Study Design

This study used a randomized pre-test-posttest time-series design that compared the performance of three groups of participants across three time periods on five outcome measures related to receiving unexpected money, the primary dependent variable (DV).

Unexpected money was defined for participants as, “Money that comes into your hands surprisingly and suddenly (‘out of the blue’), without your earning it, soliciting it, specifically seeking it out, or otherwise expecting it.” Examples of unexpected-money events included finding cash on the street, inheriting money unexpectedly, winning money in a lottery or contest, receiving money through an unexpected credit or class action settlement. Even windfall capital gains, wage increases, new work, or return of monies borrowed were considered valid events if there was some aspect to their receipt that felt highly out of the ordinary.

Events that were not considered unexpected money included gains from illegal activities, and awards or credit from litigation, insurance settlements, or other disputes in which the participant was an active plaintiff, defendant, or victim.

Immediately after enrollment, eligible participants were randomly assigned to equally sized directed, individual, and control groups and asked to complete a baseline questionnaire that, among other measures, asked about their experiences with unexpected money during the previous month.

During the first 2 weeks of the study (period 1), all participants were asked to pay attention to receipts of unexpected money.

During the second 2 weeks of the study (period 2), the directed and individual groups were asked to perform separate practices to help focus their intention on receiving unexpected money. In contrast, the control group was instructed to continue paying attention. All three groups reported results on a final questionnaire.

Study Hypotheses

The research question posed in SUM was: Will participants receive unexpected money or other related financial events more regularly when they are asked to focus more intention on such events than they did before? If so, which activities or characteristics of the participants — if any, besides the fact of being involved in an experiment — are significantly associated with the occurrence of money events?

The null hypothesis was that there would be no change in receipts of unexpected money across groups and times, demonstrating no association between the application of intentionality and the outcome of outer events. The alternative (two-tailed) hypothesis was that there would be statistically significant changes in either direction across groups and times, demonstrating an association between applying intention and experiencing the target events. Specifically, participants performing intentional practices (ie, those assigned to the directed and individual groups) would experience a significantly greater number of money events than the control group (which was instructed only to “pay attention”). They would also experience significant improvements in their own performance during Period 2 when they were performing the designated intentional practices. The directed group, which was presented with a candle unexpectedly and asked to perform a ceremony, would prevail over the individual group, which was asked to perform personally meaningful practices. This was hypothesized because the ceremony they had performed has long held numinous associations. However, if members of the individual group were already experienced practitioners of spiritual or intentional activities, then they would attain the highest numbers.

The control group participants would experience no significant differences in their performances over the three measuring periods. The directed group would prevail over the individual group (that was told to perform personally meaningful practices) because they were unexpectedly presented with a candle and asked to perform a ceremony that has long held numinous associations for people, but members of the individual group, who were already experienced practitioners of spiritual or intentional activities, would attain the highest numbers. The control group participants would experience no significant differences in their performances during the three measuring periods.

A significance level of P < .05 was chosen to reject the null hypothesis for all outcome measures.

Study Method

Results

Kolmogorov-Smirnov Normality Tests were first performed to ascertain whether a normal distribution of differences on each variable existed in the study participants. Distributions on all variables, including the dependent variables, fell into a normal (P > .05) bell-curve distribution, with the exception of two wagering questions (“Have you ever spent money on wagering?” and “In the past year, have you spent money on wagering?”) in which responses were significantly skewed toward the abstention end of the scale (for historical: x2 [2, n = 60] = 10.800, P = .009; for present: x2[2, n = 60] = 8.533, P = .028). This uneven distribution, however, proved to be evenly distributed across the groups when one-way analysis of variance (ANOVA) tests were run (historical: F[2,57] = 0.373, P = .690; present: F[2,57] = 0.095, P = .909).

To determine whether the randomization process was successful in producing a normal demographic distribution across groups, means between groups on continuous variables were compared using one-way ANOVA tests. The tests revealed uneven distributions on only two variables: age and the degree of perceived effectiveness of personal prayer practices.

The nonnormality in age distribution was most likely caused by outliers in both the individual and control groups. Once this variable was recategorized into six age groups of a decade each and “elderly” (age 70 years and older), chi-square analyses revealed x2(10, n = 60) = 16.834, P = .078, which means that, statistically speaking, age categories were normally distributed across the groups. The prayer efficacy variable was similarly recategorized into groups of “low” (those who rated the efficacy of employing “indirect methods” between 1 and 3), “middle” (4 to 7) and “high” (8 to 10), yielding x2(4, n = 51) = 6.510, P = .160, meaning that the groups were more or less evenly distributed with regard to categories of response.

Chi-square analyses were also conducted on the categorical measures (eg, religion and occupation) to determine whether traits were distributed evenly among the groups. They yielded P values more than .05 on all traits, indicating the randomizing process had produced statistically equivalent groups on these traits.

Correlations were then conducted on all independent and dependent continuous variables. The primary discovery from this procedure were the highly positive — and significant — correlations between the baseline expectations outcome measure and the incidences outcome measure for all three time periods (in the first, r = 0.485, P < .0001; in the second, r = 0.644, P < .0001; in the third, r = 0.487, P <.0001). Other less pronounced — but still positive and significant — correlations were found between the unexpected money outcome measures (incidences and quantity) and high prayer efficacy, frequent prayer practice, and high financial and wagering luckiness.

Regarding the initial expectations measure, another significant observation was made: judging by their erratic responses to this measure at the end of periods 1 and 2, participants appeared to be misinterpreting the question (“To what extent were your expectations met?”) because they were not recalling their responses on the first questionnaire. Instead, many thought the questionnaire was asking about their success in receiving unexpected money during that particular period, and so responses on this variable did not follow expected patterns. For example, some who had scored expectations low on the baseline scored it high in period 2 because they had received some unexpected money, when a low-score response was again anticipated because the initial “low” expectations had not been met. As a result, period 1 and period 2 expectations responses were discarded from the analysis.

Main Analyses

Before performing the main analysis, one-way ANOVAs were conducted on all outcome measures by group for every time period to spot trends in the data. Significant differences between the groups were spotted on the first outcome measure (incidences) for baseline and period 2. Differences between the groups also approached significance in period 1 (F[2,57] = 2.765, P = .071). Tukey’s significant difference post-hoc comparisons revealed that the significant differences occurred in differing pairs of groups in the two time periods where significance was attained: At baseline, the directed group reported significantly more incidences than the individual group (P = .041), whereas in period 2, the control group had significantly more incidences than the individual group (P = .045).

An ANOVA with repeated measures statistical procedure to compare participants’ success on the four main measures (incidences, quantity, events, and event impact) across the three time periods and three groups yielded a significant result for the incidences outcome, nonsignificant for the other three. The results of this statistically significant test are unequivocal regarding the association of group membership with outcome because neither the effect of time nor the interaction between group membership and time were significant.

The Games/Howell post-hoc comparison test revealed that both the control and directed groups significantly outperformed the individual group. The control group participants actually received unexpected money more times overall than the directed participants, but this difference was not significant. With the significant difference discovered between the individual and directed/control groups, the incidences variable became the outcome measure of interest, and the focus of all subsequent tests that analyzed the influence of other variables that correlated highly with this outcome measure or with each other.

As the correlations predicted, by far the most significant results emerged from the analysis comparing initial expectations scores to the outcome measure. Re-categorizing participant responses into “low expectations” (those who responded between 1 and 4 on the Likert scale); “middle expectations” (5 to 7); and “great expectations” (8 to 10); and comparing these attitudinal groups’ performance on the incidences variable across the time periods, this test revealed substantial significant differences.

The Games/Howell post-hoc comparison test revealed that the significant differences occurred between the low and great, and middle and great groups. On average, those reporting the highest expectations received unexpected money a significantly greater number of times. On average, participants in the low group reported receiving unexpected money less than twice during the baseline and period 2, and 2.5 times in period 1; those in the middle group received money between two to three times during each period; and those in the great group received money more than three times. As in the main analysis of the groups, the time effect and interaction between expectations and time both proved insignificant, further validating the association between expectations and the incidences variable.

Although such numbers suggest that initial expectations were more important than group membership in predicting outcome on the incidences dependent variable, the fact that participants were statistically evenly distributed across groups with respect to expectations (F[2,57] = 1.351, P = .267) indicates that the effect of the group must also be taken into consideration when evaluating overall results.

Discussion

The significant findings for the main effect of group on the incidences variable, as well as the supplemental analysis of the effect of expectations on the incidences variable, reveal a new pattern of events that emerged during the study. According to the probability rules, which provide theoretical and statistical bases for hypothesis testing, equal groups should yield equal results, unless another factor is influencing the situation. With these exceptions noted, the experimental groups were equal regarding demographics, attitudes, wagering and prayer practices, and initial expectations. Consequently, the nonnormal distribution of receipts of unexpected money across groups and expectation levels is notable and, as such, calls for a closer investigation.

Although the statistically significant results support the alternative hypothesis, they nevertheless confounded research expectations that sought to affirm the effectiveness of applying conscious intentionality, owing to the lower than expected performance of the individual group. That success in this study is associated with factors other than conscious intentionality, indicated by the high significance of the effect of expectations, as well as results in the group membership test. At baseline, the directed participants were the most successful, reportedly receiving money an average of slightly more than two times during the previous month, whereas the controls received money fewer than two times, and the individuals only one time. Yet, by the end of the study, the controls had caught up and outperformed both groups on average, and the individuals, by a statistically significant amount.

The fact that the controls prevailed seems to indicate that the act of paying attention was enough, and that performing any additional intentional activities actually deterred participants from achieving the desired outcome. These findings support those from prayer studies ⁸ and PK experiments, ⁹ which have discovered that the best results often are achieved when participants are not applying rigorous intentionality to the psi task. Dossey, in fact, devoted much attention to this question and cited the Spindrift studies, that compared the efficacy of directed and nondirected prayer for healing of living organisms. ⁸ Although they found both methods to be effective, nondirected prayer (ie, not focused on a particular outcome) reaped the most impressive results. Healing was found to be most pronounced when the prayers were free of visualizations, associations, or specific goals. The high significance of the initial expectations variable’s association with success on the incidences variable corroborates the results of other research, ⁷ and aligns with Bandura’s well-known theory of self-efficacy, which posits that people with “efficacy expectations” believe they “can successfully execute the behavior required to produce the outcomes.” ¹⁰ In this study, participants’ ability to apply intentionality toward increasing the incidence of money events appeared to increase with higher levels of self-efficacy.

However, it must still be noted that the directed group also significantly outperformed the individual group, indicating that certain types of intention can be more effective than others. These invite closer study, even though the mean success of the directed group on the incidences variable actually declined — albeit nonsignificantly — after they were introduced to the candle activity (from a mean of 2.90 events in period 1 to 2.70 in period 2). The means of both the directed and individual groups on this measure were actually highest in period 1.

Results reveal that intentional practices were only partially associated with success on the incidences variable. However, they also reveal that participants had more incidences while they were engaged in the study than they had during the baseline period. The argument that people were just noticing more pennies on the ground after they entered the study is easily refuted by the fact that not all groups — equally matched in attributes, attitudes, and practices — experienced the same pattern of increase in the incidences variable.

Synchronicity and the Study

Every time a participant received unexpected money when enrolled in a study of such events represented an instance in which the outer world seemed to collude with an inner thought, and would probably be considered a synchronicity by Jung, ⁹ though perhaps not Mansfield ¹² because it did not aid the individuation process. If incidences of receiving unexpected money can indeed be considered examples of AMMCs, these statistical results seem to contradict Jungian beliefs that these events cannot be measured statistically.

Although some participants experienced no events in the study, the statistically significant results on the incidences variable and association of the baseline expectations variable with the incidences variable reveal that a new pattern of events emerged for many participants. Many participants received unexpected money more times while engaged in the study than they had the month immediately before enrollment. Follow-up interviews revealed that some who did not receive unexpected money experienced other financial events. Although part of this change could no doubt be attributed to fallacy of memory and heightened sensitivity, the fact that results were significantly uneven across the groups indicate that paying active attention rather than engaging in an effortful intentional practice was more frequently associated with the occurrence of AMMCs.

However, the study affirmed the acausal, random character of AMMCs postulated by Jungians because the general pattern of findings did not apply in every individual case, and they did not manifest in the same form. For instance, some participants who scored high on baseline expectations did not receive unexpected money; others who scored quite low did.

Also, in accordance with Jung’s beliefs, all groups averaged more success during period 1 than they did during period 2, perhaps because the experience of being in the study was still novel. This supports widespread findings in psi research discussed previously ² that participants score the most hits at the beginning of trial periods, and then suffer “decline effects” as the trial proceeds.

The fact that this success occurred before the experimental groups were introduced to their intentional activities indicates that the exercise of active intention was not associated consistently with success in this study. All groups had more incidences of unexpected money during the two study periods than they had the previous month (the baseline). The individual and control groups actually had more incidences during period 1 than in the previous month. The directed group, however, had more incidences during baseline than during period 1.

Even though this study did not hypothesize that any particular mechanism would cause the money events to occur, it did hypothesize that an association would be observed between the application of intentionality and the occurrence of money events.

The new finding here is not that the exercise of intentionality was associated with outer-world events, but that it was achieved in the context of a consciously wrought experiment in a pattern that was statistically discernible.

Further Research

This study, as the first of its kind this researcher has encountered in the literature, must be repeated for its findings to be meaningful to the growing body of research in coincidence studies.

Jungians maintain, however, that, just as each moment in time carries its unique archetypal quality, an exact replication of this study — or of any study, as a matter of fact — could be impossible. The next SUM will perhaps be performed by a different experimenter using different people, a slightly modified study design, and could take place in a different year characterized by its own unique economic, social, political, cultural, and astrological features.

If correspondences are still discovered, however, especially if they resemble the patterns found in this study, the psi field will be closer to establishing an empirical basis for the Jungian assertion that an intrinsic relationship exists between inner states and outer events and that the practice of attention/intention is mediated by the archetype.