Printer Friendly



ABSTRACT: We tested a theoretical model and attempted a conceptual replication of a previous study suggesting that people may nonintentionally use psi to help others avoid unpleasant outcomes and instead experience pleasant outcomes. The model predicted: overall helping, measured by performance on a nonintentional psi task; more helping between friends than between strangers; less helping when it was costly to do so; a negative correlation between neuroticism and helping; and a positive correlation between openness and helping. Forty friend pairs and forty stranger pairs participated. Each participant did a nonintentional psi task, performance on which could help the other. There was a nonsignificant tendency towards helping on the nonintentional psi task t(79) 1.117. The effect of the relationship between participant pairs was nonsignificantly contrary to the predicted direction, F(1, 76) = 0.355, and there was no significant effect for cost of helping, F(1, 76) = 0.003. There was a weak correlation in the p redicted direction between neuroticism and helping, r,(79) = -.183, and no correlation was found between openness and helping, r,(79) = -.006. Exploratory analysis compared nonintentional psi and perceived luckiness.

The phrase "you scratch my back, I'll scratch yours" crudely conveys the concept that individuals within species may behave altruistically to one another if there is some kind of adaptive value or payoff for doing so. The influence of evolutionary pressures, and our adaptive response to them, on the development and expression of human characteristics provides a powerful perspective with which to expand our understanding of human psychology. Evolutionary psychology provides a fruitful new approach to many domains, including human social behavior and even human thought and knowledge (e.g., Crawford & Krebs, 1998; Plotkin, 1994).

Within parapsychology too, researchers (e.g., Broughton, 1988) have considered the role of evolutionary factors in the likely characteristics of "psi"--a postulated form of anomalous perception or anomalous communication in which an organism appears to exchange information or influence with its environment without the use of the currently understood senses or inference. This study aimed to test predictions arising from the Psi-Mediated Instrumental Response (PMIR) Model (Stanford 1974a, 1974b, 1977, 1990). The model assumes that the organism uses psi in everyday life to respond to its environment adaptively, in a way that fulfills its needs. For example, commuting home from work, one might follow an impulse to deviate from the normal route, and by so doing avoid the scene of an accident that had blocked the road. According to the PMIR model, psi is considered to operate often in daily life, and largely nonintentionally.

It would be of great value to parapsychologists to identify a theoretical model of psi that might productively guide and focus research efforts, and there have been several attempts explicitly to test the PMIR model's assumptions, as reviewed by Stanford (1990). Stanford's review found that, although supportive of PMIR, many of the studies were flawed. Some of these flaws concern failures to rule out sensory information leakage, such as inadequate procedures for random target selection (the "target" is the concealed information that the organism is hypothesized to perceive anomalously). Other flaws relate to other areas of design, such as the failure to ascertain the efficacy of experimental manipulations of theoretical constructs such as need-strength. In addition, numerous assumptions and predictions from the model have rarely, or never, been tested.

One of these little-tested aspects of the PMIR model concerns the social dimension of nonintentional psi. Based on real-life accounts of "good luck," in which people have helped a friend or relative by following an impulse, Stanford has suggested that interpersonal factors may be important in moderating the operation of nonintentional anomalous psi. The only laboratory study of the importance of social, interpersonal factors in the use of nonintentional psi (Stanford & Rust, 1977) looked at the possibility of using nonintentional psi to help another person avoid an unpleasant task and instead experience a pleasant task. The nonintentional psi measure was a word association task, through which the helper could assist, or hinder, the person with whom they were unknowingly paired. Word association was chosen because the PMIR model postulates that nonintentional psi manifests through normal psychological processes such as memories, preferences and associations, and Stanford's previous experience with this task s uggested it could be a useful vehicle for testing PMIR. On this task, one of the words was randomly designated as the "key" word (the identity of which was unknown to the experimenter or the helper at the time of the task). The experiment was designed with a predetermined contingency: if the helper produced his or her shortest reaction time to the key word, the helpee was directed to the pleasant task; otherwise, the helpee was directed to the unpleasant task. Significantly more individuals were directed to the pleasant task than would be expected by chance, exact binomial p = .026, one-tailed, N = 30; effect size, Cohen's h = .36, thus supporting Stanford's prediction and giving some initial support to the idea that people may quite unknowingly use their psi in order to assist others.

Ways in Which the Present Study Differs from Stanford and Rust

The present study is a conceptual replication and extension of Stanford and Rust's (1977) successful study. Participants in S&R's study had a single role, either helper or helpee. No data were collected from helpees, which we felt was a waste of a valuable resource. Therefore, in our study, each participant fulfilled two roles in a reciprocal way: helper first and then helpee, thus enabling us to gather PMIR data from every participant. We also added two manipulations based on theoretical predictions. Participants were either friend pairs or stranger pairs: the PMIR model predicts that there is more incentive to use psi to help a friend than to help a stranger. Second, we systematically manipulated the "cost" of helping: PMIR predicts less incentive to help when it is costly to the organism to do so. In addition, we devised a new nonintentional psi task, detailed below, designed to avoid statistical problems created in the original task when reaction times were tied. The new task was intended to be conceptua lly similar to S&R's. The PMIR model assumes that nonintentional psi operates through triggering behaviors, memories, feelings, and desires that are already in the organism's repertoire (Stanford, 1990, p. 102). We therefore considered that a task measuring aesthetic preference would provide a valid vehicle for PMIR. Whereas there was a 10% chance of "helping" in S&R's study, we chose to arrange the contingency so that there was a 50% chance of helping. The reason for this was to improve statistical power to enable comparisons between the four different conditions, with similar numbers in each condition. With only 10% of people helping, this would have created a substantial imbalance between the numbers in each condition, making it more difficult to look at our additional theoretical manipulations.

Finally, S&R report two different outcome measures: a discrete, "direct hits" measure, and a continuous z-score measure. Significant results were found for the discrete measure but not for the continuous measure. S&R do not state which of these measures was intended as the primary outcome measure. We chose to use a continuous measure as the primary outcome because we were persuaded by Stanford's argument elsewhere (Stanford & Associates, 1976) that a continuous measure would be more sensitive, both on theoretical and psychometric grounds. However, in order to allow direct comparison with S&R, we also include analyses based on a 'direct hit' indicator for hypotheses 1,2, and 3 (in which participants may be clearly categorized depending on their condition allocation).

We included a questionnaire in order to test predictions relating PMIR and personality. The main personality dimensions of interest to us were neuroticism and openness. Considering the limiting factors to the use of PMIR, Stanford (1990) postulates that the mental factors that in everyday situations favor self-defeating or harmful behavior, such as neuroticism, would also dispose towards systematic misuse of PMIR. Also, considering individual differences that might enhance the tendency to respond to nonintentional psi-mediated impulses, we felt that higher openness (to fantasy, aesthetic experiences, feelings, for instance) might be associated with higher nonintentional psi scores.

The operation of PMIR in everyday life might resemble good luck. Therefore, we included a luckiness questionnaire to explore whether perceived luckiness would relate to nonintentional psi.

Finally, we addressed one of Stanford's criticisms of earlier attempts to test PMIR by administering a post-session validation check questionnaire.

Hypothesis 1. There would be an overall positive nonintentional psi effect, as defined in the Design section below. Planned analysis: single-mean t-test vs. MCE; one-tailed.

Hypothesis 2. Nonintentional psi would be greater when helping a friend than when helping a stranger. Planned analysis: 2-way unrelated ANOVA.

Hypothesis 3. Nonintentional psi would be greater when the cost of helping was low than when the cost of helping was high. Planned analysis: 2-way unrelated ANOVA.

Hypothesis 4. Neuroticism would correlate negatively with nonintentional psi. Planned analysis: Spearman correlation; one-tailed.

Hypothesis 5. Openness would correlate positively with nonintentional psi. Planned analysis: Spearman correlation; one-tailed.

Exploratory Question. Perceived luckiness and nonintentional psi.



A between-subjects design was used. There were two independent variables: relationship between pairs (friends versus strangers) and cost of helping (high versus low), giving four conditions. The design was orthogonal in that there were 20 participants in each condition: i.e., 20 strangers in the high-cost condition, 20 low-cost strangers, 20 high-cost friends, and 20 low-cost friends, giving a total of 80 participants. The dependent variable (see below) was a measure of performance on a nonintentional psi task, where participants were required to rate Kanji (Japanese language) characters for their aesthetic appeal. The study was preplanned to terminate once 20 participants had been tested in each condition. Participants were allocated to the High-Cost or Low-Cost Condition according to a counterbalanced schedule prepared in advance of the study. Half the participants were allocated to the "low-cost" condition in which the participant received [pound]20 in cash at the end of the experimental session if they h elped. Half the participants were allocated to the "high cost" condition, receiving no reward if they helped. The order of testing friend pairs and stranger pairs was determined by when the pairs were available to take part. In practice, this led to a reasonably even mix of "friend" and "stranger" sessions throughout the study: the first ten sessions consisted of 7 friend and 3 stranger pairs; the next ten had 3 friend and 7 stranger pairs; the next ten had 6 friend and 4 stranger pairs; and the final ten had 4 friend and 6 stranger pairs.

The Dependent Variable: Measure of Nonintentional Psi. For each trial, a Kanji character was randomly selected (see Procedure section) as the "key" or target Kanji. The nonintentional psi measure was a standardized score calculated from the Kanji ratings for each participant's session (see Results section).

The "Helping" Contingency. A participant helped by performing on the nonintentional psi task in a way that resulted in the other participant experiencing the pleasant outcome. If a participant did not help, the other participant was directed to the unpleasant task. The requisite performance for "helping" was defined arbitrarily by an a priori rule: If the key Kanji was an odd number and the participant's highest-rated Kanji was also an odd number, this was defined as helping and the other participant was directed to the pleasant outcome; and if the key was an even number, and the highest-rated was an even number too, this was helping. Otherwise, the participant's performance on the nonintentional psi task had the consequence of "not helping," and the other participant was directed to the unpleasant outcome. Therefore, there was a 50% likelihood of helping by chance alone and 50% of helpers were in the low-cost condition; therefore, one in four participants (25%) had a chance of winning [pound]20.


Eighty volunteers participated in the study: 39 males, 41 females, mean age 26 years, range 15-51 years. Approximately half of the participants were students, about 10 were unemployed, and the remainder had a variety of occupations. Participants were recruited via email to student mailing lists and more general discussion groups, word of mouth, and advertising in a widely available magazine. The study was described as one into "personality and luck," and participants were correctly informed that there would be a 25% chance of gaining [pound]20 for participating. They were informed that the payment would be randomly determined, but not that the payment was part of the experimental manipulation. They were not informed that there was a parapsychological component to the study. Participants were asked either to bring a friend, or to come singly, knowing that another person coming singly would be scheduled at the same time as them. Strangers arrived separately at the prearranged time, were briefly introduced to o ne another by the experimenter, and were then taken to separate testing rooms. In no case did it turn out that strangers actually knew each other. There were three experimenters in this study: CW, JR, and VM. Two of these acted in each session, selected on an ad hoc basis. E1 (CW, JR, or VM) dealt directly with participants, while E2 (either CW or JR) recorded session information and directed E1 on procedural details during the session. CW and JR were aware of the nature and design of the study; VM was not aware of this.


NEO-FFI. The Neo-FFI (Costa & McCrae, 1991) is a 60-item shortform of the NEO-PI-R, a personality inventory that gives measures of neuroticism, extraversion, openness, agreeableness, and conscientiousness.

Perceptions of Luck Questionnaire. The Perceptions of Luck Questionnaire (Smith, Wiseman, & Harris, 1997) was used. The first question of this scale asks people to indicate their perceived luckiness, the second their perceived unluckiness. Further questions seek more information on what respondents think are the factors underlying luck, for example, random chance.

Nonintentional Psi Task. This was a rating task in which participants were asked to indicate their aesthetic preferences for each of ten Kanji characters. As in Stanford's word association task, these ten were preceded by three "dummy" characters, whose function was to familiarize participants with the task; scores on these were not included in the analysis. The ten Kanji characters were selected on the basis of a previous pilot study with the aim of identifying those characters to which most raters gave relatively neutral ratings (on a scale from 1 to 7, the chosen characters had mean ratings ranging from 3.9 to 4.2, where 4 = neutral), in order to avoid response bias.

Response booklets were prepared, each consisting of 13 pages with a character and rating scale on each page. The first three pages contained the dummy characters. The remaining pages contained the 10 experimental characters, in a different random order for each booklet. When doing the rating task, participants were allowed to look through the booklet as they wished and were not permitted to give any tied ratings. Participants were asked to rate the characters in terms of "aesthetic appeal" on a scale that ranged from 0 "Totally Displeasing" thru 50 "Neutral" to 99 "Totally Pleasing."

Post-Session Validation Cheek Questionnaire. In order to check the validity of our attempts to manipulate the relationship between participants, cost of helping, pleasantness and unpleasantness of tasks, and knowledge about the study, we asked participants to rate various parts of the experiment using seven-point scales. They were asked to rate a tracking task (see Procedure, below) and a tape or computer game (if they experienced it) for pleasantness, and a "leave early" option (if they took it) for how much they valued it. Further questions asked how well they knew the other participant, and how emotionally close they were. We also asked what they thought the experiment was about. Finally, participants were asked to rate how much they valued winning [pound]20 in the study (note that this judgment was made before they knew whether they would get [pound]20).


Designation and Security of the "Key" Kanji Characters. Prior to the start of the experiment, an assistant [2] otherwise uninvolved with the study produced a sequence of random numbers from 1 to 10. On a day known only to her, she derived an entry point into the Rand Corporation (1955) random number tables by applying an entry rule based on the temperature figures of the first six world cities, published in a daily newspaper. The randomizer kept a duplicate list of the selected random numbers in a locked filing cabinet in her office. The random numbers were placed in individual envelopes numbered from 1 to 100, kept in a locked box in the experimenters' office. [3]

Just prior to each testing session, E2 unlocked the target box and removed two numbered envelopes, each containing the designated target for each of the two experimental participants, and kept the envelopes on his or her person. E2 kept a separate written record of the target identity, once this was known. Following the experiment, CW double-checked that participants had been linked with the correct target identity, and that the subsequent manipulations (whether pleasant or unpleasant tasks were experienced, allocation to high cost or low cost of helping condition, payment of reward) were correct. In one case, it was found that the wrong target identity had been linked to the participant, resulting in the session being conducted as if they had "not helped" when in fact they had "helped." The data from this session were not included in the study, and an additional session was run.

Session Procedure. Three nonadjacent rooms in the basement of the Department of Psychology were used for testing. One was occupied by E2, who had no contact with participants during the session. El allocated each of the participants to the other two rooms, according to alphabetical order of their surnames. The two rooms each contained a Macintosh computer, a separate desk and chairs, an easy chair, a tape recorder, some reading material, and the session materials.

E1 alternated between the two participants' rooms, taking them individually through the procedure. First, the contents of the session were described (without revealing the nature of the rating task, nor the contingency between the rating task and the other participant's subsequent task), and the participant's questions were answered as far as possible. The participant then signed a consent form and was directed to complete the rating task. El inspected the rating booklet for tied ratings, and asked the participant to rerate to break any ties. (Note that at the time of rating, El was blind as to the key Kanji identity and to the "cost of helping" condition allocation). El then identified which character had received the highest rating and communicated this to E2. While the participant had been doing the rating task, E2 had opened the two envelopes containing the target designation for each of the participants. When contacted by El, E2 told El which Kanji character (the key character) each participant was to p ractice drawing, as a decoy task. While the participants were practicing drawing their key Kanji, E2 recorded whether participants' highest rated character matched their key Kanji and, therefore, whether they had "helped" or "not helped" each other. E2 also noted what condition the participants had entered (low cost of helping or high cost of helping).

E2 instructed El on what was to follow for each participant. If a participant "helped," the other member of the participant pair was directed to do a pleasant task. This was a choice of three options; the ability to choose was itself regarded as a positive feature of the pleasant task. The choice was either to leave the experiment early, to listen to a relaxing audio tape over headphones in a comfortable chair, or to play an "asteroids" style computer game. These options were intended to be relatively pleasant, even for participants with quite varying individual preferences. If a participant did not "help," the other member of the participant pair was directed to do a tedious and rather tiring computer tracking task, using the computer mouse to track a slowly moving target on the screen. If the mouse drifted off target, an auditory tone sounded. The task was also unpleasant in that there was no choice (unless the participant opted to abort the session altogether), and the participant was not informed how lon g the task would take (15 minutes).

After finishing their pleasant or unpleasant tasks, participants completed the post-session validation check questionnaire and received payments, if necessary. They were asked not to discuss the study with other potential participants. Participants were not debriefed as to the nature of the study at this time, because of the need to ensure security of the nonintentional psi task. However, following completion of the study, all participants were sent a three-page feedback letter which explained fully the nature of the study and invited participants to contact GW or JR if they wished any further information about the study.


Validation Check Questionnaire

Before considering the principal hypotheses, we must establish whether our experimental manipulations had the desired effect.

Pleasant and Unpleasant Tasks. Forty participants did the unpleasant tracking task. On a scale of 0 (not at all pleasant) to 6 (very pleasant), the mean rating was 1.325 (SD = 1.575), suggesting that participants did indeed find this an unpleasant task. Forty participants had the three pleasant options. Eight participants chose to leave early. Asked to rate the value to them of leaving on a scale from 0 (very little) to 6 (a lot), the mean rating was 4 (SD = 1.512), suggesting that this was a moderately pleasant option. Eight participants chose to listen to music. On a scale from 0 (not at all pleasant) to 6 (very pleasant), the mean rating was 4.6, (SD = .916) suggesting that participants found the music quite pleasant. Twenty-four participants chose to play the computer game. On a scale of 0 (not at all pleasant) to 6 (very pleasant), the mean rating was 4.6 (SD = 1.139), suggesting that the game was quite pleasant to participants. A Mann-Whitney comparison of the 40 pleasant task ratings versus the 40 unp leasant task ratings gives a highly significant difference, z score (based on U) = 6.586, thus validating the intended effects of the pleasant and unpleasant tasks.

Cost of Helping. The potential payment of [pound]20 was intended to be of value to participants. Prior to knowing whether they would receive [pound]20, all participants rated how they valued [pound]20 on a scale from 0 (very little) to 6 (a lot). The mean rating was 4.2 (SD = 1.679), indicating that [pound]20 was quite valuable to participants. However, several participants reported that they found participating in the experiment intrinsically rewarding and that the potential financial gain was of only secondary importance.

Participants' Relationship. Asked to rate how well they knew each other, on a scale from 0 (not at all well) to 6 (very well), the "stranger" participants gave a mean rating of 0.1 (SD = .656), and the "friend" participants gave a mean rating of 5.5 (SD = .951). Asked to rate how emotionally close they were, on a scale from 0 (not at all close) to 6 (very close), strangers gave a mean rating of 0.2 (SD = .721), and friends gave a mean rating of 4.9 (SD = 1.413). These ratings clearly indicate the validity of our identification of friends and strangers.

Nature of the Study. Asked to indicate what they thought the study was about, only one participant gave any indication that they suspected this was a parapsychological study. That participant had previously taken part in parapsychological research involving CW. No participant guessed the nature of the link between the participant pairs; indeed, several queried why it was necessary to have two participants at once because what they did seemed to be independent. We are therefore confident that the psi task, and the helping contingency, remained nonintentional.


Participants' preference ratings for the 10 Kanji characters were normally distributed. For each participant's session, a standardized rating score was calculated from the difference between the participants' rating on the key Kanji and their mean rating for all 10 Kanji, divided by the SD of all 10 ratings. Positive scores indicate higher than average preference ratings for the randomly selected key Kanji, that is, positive nonintentional psi performance. Negative scores indicate lower than average preference ratings for the key character, and a score of 0 is equivalent to exactly chance performance. Apart from this primary outcome measure, we also give descriptive results for a "direct hit" measure (where participants' highest rated Kanji exactly matched the hidden target Kanji), to allow comparison with Stanford and Rust's results.

Hypothesis 1. For all 80 trials, the mean standardized rating score was 0.117 (SD = .941), t(79) [vs. MCE] = 1.117, p = .13, one-tailed, effect size (Cohen's d) = .25. This gives no statistically significant support to Hypothesis 1, that nonintentional psi would occur in the rating task. For the secondary, direct hits measure (with 80 trials and a 10% probability of a hit), we found 10 direct hits. This is positive but nonsignificant, N = 80, exact binomial p = .276, effect size (Cohen's h) = .08.

Hypothesis 2. It was predicted that friends would have higher rating scores for the key Kanji than strangers. Table 1 gives mean standardized rating scores for the four testing conditions, and shows little difference between friends and strangers on this measure. For the secondary direct hits measure, friends had three hits, N = 40, exact binomial p = .78, ES (h) = -.09; strangers had seven hits, N = 40, exact binomial p = .1, ES (h) = .22.

Hypothesis 3. It was predicted that individuals in the low cost of helping condition would have higher rating scores than those in the high cost of helping condition. Table 1 shows little difference in the results for the two conditions. For the secondary direct hits measure, there were seven hits in the high cost condition, N = 40, exact binomial p = .1, ES (h) = .22; 3 hits in the low cost condition, N = 40, exact binomial p = .78, ES (h) = -.09.

As planned, Hypotheses 2 and 3 were tested using a two-way unrelated Analysis of Variance of standardized rating scores, in order to detect possible interaction effects between the cost and relationship conditions. The results shown in Table 2 confirm that no significant effects were found for the two principal manipulations in this study. For the direct hits measure, six hits were scored in the high-cost stranger condition, two in the low-cost friend condition, and one each in the low-cost stranger and high-cost friend conditions. This distribution does not significantly differ from chance, Fisher's exact p = .825.

These results give little evidence that participants nonintentionally used their psi to help each other (Hypothesis 1), and give no support to the detailed predictions made by the PMIR theory (Hypotheses 2 and 3).

Hypothesis 4. It was predicted that scores on the personality factor neuroticism would be negatively related to nonintentional psi. Mean neuroticism score = 21.9 (SD = 9.190). Correlating the standardized rating scores with neuroticism gave [r.sub.s] (79) = -.183, a weak effect in the predicted direction, just missing significance at the .05 level one-tailed.

Hypothesis 5. It was predicted that scores on the personality factor openness would be positively related to nonintentional psi. Mean openness score = 34.5 (SD = 5.071). No relationship was found between Openness and standardized rating scores, [r.sub.s] (79) = -.006.

Exploratory Question: Luckiness and nonintentional psi. For each participant, a "net luck" score was calculated by subtracting the participants' self-rated "unluckiness" from their self-rated "luckiness." Mean netluck score was 1.438 (theoretical and actual range from -6 to +6; SD = 2.480). No correlation was found between netluck and performance on the nonintentional psi task, [r.sub.s] (79) = .046.


This study attempted to conceptually replicate and extend Stanford and Rust's (1977) successful study showing that participants could nonintentionally use their psi to help another person avoid an unpleasant task and instead experience a pleasant task. The underlying PMIR model was tested by varying the cost to the participant of helping another and the relationship between the participants. The model also predicted relationships between nonintentional psi, neuroticism, and openness. Although results were in the predicted direction for the overall measure of nonintentional psi and for neuroticism, none of the predictions received statistically significant support.

Therefore, in terms of finding statistically significant effects, our results did not support the PMIR model and did not replicate Stanford & Rust's (1977) helping study for the secondary discrete (direct hits) outcome indicator. However, it would seem to be worthwhile to look again at these questions, using greater statistical power. The validation check questionnaire confirmed that the experimental manipulations had the desired effect, suggesting this was a valid test of the theory. However, it is worth noting that Stanford and Rust's study found a small effect size (Cohen's h) of .36, based on their discrete outcome indicator. Our study also found a small effect size (Cohen's d = .25 based on the continuous outcome indicator Cohen's h = .08 for the discrete outcome indicator; .20 is a small effect, .50 a medium effect, and .80 a large effect; Rosenthal & Rosnow, 1991), for the basic question of whether scores on the nonintentional psi task were above chance. Based on published power tables (Cohen, 1988, p . 31) our study, with 80 participants overall, only had a statistical power of around .48 in order to detect a small nonintentional psi effect at the .05 level one-tailed (that is, less than a 50% chance of detecting any effect). We would have needed to have over 200 participants to raise power to approximately .75 (Cohen, 1988, p. 31). In addition, our nonintentional psi task was newly devised, and this highlights the general problem that very little research has been done on the question of nonintentional psi. If parapsychologists are to make progress in devising theoretical models of psi, much more systematic research needs to be done into the underlying assumptions of such models.

Smith, Wiseman, and Harris (1997) argue that luck may be a form of PMIR and recommend comparing luckiness and performance on a nonintentional psi task. We found no relationship between luckiness and nonintentional psi. However, given that the psi task in this study was to help another, rather than to use one's psi to benefit oneself, one could argue that the psi task was not analogous to the experience of luck in everyday life. A subsequent study has compared luckiness with performance on a PMIR task in which participants could use their psi to benefit themselves, with encouraging results (Watt & Nagtegaal, 2000). The study of luckiness and nonintentional psi may provide fertile ground for a fruitful interaction between laboratory parapsychology and the study of everyday psi experiences.

(1.) We gratefully acknowledge financial support from the Perrott-Warrick Fund of Cambridge University and the Moray Fund of the University of Edinburgh. This paper has benefited from many helpful suggestions from Julie Milton and Fiona Steinkamp, and anonymous referees, for which we are grateful. We also thank Rex Stanford for his encouragement, assistant experimenter Valerie MacGregor for her invaluable help, and Nancy Zingrone and Carlos Alvarado for statistical help.

(2.) We would like to thank Claire Brady for her help in this role.

(3.) Shortly before the start of the study, entry was forced into over 60 rooms in the Department of Psychology, including the experimenters' office. Cash and portable electronic equipment were stolen. The box containing the target designations was broken into. We found no sign that the target envelopes had been opened. Therefore, we did not re-randomize the targets, but re-housed them in a new locked box.


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

COHEN,J. (1988). Statistical power analysis for the behavioral sciences. (2nd Ed.). Hillsdale, NJ: Lawrence Erlbaum Associates.

COSTA, P. T., & MCCRAE, R. R. (1991). Neo Five-Factor Inventory. Odessa, FL: Psychological Assessment Resources, Inc.

CRAWFORD, C. B., & KREBS, D. L. (EDS.), (1998). Handbook of evolutionary psychology: Ideas, issues and applications. Mahwab, NJ: Lawrence Erlbaum Associates.

PLOTKIN, H. (1994). Darwin machines and the nature of knowledge. London: Penguin.

RAND CORPORATION (1955). A million random digits with 100,000 normal deviates. London: Collier-MacMillan.

ROSENTHAL, R., & ROSNOW, R. L. (1991). Essentials of behavioral research: Methods and data analysis. (2nd Ed.). New York: McGraw-Hill.

SMITH, M. D., WISEMAN, R., & HARRIS, P. (1997). Perceived luckiness and the UK national lottery. Proceedings of Presented Papers: The Parapsychological Association 40th Annual Convention, 387-398.

STANFORD, R. G. (1974a). An experimentally testable model for spontaneous psi events. I. Extrasensory events. Journal of the American Society for Psychical Research, 68, 34-57.

STANFORD, R. G. (1974b). An experimentally testable model for spontaneous psi events. II. Psychokinetic events. Journal of the American Society for Psychical Research, 68, 321-356.

STANFORD, R. G. (1977). Conceptual frameworks of contemporary psi research. In B.B. Wolman (Ed.), Handbook of parapsychology (pp. 823-858). New York: Van Nostrand Reinhold.

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, Vol. 6 (pp. 54-167). Jefferson, NC: McFarland.

STANFORD, R. G., & ASSOCIATES (1976). A study of motivational arousal and self-concept in psi-mediated instrumental response. Journal of the American Society for Psychical Research, 70, 167-178.

STANFORD, R. G., & RUST, P. (1977). Psi-mediated helping behavior: Experimental paradigm and initial results [Abstract]. In J. D. Morris, W. G. Roll, & R. L. Morris (Eds.), Research in parapsychology 1976 (pp. 109-110). Metuchen, NJ: Scarecrow Press.

WATT, C., & NAGTEGAAL, M. (2000). Luck in action? Belief in good luck, psi-mediated instrumental response, and games of chance. Journal of Parapsychology, 64, 33-52.
COPYRIGHT 2000 Parapsychology Press
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2000 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Psi-Mediated Instrumental Response
Publication:The Journal of Parapsychology
Geographic Code:1USA
Date:Mar 1, 2000

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters