The Global Consciousness Project: 28 Years of Worldwide RNGs

Roger Nelson was a senior researcher at Princeton's PEAR lab when he proposed an ambitious follow-up: instead of one person trying to influence one Random Event Generator in a lab, what if you placed 70+ REGs around the world and looked for output deviations during moments of intense, coherent global attention?

He launched the Global Consciousness Project (GCP) in 1998. It is still running in 2026. The cumulative result: a z-score of 7.31, equivalent to p < 10⁻¹².

This article walks through the protocol, the most-cited events, and the selection-bias critique that defines the field's response.

What does the GCP measure?

The GCP operates 70+ Random Event Generators (REGs) at fixed locations across more than 40 countries. Each REG produces a continuous stream of binary digits driven by quantum-mechanical noise. The data streams are sent to a central server at Princeton and aggregated in real time.

The hypothesis is that during global events that produce intense, coherent collective attention, the network's combined output should deviate from chance in detectable ways.

Importantly, the analysis is pre-registered: events are designated as "formal" tests in advance, with start and end times set before the data is analyzed. This pre-registration began in 1998 and is publicly logged.

What is the cumulative result?

Across 500+ pre-registered global events, the cumulative z-score is 7.31. In statistical terms, this corresponds to a probability of approximately 10⁻¹² that the observed pattern arose from chance alone — one in a trillion.

[Nelson R, Bancel P. (2011). Effects of Mass Consciousness: Changes in Random Data During Global Events. Journal of Scientific Exploration, 25(3), 327-350.]

What were the most-cited events?

September 11, 2001 (the World Trade Center attacks): Among the largest single-event deviations on record. The cumulative deviation across the network rose sharply in the hours surrounding the attacks and remained elevated for approximately 48 hours.

New Year's transition events: Each year's midnight crossing across time zones produces a measurable cumulative signature that has been reported across more than 20 New Year cycles.

Princess Diana's funeral (1997): A pre-test of what became the formal protocol; produced significant deviation.

Major sporting events, election nights, natural disasters, religious observances: Mixed results, with significant effects in some categories and not others.

[Nelson R. (2002). Coherent Consciousness and Reduced Randomness: Correlations on September 11, 2001. Journal of Scientific Exploration, 16(4), 549-570.]

What is the selection-bias critique?

The principal scientific critique of GCP is that the cumulative significance reflects post-hoc event selection rather than a genuine effect.

The argument: across continuously running RNG data, you will always find anomalous deviations somewhere. If you can pick which time windows to analyze after the fact, you can guarantee statistical significance. The GCP's response is that events have been pre-registered with start/end timestamps before analysis since 1998, and that the cumulative significance survives even when the analysis is restricted to formally pre-registered events.

Skeptics counter that the selection of which events to designate as formal still leaves room for unconscious bias. The argument is that researchers might (consciously or not) designate as "formal" only those events that look likely to produce a deviation.

Defenders counter that the event-selection criterion has been published since 1998 and predates most of the events tested.

This back-and-forth — pre-registration vs. selection bias — is the unresolved methodological frontier of the GCP literature.

Can we say the GCP proved collective consciousness affects RNGs?

No. The GCP provides a cumulative anomaly that survives basic methodological audit. The interpretation of that anomaly is the open scientific question.

What we can say:

• A statistically significant deviation correlates with pre-registered global events
• The deviation has been reported across 28 years of continuous data collection
• Standard methodological audits identify selection concerns but do not eliminate the cumulative significance

What we cannot say:

• That the mechanism is "collective consciousness" or any other specific cause
• That the result rules out subtle confounders (e.g., infrastructure load on RNGs during major events)
• That the effect would replicate under stricter blinded protocols

Where does this fit in the pillar?

The GCP extends the PEAR REG approach to a worldwide network. It does not prove the same thing as PEAR — but it converges on the same uncomfortable observation: focused human attention correlates with measurable RNG output deviations across multiple independent protocols.

For the full evidence base, see the Psi Research Evidence pillar.

Sources

• Nelson R, Bancel P. (2011). Effects of Mass Consciousness. Journal of Scientific Exploration
• Nelson R. (2002). Coherent Consciousness and Reduced Randomness: 9/11. JSE
• Global Consciousness Project results page
• Bancel PA. (2017). Searching for Global Consciousness: A 17-Year Exploration. EXPLORE
• Nelson RD. (2019). Connected: The Emergence of Global Consciousness. ICRL Press