DOC PREVIEW
Replication and Interpretation

This preview shows page 1-2-3-27-28-29 out of 29 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 29 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 29 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 29 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 29 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 29 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 29 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 29 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

The MegaREG Experiment:Replication and Interpreta tionY. H. DOBYNS,B.J.DUNNE,R.G.JAHN, AND R. D. NELSONPrinceton Engineering Anomalies ResearchSchool of Engineering and Applied SciencePrinceton UniversityPrinceton, NJ 08544-5263Abstract—Anomalous effects of human intention on the output of electronicrandom event generators (REGs) have been well established at the PEARlaboratory and elsewhere. A simple model of this effect as a change in thebinary probability of the REG digits would predict that larger statistical yieldcan be achieved simply by processing more bits. This hypothesis was exploredpreviously using protocols ranging from 20 to 2000 bits per trial, with resultsthat were consistent with the bitwise model, but had too little resolution to ruleout many competing models. More recently, a ‘‘MegaREG’’ experiment wasdeployed to test this hypothesis using 2-million-bit trials interspersed with 200-bit trials in a double-blind protocol.In the initial phase of MegaREG, the 200-bit trials produced outcomescomparable with our standard experiments, while the 2-million-bit trialsproduced an effect somewhat larger in absolute scale, but inverted with regardto intention. A subsequent replication phase reproduced these findings, exceptfor statistical ly nonsignificant quantitative changes. These appear to besecondary consequences of a statistically significant di fference be tweenoperators having, and lacking, prior experience in REG experiments, therelative proportions of which account for the differences between theseexperimental phases. Other operator population distinctions, such as gender,and various secondary protocol parameters, had no significant effects.A related experiment called ‘‘MegaMega,’’ differing from MegaREG only inthat all data used 2 million bits per trial, with no interspersal of a second datatype, produced a reversed intentional effect of the same scale. It also displayeda significant asymmetry between the intentional runs and the non-intentionalbaselines, which was not seen in MegaREG.The combined result of all high-speed experiments was an effect size per trialof 2.77 6 0.69 times that seen in earlier REG experiments, but given thelarger number of bits per trial, the bitwise probability change was some 30times smaller. The composite score for the intentional effect in high-densitydata across all experiments was T ¼4.03 (d.f. . 105), p ¼ 5.65 3 105(2-tailed). The causes of the change of scale, and of the inversion of sign in theeffect, remain unknown. Explanations that can be ruled out with a high degreeof confidence include statistical artifact, the change in the source, the use ofdifferent operator pools, and the double-blind interspersal of data types.Testable explanations that remain potentially viable include increased taskcomplexity, inherent timing or rate limits on anomalous functioning, andchanges in the psychological environment.Journal of Scientific Exploration, Vol. 18, No. 3, pp. 369–397, 2004 0892-3310/04369Keywords: consciousness—anomalies—human/machine—interaction—ran-dom event—generators—replicabilityIntroductionThe Princeton Engineering Anomalies Research (PEAR) laboratory has con-ducted extensive experimental and theoretical study of anomalous eff ects ofhuman consciousness on various types of random event generators (REGs) since1979 (Jahn et al., 1997). This work builds on previous and ongoing studies bymany other researchers, and is particularly close in design and protocol to that ofSchmidt (1970a,b). An extensive meta-analysis by Radin and Nelson (1989)found that most of these experimental programs, including PEAR’s, producedanomalous effects of broadly similar nature and scale.The anomalous effect seen in these experiments consists of a shift in the meanoutput level of REGs that correlated with pre-stated human intention. Thus,regardless of mechanism or model, any database containing an anomaly displaysa change in the empirical probability of the individual binary events comprisingthat database. Most of the PEAR data are consistent with the hypothesis that theanomalous effect is in fact nothing more nor less than an alteration of theprobability of elementary binary events in the experiment, rather than some morecomplicated process which would produce the empirical probability shift asa consequence (Dobyns, 2000; Jahn et al., 1991). For reasons of protocolstandardization, however, many physical and psychological variables were heldconstant in these experiments, which means that the consistency with theprobability-change hypothesis could be an artifact arising from the uniformity ofsome other key parameter throughout the experiments. If the anomalous effect istruly a shift in the elementary bit-level probability, however, the statistical yieldof an experiment should be increased by increasing the number of bits processed,while holding all other factors constant.This prospect was explored in a preliminary fashion using various protocols.The standar d ‘‘REG200’’ protocol collects the sum of 200 random bits intoa single ‘‘trial,’’ at a sampling rate of 1000 bits/second. Thus, data collection isactive for 0.2 second, a period of time easily perceptible to the operator. Anintervening pause of approximately 0.7 second leads to a mean data generationrate of approximately 0.9 second/trial. Two exploratory variants, labeled REG20and REG2000, collected respectively 1/10 and 10 times as many bits per trialfrom the same noise source, with the sampling rate changed correspondingly sothat the periods of trial accumulation were the same. Although the amount ofdata accumulated in these protocols was small relative to the primary REG200experiment, the results seemed compatible with a bitwise effect hypothesis(Dobyns, 2000; Jahn et al., 1997). These previous explorations are summarizedin Figure 1, showing all three data points well within a 95% confidence interval370 Y. H. Dobyns et al.(1.96 times as wide as the 1r error bars drawn) of the sloping dotted lineindicating constant effect per bit.The original ‘‘MegaREG’’ experiment (Ibison, 1997) attempted a morethorough test of the bitwise effect hypothesis. The concept was to increase thestandard trial size from a sum of 200 bits to a sum of 2,000,000 bits. If the effectwas a direct alteration of binary probability, then increasing the number of bits perexperiment by 104, while keeping all other parameters of the experiment constant,should


Replication and Interpretation

Download Replication and Interpretation
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Replication and Interpretation and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Replication and Interpretation 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?