Printer Friendly

Effect of a threatening subliminal stimulus on the perceptual ESP test: a partial replication.

When I was working in the Netherlands, Martin Johnson and I conducted an elaborate experiment that used a perceptually based ESP test (Palmer & Johnson, 1991) inspired by a procedure of Barbara Lovitts (1981). A slide with a quasi-random assortment of the lowercase letters a-d was flashed to subjects for 150 ms, and they were asked to respond with the letter that seemed most salient to them when the slide flashed. This was probably, but not necessarily, the letter that was in the center of the visual field. The idea was that a paranormal process would guide the subject's eye movements in such a way that the fixated or otherwise most salient letter would match a randomly selected target letter being viewed by an agent. The broader purpose of the procedure was to minimize the often intellectual kind of mental activity associated with more traditional ESP tests in which subjects are required to conjure up an image or impression. There are indications in the literature that such thought processes are psi-inhibiting, especially in forced-choice ESP tests (Stanford, 1975).

Immediately prior to the presentation of the letter slide, a subliminal picture

was flashed to the subject. The picture was either the standard Defense Mechanism Test (DMT) slide (Kragh & Smith, 1970) in which a threatening male figure is peering down at a boy seated at a desk, or a control slide in which the countenance of the male figure in the background is pleasant and smiling. We expected more negative ESP scores when the face was threatening than when it was smiling. The hypothesis was not confirmed, but two post hoc findings suggested to us that there might have been some ESP in the data. First, there was overall tight variance, which means that the scores clustered more closely around mean chance expectation (MCE) than one would expect by chance. Tight variance has been associated with a negative mood in the literature (Rogers, 1966, 1967). The puzzling aspect of this finding was that the tight variance was present with the smiling subliminal picture as well as the threatening one. We offered two possible explanations. First, the smiling picture may not have been as benign as we originally thought; for example, we thought in retrospect that it could be construed as lecherous, especially by females. Second, we thought that a possibly negative context of the experiment as a whole could have caused the tight variance. For instance, to fulfill the requirements of a visual field manipulation, subjects had to wear uncomfortable goggles and place their chin on a chin rest during the ESP task.

Second, there was a confirmation of Stanford's (1975) response-bias hypothesis. In cases where a particular letter was called six times or less out of the 40 trials, the number of hits on these undercalled trials was significantly above chance and significantly greater than on the remaining trials.

Because one purpose of the perceptual ESP test is to minimize linear thought processes, I had hoped that there would be less evidence of response biases with this procedure than with a traditional forced-choice ESP test. It seemed that such biases were indeed reduced, particularly the undercalling of XX doublets, but the biases were still present to some degree.

The present experiment is a partial replication of the Palmer and Johnson experiment. The intent was to reproduce the situation in which scoring on the perceptual ESP test can be influenced by the immediately prior subliminal presentation of a threatening stimulus, the standard DMT slide. It was also intended to make the test situation more pleasant for subjects than it was in the Palmer and Johnson experiment, primarily by eliminating the visual field manipulation. Also, the sample was restricted to persons genuinely interested in the research; most subjects in the earlier study participated to receive course credits in a psychology class. Thus, one alternative explanation of the tight variance effect found by Palmer and Johnson was removed. Other modifications in the procedure were also introduced, but there was no reason to expect that any of these changes would affect the hypothesized process, namely, that the subliminal presentation of a threatening stimulus immediately before each ESP trial would produce tight variance as a defensive response to the stimulus.

These other modifications are:

1. A clairvoyance procedure was used; there was no agent. My motive here was purely a matter of economics; the clairvoyance procedure did not require us to recruit agents, and no staff members were available for this role.

2. There was no break between the two 20-trial runs, leaving a single run of 40 trials. In the Palmer and Johnson experiment, a break was considered necessary to relieve subjects from having to continuously place their chin on the chin rest.

3. The targets were changed from letters to typewriter carets, or Vs, pointing up, down, left, or right. I chose the carets over the letters because they were physically more uniform, and thus I thought they might reduce response biases. I also thought they might have fewer semantic associations than the letters, especially for students, for whom the letters corresponded to academic grades.

4. There was no DMT slide with the smiling face. In the Palmer and Johnson experiment, this slide had the same effect for tight variance as the standard DMT slide did. Moreover, the slide with the smiling face was not available.

5. On trials where subjects missed the flash of the target slide or could not decide on a response, they were given the opportunity to pass.

6. There was no exchange of token objects.

There was also little reason to expect that any of these modifications would eliminate the response-bias effect, with one possible exception. If the choice of carets over letters indeed reduced response biases, it is conceivable that a necessary condition for Stanford's response-bias effect would be removed. The response-bias effect presupposes that there are some target alternatives that are purposefully (even though perhaps unconsciously) avoided. If this condition is not met, some targets may still be undercalled occasionally, but this would be part of the vagaries that occur in any truly random sequence from time to time and thus would not be expected to favor hits on "counterbias" responses. Fortunately, the likelihood of this condition's not being met can be assessed empirically by simply noting if response biases are weaker in the present experiment than in the previous one.

In conclusion, the following two hypotheses were proposed:

1. The variance of subjects' ESP hit totals will be significantly low.

2. The number of ESP hits will be significantly above chance on targets called six times or less in the 40-trial run.

A significant outcome for these hypothesis was defined as p |is less than~ .05, one-tailed. The criterion of significance for all other analyses was to be p |is less than~ .05, two-tailed.



The number of subjects was set in advance at 40. Although this is only half the number of subjects in the Palmer and Johnson experiment, I did not wish to commit additional resources until I could establish that the results were promising. Subjects were recruited by ads in the Duke University student newspaper, fliers put up on the Duke campus, and talks on parapsychology given by FRNM staff to student dormitory groups. They were told only that the experiment concerned ESP and subliminal perception. There were 20 males and 20 females.

General Layout

There were two test rooms, both located on the second floor of the Institute for Parapsychology. The subject was seated in a reclining chair in Room 1. The experimenter (J. P.) was located in Room 2, which is adjacent to Room 1 and separated from it by a door and a 60 mm X 45 mm double-glass window. Room 2 housed the projector and computer. The door between the two rooms was closed during the test sessions. There was two-way intercom communication between the rooms, but the experimenter had no knowledge of the target sequences.

Equipment and Specific Layout

Stimuli were presented to the subjects by means of a two-channel projection tachistoscope (Gerbrands, model G1170). Each channel consists of a standard Kodak Ektagraphic III B slide projector fitted with a tachistoscopic shutter. Both channels were controlled by an IBM-PC-type microcomputer and custom-made interface hardware.

The projector presented visual stimuli to the subject in Room 1, using a 250-W halogen lamp and the low-brightness setting. One channel projected the subliminal stimulus, and the other projected the caret slide. The stimuli first passed through two pieces of polarizing filter, the axes of which subtended an angle of 72|degrees~. These were pasted directly on the shutter housing. Images next passed through the double window into Room 1 and then through a 12-mm-thick ground-glass (milk) screen mounted on a stand inside the room. The double window was covered in Room 1 by a wooden screen into which a 9 cm X 12 cm rectangle had been cut. The milk screen was covered by a piece of black cardboard into which had been cut a 12 cm X 14 cm rectangle that defined the screen where the stimuli were exposed. The subject was seated 2 m in front of the screen.

Stimulus Slides

The caret slide was made by photographing a sheet of white paper onto which had been rubbed black caret symbols (Geotype GS-106). The slide was darkened by photographing the paper at a high f-stop (f-11), using black and white film. It consisted of a display of 28 carets arranged in staggered rows, so that 7 pointed up, 7 pointed down, 7 pointed left, and 7 right in a quasi Latin-Square format that resulted in each of the directions being represented in all areas of the slide.

The main subliminal stimulus was a standard DMT slide depicting a boy seated at a desk and a man of threatening demeanor peering over his right shoulder from the rear. The slide was darkened to about the same degree as the caret slide. A second subliminal stimulus, which is used as a control in the DMT, depicts two boys, one with a red shirt, standing against a wall. This slide was used for practice trials.


All targets for the experiment were selected by J. P., using a pseudorandom algorithm implemented by Turbo-Basic software. The sequence was seeded by the number of microseconds between two keyboard button-presses. The total number of targets generated was 1,760. This total consists of 40 trials for each of the 40 subjects, plus four additional 40-trial sequences, which were added so that if a particular run had to be aborted, a fresh target sequence could be substituted.

The total sequence was evaluated by chi-square tests for singlet bias, singlet bias summed over runs, doublet bias, and bias of XX doublets. These analyses were performed by the computer without the individual targets' being exposed. The system is set up such that if any of the chi-square values approaches or reaches significance, a new set of targets can readily be generated. This procedure was repeated two or three times until all statistical tests gave outcomes equivalent to less than |+ or -~ 1.5 standard deviations from chance expectancy.

Caret Test

The procedure for the caret test was as follows: Subjects initiated each trial by pressing a green button in the middle of a response box resting on their lap and then immediately looking up at the screen. Two seconds after the button-press, the subliminal DMT stimulus was flashed for a duration of 10 ms. Subjects were then asked to relax their gaze and allow their eyes to wander around the screen. Two seconds after the DMT slide flashed, the caret slide flashed for 150 ms. The subjects then chose an ESP response based on which particular caret was most salient in their visual field when the slide flashed. For example, if the most salient caret was pointing up (irrespective of its location on the slide), the subject would press the "up" button on the response box. If the subject missed the flash or for some other reason was unable to make a choice, a "pass" option was available: By pressing a red button on the response box, the subject could cause the trial to recycle with the target remaining the same. The run consisted of 40 successive trials. The end of the run was signified by the lighting of the red button by itself.


Each subject completed two questionnaires as part of the procedure. The first was the 26-item Questionnaire of Experiences of Dissociation (Riley, 1988). The second was the 20-item trait portion of Spielberger's State-Trait Anxiety Inventory (Spielberger, 1983). These questionnaires played a prominent role in a separate caret test experiment that has not yet been published. They were introduced here for purposes of comparison with that study.

The subjects also filled out an 8-item rating scale on which they described their reactions to, and approach in, responding to the caret test, as well as their guesses as to the content of the DMT slide.


Most subjects were first given a tour of the lab by staff member Kathy Dalton, which included a few trials on one of our computerized psi games. Then the subject was greeted by J. P. and accompanied to Room 1. After a few minutes of friendly conversation, J. P. explained the procedure to the subject. It was mentioned that prior to each trial a subliminal picture would be flashed on the screen, but nothing else about the picture was divulged. Then the caret test was described. Finally, the subject signed a consent form and completed the dissociation and anxiety questionnaires.

As soon as the subject signalled completion of the questionnaires, J. P. dimmed the lights slightly so that the ambient lighting was comparable to that in the Palmer and Johnson experiment. Prior to the formal 40-trial ESP test, 10 "warm-up" trials were undertaken with the neutral subliminal picture.

After the run, J. P. brightened the lights in Room 1 and requested the subject to complete the rating scale. When this was finished, J. P. dimmed the lights and revealed the two subliminal slides to the subject at full exposure, asking the subject in each case to describe the picture. For the experimental DMT slide, the subject was probed about the expression or mood of the man in the background, unless such comments were offered spontaneously. J. P. then showed a copy of the ESP results to the subject, explained the rationale of the study, and answered any questions the subject might have. Finally, he thanked the subject for participating and asked that he or she not discuss details of the experiment with any possible future subject.


Subliminality of DMT Slide

An effort was made to duplicate the level of identifiability of the DMT slide that was present in the Palmer and Johnson experiment, but in that experiment subjects were not asked to identify the content of the slide following the ESP task. In the present experiment, 28 of the 40 subjects (70%) were able to identify at that time the large image in the foreground of the slide (the boy) as a person. However, none of the subjects were able to identify the image in the background even as a living being, and no one interpreted the scene as depicting any kind of negative emotion or threat. The one subject who attributed affect to the slide described it as "happy." Thus, the essential aspects of the slide as a depiction of threat were subliminal for all subjects, even after repeated exposures.

Basic Parameters and Hypotheses

The mean ESP hit score of the 40 subjects was 10.33 (SD = 2.28). Compared with the MCE of 10, this mean is in the psi-hitting direction but clearly nonsignificant, CR = 0.90.

Tight variance. The variance of the scores around MCE was 5.31. This value is lower than the expected variance of 7.5 and thus is in the direction predicted by the first hypothesis. However, the value does not quite reach significance, CR = 1.46, p = .072, one-tailed.

The results of a chi-square goodness-of-fit test also approached significance,(1) ||chi~.sub.2~ (4) = 9.05, p = .060. Table 2 illustrates the corresponding outcome from the Palmer and Johnson experiment. The two patterns are quite similar structurally. In particular, in both studies there is a considerable excess of |+ or -~ 1 deviation scores that is not matched either by 0 deviation scores or by |+ or -~ 2 deviation scores.


Response bias. Of the 40 subjects, 18 (45%) called at least one of the symbols six times or less during the run. The total number of counterbias responses was 108, of which 31 were hits, 28.7%. Although this is in the direction predicted by the second hypothesis, the value is clearly nonsignificant, CR = 0.78.

A chi-square goodness-of-fit test was computed for each subject's distribution of calls by comparing the observed number with the expected number of each call. This chi-square value provided a measure of the subject's degree of call imbalancing and is identical to RB1 in the Palmer and Johnson experiment. The mean of these RB1 values in the present experiment was 4.28, compared with 3.89 in the previous study.

The pooled data also revealed a significant tendency for some caret directions to be called more frequently than others. The down-pointing caret was called most frequently (457 times, or an average of 11.43 times per test). The left-pointing caret was called least frequently (325 times, 8.13 times per test). The other directions were called with close to chance frequencies: 403 times (10.08) for "up," and 415 times (10.38) for "right." A chi-square goodness-of-fit test revealed that this distribution differed from that expected by chance TABULAR DATA OMITTED to a marked degree, ||chi~.sub.2~(3) = 22.97, p = 4 X |10.sup.-5~, as it did in the Palmer and Johnson experiment.

Post Hoc Analyses

For the purpose of these analyses, a "variance" score for each subject was computed by taking the square root of the absolute deviation of the number of ESP hits from the MCE of 10. (This square root measure was used to normalize the distribution of variances.) It was also decided in advance that a nonparametric statistical test would be used whenever any continuous variable involved in the relationship had either a skewness or a kurtosis value exceeding z = |+ or -~ 1.50.

There were no significant relationships between either ESP hits or ESP variances and the two personality scales that measured dissociative tendencies and anxiety, respectively. However, the latter two measures did correlate significantly with each other, |r.sub.p~(35) = + .43, p = .008.

Interpretation of the DMT slide. Beginning with the third subject, J. P. wrote down subjects' descriptions of the threatening male figure when the DMT slide was fully exposed to them.(2) One subject could not see the man's expression well enough to make a judgment. Although none of the remaining 37 subjects used the word threatening in their descriptions, 23 of them (62%) used related words such as angry, scary, scowling, or stern. The remaining subjects described the man as sad (5), neutral (2), curious (2), afraid (2) serious (1), thoughtful (1), or authoritative (1).

The 23 subjects who attributed hostility to the threatening male figure had a mean ESP variance score of 1.44 (SD = 0.46). The 14 subjects who did not attribute hostility to him had a mean of 0.98 (SD = 0.60), which is significantly lower than the mean of the first group, t(35) = 2.64, p = .012. The actual variance around MCE for the non-hostility group was 2.92, ||chi~.sup.2~(14) = 5.07, p = .015, whereas the variance for the subjects who did attribute hostility to the man was 7.54, very close to the MCE. These results mean that the tight variance is attributable to those subjects who did not see the man in a hostile light.

Difficulty of the test. As planned prior to the study, all rating-scale items were correlated with the ESP hit and variance scores. Two of these items concerned how difficult it was for the subject to take the caret test. The first of these asked, "How difficult did you find it to see individual carets when the caret slide flashed?" (DIFF-1). The second asked, "How difficult did you find it to choose a particular response for each trial?" (DIFF-2). Both answers were measured by having subjects make a mark along a 100-mm line, the score being the number of millimeters between the mark and the left edge of the line. These items were selected to differentiate two aspects of a more global difficulty question that correlated significantly and positively with ESP hits in a previously conducted, not yet published, experiment involving the caret test. Not surprisingly, DIFF-1 and DIFF-2 correlated significantly with each other in the present experiment, |r.sub.s~(38) = + .47, p = .002.

The ESP variance scores correlated significantly with both DIFF-1, |r.sub.s~(38) = + .35, p = .027, and DIFF-2, |r.sub.s~(38) = + .34, p = .033. If the two scales are combined by taking the average of the summed DIFF-1 and DIFF-2 ratings, the correlation with ESP variance is raised to |r.sub.s~(38) = + .41, p = .008. Because the mean variance was well below chance, this finding can best be interpreted to mean that those subjects who found little difficulty in taking the caret test produced tight variance. The mean rating on DIFF-1 was 22.95 (SD = 3.35), and the mean rating on DIFF-2 was 29.50 (SD = 21.08), both on 100-point scales.

DIFF-1 also correlated significantly with ESP hits, |r.sub.s~ (38) = + .32, p = .044, but DIFF-2 did not, |r.sub.s~(38) = + .20. The correlation between ESP hits and the combined difficulty items (DIFF) was not quite significant, |r.sub.s~(38) = + .28, p = .083.

Imagery. An item included on the rating scale for the first time asked subjects: "Did you ever get a visual or mental image of one of the target alternatives before the caret slide flashed?" This question was answered affirmatively by 19 subjects, who had a mean ESP hit score of 11.21 (SD = 1.87), t(18) = 2.82, p = .011. The other 21 subjects had a nonsignificant mean hit score of 9.52 (SD = 2.36). The difference is significant, t(38) = 2.49, p = .017.

Subjects who answered this imagery question affirmatively were asked to estimate the percentage of trials in which such images occurred. The estimates from the 18 subjects who gave them ranged from 5% to 55%, with a mean of 19.22 (SD = 14.78); however, the distribution was positively skewed. This measure did not correlate significantly with either ESP hits, |r.sub.s~ = + .11, or ESP variance, |r.sub.s~ = + .26. However, it did correlate significantly with call imbalancing (RB1), |r.sub.s~ = + .62, p = .011.

These subjects were then asked to estimate the percentage of trials in which their images determined their responses. The estimates of the 15 subjects who ventured them ranged from 0% to 30%, with a mean of 12.03 (SD = 8.37). These estimates correlated significantly with both ESP hits, |r.sub.p~(13) = + .66, p = .008, and ESP variance, |r.sub.p~(13) = + .70, p = .004. The exceptional convergence of results for these two ESP measures derives from the fact that in this subsample they were highly correlated, |r.sub.p~(13) = + .71, p = .003, whereas in the total sample they were not, |r.sub.p~(38) = - .01. The former correlation in turn seems attributable to the fact that the relevant subsample consisted of those subjects who reported pretrial imagery, among whom were a disproportionate number of relatively extreme psi-hitters.

There were no other significant correlations between rating-scale items and ESP scores.


The attempt to replicate the tight variance effect found in the Palmer and Johnson experiment fell short of significance, although it came tantalizingly close. The failure cannot be attributed exclusively to the smaller sample size in the present experiment because the variance of 5.31 is higher than the variance of 3.97 obtained by Palmer and Johnson. On the other hand, if the variance of 5.31 had been maintained over 80 subjects, the outcome would have been significant, CR = 2.10, p = .036, one-tailed. On balance, the present result can be seen as encouraging, and as justification for testing more subjects to see if the effect is confirmed. Moreover, the present results suggest that the negative test conditions introduced by the visual field manipulation and poorly motivated subject sample were not a major factor in producing the significant tight variance found by Palmer and Johnson.

Simply labeling this outcome as tight variance, however, may not be an adequate description. How can one sensibly account for the disproportionate influence of the |+ or -~ 1 deviations on the outcomes in both studies? A close inspection of Tables 1 and 2 reveals that the cause of the failure to replicate was too many scores in the |+ or -~ 4 category in the present experiment. The excess of observed over expected scores in the |+ or -~ 1 category is actually larger in the present experiment than in the previous one, 63.6% vs. 54.5%. If the other categories are collapsed and compared with the |+ or -~ 1 category for the two experiments combined, the result is ||chi~.sup.2~ (1) = 15.09, p = 1.0 X |10.sup.-4~. This magnitude of effect cannot be easily ignored, even though it is post hoc.

I do not see at the present time how this finding could be an artifact. The expected values in Tables 1 and 2 are admittedly rounded, but the effect is more than a rounding error. One can conceive that subjects who knew just a few of the targets psychically could bias their scores toward MCE, but tilting them toward a precise deviation would seem to require more thorough knowledge of the target sequence. Also, why would a subject want to produce a |+ or -~ 1 deviation? If the motive is to produce no measurable psi, might subjects reason unconsciously that a 10 would be too perfect a chance score?

The tight variance effect, such as it is, has tentatively been interpreted as a response to subliminal presentation of a threatening stimulus. However, this interpretation cannot be considered verified until control conditions are introduced that involve either supraliminal presentation of the DMT slide or subliminal presentation of a nonthreatening slide, and these control conditions are shown to eliminate the tight variance effect. When the smiling-face control picture was used in the Palmer and Johnson experiment, the tight variance effect was not eliminated, but this picture may not have been nonthreatening. I did not include a control condition in the present experiment because I did not want to commit additional resources until I had some reason to believe that the basic effect was replicable.

The threat interpretation of the tight variance receives some support, however, from the relationship between perceptions of the DMT slide at full exposure and the ESP variance scores. From a psychodynamic point of view, it can be argued that failure to see the ostensibly threatening male figure in a hostile light is reflective of repression or denial, and it was the subjects who so responded that manifested the tight variance. According to the experimental hypothesis, tight variance is much the same kind of defensive response to threat as is repression or denial.(3) Thus, it makes theoretical sense that the two effects coincide.

The attempt to replicate the response-bias effect from the Palmer and Johnson experiment was more clearly a failure, although if one combines the CRs from the two experiments (+ 2.91, + 0.78) by the Stouffer method, the result is still significant, CR = 2.61, p = .009. The weaker effect in the present experiment cannot be attributed to less biased calling, as the mean RB1 score was actually higher in this experiment. It is difficult to determine the broader evidential status of the response-bias effect, as well as its boundary conditions if real, because the appropriate analyses are reported in only a handful of forced-choice ESP experiments in which they are applicable. With the aid of computers, these analyses are not difficult to do. The response-bias effect, if reliable, is important, both for theoretical reasons and as a means of identifying psi-prone trials.

The preceding discussion also reveals that the change in targets from letters to carets did not reduce subjects' response biases, but neither did it increase them to a notable degree. Perhaps response biases are simply intrinsic to this kind of ESP task, whatever the targets.

Moving on to the post hoc analyses with the rating scale, we found that the tight variance was attributable primarily to subjects who experienced little difficulty in taking the caret test and in choosing their responses. The caret test is supposed to be easy and to require little cognitive exertion to determine a response. Thus, the low-DIFF subjects were the ones who were taking the test the "right" way, or at least in the intended spirit. Inasmuch as tight variance is the closest thing to a significant global effect in the present experiment, this finding could simply mean that the low-DIFF subjects were indeed the ones who produced psi in the test, and because of the DMT stimulus the kind of psi they produced was tight variance.

In any case, the difficulty variable seems to be important. A similar item was the strongest single correlate of overall ESP deviation scores in a major and more complex experiment (unpublished) that also used the caret test. I will postpone discussion of how and if this finding can be reconciled with the DIFF result of the present experiment until the earlier experiment is published.

The conclusion that the only subjects producing psi in this experiment were the low-DIFF subjects must be qualified by the results with the imagery item. Subjects who reported spontaneous target imagery during the test scored significantly above chance and significantly higher than subjects who did not report such imagery. Moreover, among the subjects who reported imagery, the more frequently they based their responses on these images, the higher their ESP scores. The proportion of trials where the imagery applied was uniformly low, but only a couple of real ESP hit trials per run would be necessary to account for the findings.

These findings suggest that subjects sometimes got images of a target spontaneously, and when they acted on those images they tended to be correct and raise their ESP scores. Again, this is not the way one is supposed to take the caret test, and it is conceivable that the proportion of trials reported as being influenced by these images was attenuated because of demand characteristics. Perhaps in future experiments subjects should be explicitly encouraged to base their responses on such images when they occur. More precise data could be collected by having subjects record some note of individual trials mediated by imagery, but such instructional sets might make subjects self-conscious about the images and reduce either their incidence or their effectiveness. Perhaps the main advantage of the perceptual ESP test will prove to be that it allows the emergence of psi-mediating spontaneous imagery by reducing the cognitive demands associated with more traditional forced-choice tests.

Was there any relation between the imagery and difficulty items? The data here are mixed. There was virtually no difference between the imagery and non-imagery subgroups on DIFF, t (38) = .022. However, among the imagery subjects, there was a suggestive positive correlation between DIFF and the estimated proportion of trials in which ESP responses were based on the imagery, |r.sub.s~(13) = + .46, p = .087. If anything, imagery-mediated responses were associated with higher levels of difficulty in taking the caret test as it is supposed to be taken. This would make sense, in that subjects who found the caret test difficult might be tempted to draw on their spontaneous images as an alternative response determinant.

In conclusion, the post hoc analyses seem to suggest that two different psi mechanisms were operating: (a) tight variance--a defensive response--among subjects who did not see the DMT picture as threatening under full exposure and among those who took the caret test with ease, and (b) hitting--a nondefensive response--among subjects who based some of their responses on spontaneous target imagery. However, caution must be exercised in accepting these conclusions wholeheartedly because neither of these effects has yet been replicated to a statistically significant degree. This caution is particularly essential for the imagery effect and the difficulty effect because both findings are post hoc and no replication has yet been attempted. On the other hand, the credibility of this effect is enhanced by its manifesting in two convergent correlations: (a) ESP hits and presence of imagery and (b) ESP hits and how frequently the imagers used their images to determine their responses.

1 Because this finding is a replication of a finding from the Palmer and Johnson experiment, a one-tailed test might be considered justified, which would increase the significance level to p = .030. However, this maneuver would still not formally confirm the tight variance hypothesis, because the goodness-of-fit chi-square was not the primary statistical test of the hypothesis.

2 Both the subject and I were blind to the subject's ESP score when I solicited the description of the DMT slide. However, because the analyses involving these descriptions were unplanned, I had seen the subjects' ESP scores before I classified the descriptions into the two categories. I can say, however, that I paid no attention to individual subjects' ESP scores during any of the analyses of the experiment, that I had no recollection of them during the classification task, and that they did not consciously influence my classifications, which, in any event, have a logical basis. Nonetheless, as a further check, I asked Kathy Dalton, who had not seen the ESP scores, to place the adjectives the subjects used to describe the man into two piles according to whether they did or did not connote threat, hostility, or the likelihood that the man would do something bad or hurtful to the boy. The effect reported in the next paragraph was still significant (p = .026) when her classifications were used.

3 In the Palmer and Johnson experiment, subjects completed the DMT as a psychological test of defensiveness, and the resulting scores were totally uncorrelated with ESP variance scores (|r.sub.p~ = - .00). This could mean that the comparable finding in the present study is not robust, but it could also mean that supraliminal presentation of the DMT slide measures a more relevant aspect of defensiveness than the subliminal presentations used in the DMT as such.


KRAGH, U., & SMITH, G. W. (1970). Perceptgenetic analysis. Lund, Sweden: Gleerups.

LOVITTS, B. E. (1981). The sheep-goat effect turned upside down. Journal of Parapsychology, 45, 293-310.

PALMER, J., & JOHNSON, M. (1991). Defensiveness and brain hemisphere stimulation in a perceptually mediated ESP task. Journal of Parapsychology, 55, 329-348.

RILEY, K. C. (1988). Measurement of dissociation. Journal of Nervous and Mental Disease, 156, 449-450.

ROGERS, D. P. (1966). Negative and positive effect and ESP run-score variance. Journal of Parapsychology, 30, 151-159.

ROGERS, D. P. (1967). An analysis for internal cancellation effects on some low-variance ESP runs. Journal of Parapsychology, 31, 192-197.

SPIELBERGER, C. D. (1983). Manual for the state-trait anxiety inventory (Form Y). Palo Alto, CA: Consulting Psychologists Press.

STANFORD, R. G. (1975). Response patterns in extrasensory performance. Journal of Parapsychology, 25, 153-161.
COPYRIGHT 1992 Parapsychology Press
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Palmer, John
Publication:The Journal of Parapsychology
Date:Sep 1, 1992
Next Article:Psi and the nature of abilities.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters