

On a more general note, it is hoped that the present study will encourage parapsychologists to take a flesh look at the parallels between their field and that of PWA. Although this has been done in the past, the field of PWA has since undergone a methodological revolution, of which the majority of parapsychologists are unaware. If parapsychologists are to make comparisons between psi and other forms of perception/cognition, then it is important that they are aware of developments in the relevant areas. This article is an attempt to do that within the field of PWA.

In conclusion, while the demonstration of PWA and its effects on recognition memory were successfully replicated, an attempt to obtain a similar effect using psi failed.

TABLE 1
PROBABILITY OF RESPONDING "OLD" TO A NEW TEST WORD
IN JACOBY AND WHITEHOUSE'S EXPERIMENT 1
(ADAPTED FROM JACOBY & WHITEHOUSE, 1989)

Condition         Match   Nonmatch   Control

50 ms (unaware)   0.24      0.29      0.23
200 ms (aware)    0.36      0.19      0.26

TABLE 2
OVERALL MEAN "OLD" RESPONSES TO NEW WORDS,
REACTION TIMES, AND STANDARD DEVIATIONS

                                Sending     Nonsending

Measure                         M     SD      M     SD      N

"Old" responses to new words  5.53   3.04   5.30   3.07    40
Reaction time (in seconds)    2.04   1.17   2.17   0.95    40

TABLE 3
DIFFERENCES BETWEEN MALE AND FEMALE PARTICIPANTS IN MEAN
NUMBER OF "OLD" RESPONSES TO NEW ITEMS

Mean no. of "old" responses to new words

                                Sending     Nonsending

Gender                          M     SD      M     SD      N

Male                          5.08    3.4   5.76   3.66    13
Female                        5.74   2.89   5.07   2.79    27

TABLE 4
MEAN REACTION TIME DATA FOR MALE AND FEMALE
PARTICIPANTS IN EACH CONDITION

Reaction times (in seconds)

                                Sending     Nonsending

Gender                          M     SD      M     SD      N

Male                          2.45   2.36   2.31   1.04    13
Female                        1.85   1.04   2.11   0.91    27

TABLE 5
EFFECT SIZE AND POWER CALCULATIONS FOR PSI SCORES

Effect                      Effect size ([eta.sup.2])  Power

Psi (sent vs. nonsent)                0.00              .05
Gender                                0.00              .05
Psi x Gender                          0.07              .35

TABLE 6
EFFECT SIZE AND POWER CALCULATIONS
FOR REACTION TIME (RT) DATA

Effect                      Effect size ([eta.sup.2])  Power

Psi RT (sent vs. nonsent)            0.006             .076
Gender RT                            0.036             .215
Psi RT x Gender RT                   0.067             .365

I would like to thank Fundacao Bial and the Society for Psychical Research for funding this research, Paul Stevens for creating the program in Experiment 1b, Alan Marshall for helping conduct the characterisation in Experiment 1a, and Caroline Watt, Fiona Steinkamp, Bob Morris, and Pete Lamont for commenting on earlier drafts of this article.

(1) PWA is the current term for what has previously been called subliminal/implicit/ unconscious perception and preconscious processing.

(2) In this article, unconscious is defined as a cognitive unconscious, dealing with mental processes outside of awareness. This is not necessarily the same as any form of psychodynamic unconscious as posited by Freud, Jung, and others.

(3) Prior to the creation of the program, a characterisation of the monitor was conducted. This used a BBW21 silicon photodiode, which has sensitivity similar to the human eye. A Tektronic 5000 series oscilloscope was used to measure the number of screen refreshes detected by the photodiode. A photodiode was attached to the monitor to measure light output. A variety of stimulus presentation times were then tested. It was found that, at extremely fast presentations, the actual output differed from what was requested from the program. However, the true output followed the screen refresh rate of the monitor. As the screen refresh rate was 12.5 ms, then the actual presentation times were a multiple of this. The characterisation allowed me to determine for exactly how long any particular stimulus was presented. Thus, it could be claimed with confidence that the timing of the stimulus presentation in this study was accurate.

REFERENCES

BELOFF, J. (1972). The subliminal and the extrasensory. In A. Angoff & B. Shapin (Eds.), Parapsychology and the sciences (pp. 103-115). New York: Parapsychology Foundation.

BROUGHTON, R. S. (1988). If you want to know how it works, first find out what it's for. In D. H. Weiner & R. L. Morris (Eds.), Research in parapsychology 1987 (pp. 187-202). Metuchen, NJ: Scarecrow Press.

DALTON, K. (1997). Exploring the links: Creativity and psi in the ganzfeld. Proceedings of Presented Papers: The Parapsychological Association 40th Annual Convention, 119-131.

DALTON, K. S., MORRIS, R. L., DELANOY, D. L., RADIN, D. I., TAYLOR, R., & WISEMAN, R. (1996). Security measures in an automated ganzfeld system. Journal of Parapsychology, 60, 129-147.

DALTON, K., & UTTS, J. (1995). Sex pairings, target type and geomagnetism in the PRL automated ganzfeld series. Proceedings of Presented Papers: The Parapsychological Association 38th Annual Convention, 99-112.

DIXON, N. F. (1979). Subliminal perception and parapsychology: Points of contact. Parapsychology Review, 10, 1-6.

DIXON, N. F. (1981). Preconscious processing. New York: Wiley.

EAGLE, M. (1962). Personality correlates of sensitivity to subliminal stimulation. Journal of Nervous and Mental Disease, 134, 1-17.

FEATHER, S. R. (1967). A quantitative comparison of memory and psi. Journal of Parapsychology, 31, 93-98.

FREEMAN, J. A. (1967). Sex differences, target arrangement, and primary mental abilities. Journal of Parapsychology, 31, 271-279.

HOLENDER, D. (1986). Semantic activation without conscious identification in dichotic listening, parafoveal vision, and visual masking: A survey and appraisal. Behavioural and Brain Sciences, 9, 1-23.

HONORTON, C. (1977). Psi and internal attention states. In B. B. Wolman (Ed.), Handbook of parapsychology (pp. 435-472). New York: Van Nostrand Reinhold.

JACOBY, L. L., & WHITEHOUSE, K. (1989). An illusion of memory: False recognition influenced by unconscious perception. Journal of Experimental Psychology: General, 122, 139-154.

KANTHAMANI, H., & PALMER, J. (1993). A ganzfeld experiment with "subliminal sending." Journal of Parapsychology, 57, 241-257.

KANTHAMANI, H., & RAO, H. (1974). A study of memory-ESP relationships using linguistic forms. Journal of Parapsychology, 38, 286-300.

KANTHAMANI, H., & RAO, H. (1975). The role of association strength in memory-ESP interaction. Journal of Parapsychology, 39, 1-11.

KREITLER, H., & KREITLER, S. (1972). Does extrasensory perception affect psychological experiments? Journal of Parapsychology, 36, 1-45.

KUCERA, H., & FRANCIS, W. N. (1967). Computational analysis of present-day American English. Providence, RI: Brown University Press.

MERIKLE, P. M., & JOORDENS, S. (1997). Parallels between perception without awareness and perception without attention. Consciousness and Cognition, 6, 219-236.

MILLER, J. G. (1940). The role of motivation in learning without awareness. American Journal of Psychology, 53, 229-239.

MYERS, F. W. H. (1903/1955). Human personality and its survival of bodily death. Longmans Green/Garret Publications.

NASH, C. B. (1986). Comparison of subliminal and extrasensory perception. Journal of the Society for Psychical Research, 53, 435-453.

RAO, K. (1978). Further studies of memory and ESP. Journal of Parapsychology, 42, 167-178.

ROLL, W.G. (1966). ESP and memory. International Journal of Neuropsychiatry, 2, 505-512.

RONEY-DOUGAL, S. M. (1981). The interface between psi and subliminal perception. Parapsychology Review, 12(4), 12-18.

RONEY-DOUGAL, S. M. (1986). Subliminal and psi perception: A review of the literature. Journal of the Society for Psychical Research, 53, 405-434.

RONEY-DOUGAL, S. M. (1987). A comparison of psi and subliminal perception: Exploratory and follow-up studies. Journal of the American Society for Psychical Research, 81, 141-181.

SCHMEIDLER, G. R. (1986). Subliminal perception and ESP: Order in diversity? Journal of the American Society for Psychical Research, 80, 241-264.

SMITH, M., TREMMEL, L., & HONORTON, C. (1976). A comparison of psi and weak sensory influences on ganzfeld mentation [Abstract]. In J. D. Morris, W. G. Roll, & R. L. Morris (Eds.), Research in parapsychology 1975 (pp. 191-194). Metuchen, NJ: Scarecrow Press.

STANFORD, R. G. (1970). Extrasensory effects upon memory. Journal of the American Society for Psychical Research, 64, 161-185.

STANFORD, R. G. (1990). An experimentally testable model for spontaneous psi events: A review of related evidence and concepts from parapsychology and other sciences. In S. Krippner (Ed.), Advances in parapsychological research 6 (pp. 54-161). Jefferson, NC: McFarland.

Department of Psychology

University of Edinburgh

7 George Square Edinburgh EH8 9JZ Scotland, UK

swilson@moebius.psy.ed.ac.uk

COPYRIGHT 2002 Parapsychology Press
COPYRIGHT 2003 Gale Group