The Silence Experiment - Phase 0 Statistical Power Analysis and Study Design Table
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Zenodo (CERN European Organization for Nuclear Research) |
| Published | 2026-08-05 |
| Authors | Clifton Bacon |
| DOI | 10.5281/zenodo.21807001 |
Participant count, not session count, is the binding constraint in Phase 0 of the Silence Experiment. At d = 0.3, τ = 0.10, S = 15 and α = 0.00167, Contrast 1 reaches 0.86 power with P = 10 participants. The stopping rule remains provisional until a pilot estimates τ, the participant-level variability of the deep-versus-shallow effect.
Power model and assumptions
Section titled “Power model and assumptions”Contrast 1 compares the deepest conditions, C6 and C7, with baseline and C1 through C5 at d = 0.3. The three-test family uses α = 0.005, with Bonferroni giving a one-sided threshold of 0.00167 per test.
The repeated-measures model is y(p,c,s) = μ(c) + u(p) + I[c∈deep]·s(p) + e(p,c,s). The participant intercept u(p) cancels in the within-subject contrast. Total variance is scaled to 1, with σe = √0.5 ≈ 0.71 and an intercept ICC of 0.5. Participant effect deviations follow s(p) ~ N(0, τ²).
Per-participant variance is V = τ² + (2/3)σe²/S for Contrast 1. Contrast 2 uses V = τ² + 2σe²/S. Power uses a noncentral t distribution from scipy.stats.nct with P - 1 degrees of freedom. Monte Carlo simulation provides the cross-check.
The empty-chamber baseline is one of the six shallower cells and enters Contrast 1. The thermal-dummy cell is excluded from Contrast 1, Contrast 2, the Jonckheere-Terpstra test, and the secondary protective-versus-suppression test.
Contrast 1 across participant heterogeneity
Section titled “Contrast 1 across participant heterogeneity”Power for Contrast 1 at S = 30 depends strongly on τ and participant count.
| P | τ = 0.05 | τ = 0.10 | τ = 0.15 | τ = 0.20 |
|---|---|---|---|---|
| 4 | 0.23 | 0.14 | 0.08 | 0.05 |
| 5 | 0.50 | 0.31 | 0.18 | 0.10 |
| 6 | 0.76 | 0.53 | 0.31 | 0.18 |
| 8 | 0.98 | 0.86 | 0.62 | 0.39 |
| 10 | 1.00 | 0.97 | 0.84 | 0.61 |
| 12 | 1.00 | 1.00 | 0.94 | 0.77 |
At τ = 0.05, the paper describes six participants as sufficient, although Table 1 gives 0.76 power. Eight participants reach 0.98. At τ = 0.10, P = 8 reaches 0.86 and P = 10 reaches 0.97. At τ = 0.15, P = 10 reaches 0.84 and P = 12 reaches 0.94. At τ = 0.20, P = 12 reaches only 0.77.
At the registered session floor of S = 15, P = 8 reaches 0.64 at τ = 0.10. P = 10 reaches 0.86 and is the smallest count satisfying the 0.80 rule at that floor. P = 8 reaches 0.86 at S = 30, but its power falls to 0.62 at τ = 0.15. The paper reports no feasible participant count at τ = 0.20.
Session floors and enrollment
Section titled “Session floors and enrollment”The registered target is P = 10, with P = 8 as the minimum acceptable count. Additional sessions help, but they cannot rescue every enrollment target. At τ = 0.10, P = 4 rises from 0.06 power at S = 10 to 0.16 at S = 40. P = 5 rises from 0.13 to 0.37 over the same session counts.
Larger samples respond differently. P = 8 rises from 0.48 at S = 10 to 0.91 at S = 40. P = 12 rises from 0.86 to 1.00. The paper therefore states that no realistic session count rescues a four- or five-participant study. Roughly eight or more participants are required at moderate heterogeneity.
The stopping rule must select sessions from power calculations. A fixed session count can satisfy the rule for one participant count and τ value but fail for another. This makes τ an explicit design parameter rather than an incidental assumption.
Contrast 2 remains underpowered
Section titled “Contrast 2 remains underpowered”Contrast 2 compares C7 directly with C6. Its noise term is three times the Contrast 1 term, so its sensitivity is substantially lower. At τ = 0.10, P = 10 reaches 0.82 power at S = 40. P = 12 reaches 0.86 at S = 30 and 0.93 at S = 40. P = 8 reaches only 0.59 at S = 40.
These values use the full 30-minute C7 period. The primary form is a duration-matched 15-minute C7 window, which provides less averaging time per session. The reported Contrast 2 powers are therefore upper bounds. The shorter window is underpowered as well.
Contrast 2 remains in the primary family at the corrected alpha. The paper treats its null result as unresolved rather than disconfirming. Definitive resolution is deferred to Phase 1 with a larger participant pool and the transmon sensor.
The secondary protective-sign test compares C6 and C7 with the empty-chamber baseline alone. It is not separately powered in the note. The paper expects its sensitivity to lie between Contrast 1 and Contrast 2, closer to Contrast 2. Its estimate and confidence interval are reported in all cases and classified as protective, suppression, or unresolved. Unresolved is the expected Phase 0 result, and a near-zero estimate whose interval does not exclude the protective effect size cannot be classified as suppression.
What a null rules out
Section titled “What a null rules out”The minimum detectable effect is the smallest d that achieves 80% power. For Contrast 1 at S = 30, the values are:
| P | τ = 0.05 | τ = 0.10 | τ = 0.15 | τ = 0.20 |
|---|---|---|---|---|
| 4 | 0.63 | 0.78 | 0.99 | 1.22 |
| 5 | 0.41 | 0.51 | 0.64 | 0.79 |
| 6 | 0.31 | 0.39 | 0.49 | 0.61 |
| 8 | 0.23 | 0.28 | 0.35 | 0.44 |
| 10 | 0.18 | 0.23 | 0.29 | 0.36 |
| 12 | 0.16 | 0.20 | 0.25 | 0.31 |
At P = 10, τ = 0.10 and S = 30, a null excludes d = 0.23 or larger. At P = 8 under the same conditions, it excludes d = 0.28 or larger. At P = 5 and τ = 0.15, a null excludes only d = 0.64 or larger.
Phase 0 cannot constrain effects near d = 0.05 to 0.10, as reported in prior consciousness-RNG work. That work measures RNG output statistics rather than T2 decoherence. The paper states that Phase 0 cannot reach those effect sizes in either measurement channel.
Pilot estimates and preregistration
Section titled “Pilot estimates and preregistration”The paper recommends a pilot with two to three practitioners and five to 10 sessions each. The pilot uses the same chamber and sensor to estimate τ and σe. Pilot participants are excluded from the confirmatory sample when the pool permits, and pilot data are never pooled into Phase 0. The simulation is locked against pilot values before confirmatory data collection.
The preregistration should specify the target P, a plausible τ range, and the S needed for at least 0.80 power at d = 0.3. It should also set a minimum session floor, such as S of at least 15. The stopping rule must use simulated power for Contrast 1 across the preregistered τ range.
The protocol should report the minimum detectable effect associated with a null. The target d = 0.3 is a feasibility-driven value, not a theory-derived prediction. No current theory predicts an effect size for this configuration.
Small-sample and recruitment contingencies
Section titled “Small-sample and recruitment contingencies”If P is below 8 at τ = 0.10, the paper recommends making the Bayesian track primary and reporting the frequentist test for completeness. Bayes factors are likelihood ratios and do not depend on the degrees of freedom or critical-value inflation that weaken the small-sample frequentist test.
At P = 5, a Bayesian factor from a mixed-effects model accumulates evidence with the data. The frequentist t-test faces a critical value of about 6.3, which the paper describes as potentially unreachable at d = 0.3.
If the required P exceeds the recruitable pool, the Bayesian track becomes primary. The minimum detectable effect must be recalculated and reported. Phase 0 is then reframed as a pilot for estimating τ and the noise model needed for a properly powered Phase 1.
Analytic and simulation agreement
Section titled “Analytic and simulation agreement”Monte Carlo checks used 20,000 replications per design cell. Under the null, the empirical false-positive rate was 0.0015, matching the 0.00167 target.
For P = 8, S = 15 and τ = 0.10, analytic Contrast 1 power was 0.642, compared with 0.643 from simulation. The P = 10, S = 20 design gave analytic power of 0.924 and was confirmed by simulation. The P = 6, S = 30, τ = 0.05 design gave 0.760 and was also confirmed.
The resulting design must estimate τ before fixing S and must meet the 0.80 power threshold for Contrast 1. Phase 0 can estimate effect heterogeneity and session noise for Phase 1, but its registered design does not resolve C7 versus C6.
View Original Abstract
Two supplementary documents supporting The Silence Experiment, a pre-registered protocol testing the von Neumann-Wigner interpretation of quantum mechanics against environmental decoherence theory, using EEG-verified observer cognitive state as the independent variable and T₂ decoherence on an NV-center diamond sensor as the dependent variable. This record contains the Stage 1 study design table and the Phase 0 statistical power analysis. They are deposited together because each depends on the other: the design table’s sampling plan and stopping rule are derived from the power analysis, and the power analysis is computed against the design the table specifies. Study design table. Prepared to the PCI Registered Reports Stage 1 template, landscape format. The table maps each research question to its hypothesis, sampling plan, analysis plan, and pre-registered interpretation under every anticipated outcome, covering four hypotheses and three auxiliary assumptions. H1 (Contrast 1), confirmatory: T₂ at [C6, C7] exceeds T₂ at [chamber baseline, C1-C5]. Because the empty-chamber baseline sits in the comparison group, the predicted direction commits the protocol to a coherence-protective contribution (Γcog < 0) rather than merely to the absence of a decohering one. H1b, secondary: [C6, C7] against the empty-chamber baseline alone, the only pre-registered test that separates the protective prediction from a suppression pattern in which the deep states merely return the sensor to its empty-chamber value; no primary test distinguishes them. H2 (Contrast 2): C7 versus C6, retained in the primary family but underpowered in Phase 0, computed on a duration-matched 15-minute C7 window anchored to EEG-verified state entry. H3: Jonckheere-Terpstra ordered test across C1-C7. A1 covers the frozen EEG classifier and its pilot firewall. A2 covers environmental isolation and the metrology co-product. A3 covers the pre-registered positive control, which establishes that the measurement chain is sensitive rather than merely quiet, and without which a null could not be reported as an upper bound. A further row specifies conventions common to all hypotheses: randomisation with a lodged seed and allocation concealment, permuted blocks balanced across time-of-day strata, the frozen T₂ estimation pipeline, exclusion rules at exactly three levels with no outlier rule on T₂ values, and missing-data handling. The table is design-level and host-independent; host-dependent details are supplied in the Stage 1 manuscript body once a host institution is secured. Phase 0 power analysis. A technical note computing achievable power across the Phase 0 design space and resolving the dependency between the protocol’s power section and its stopping rule. The core finding is that participant count, not session count, is the binding constraint. Each participant’s contrast estimate is modelled as a group-level effect plus a participant-specific deviation with standard deviation τ; because τ is unknown until pilot data exist, power is computed across τ ∈ {0.05, 0.10, 0.15, 0.20}. All figures use the non-central t distribution with df = P − 1, treating the participant as the unit of inference, cross-checked against Monte Carlo simulation. At moderate heterogeneity (τ = 0.10) with S = 30 sessions per condition, 8 participants yield power of 0.86 for Contrast 1 and 10 yield 0.97; at τ = 0.15, 10 yield 0.84 and 12 yield 0.94. No realistic session count rescues a four-to-five person study. Contrast 2 is substantially less powered at every design point, because comparing one condition against one triples the per-participant noise variance; even P = 12, S = 40 reaches only 0.93. A null on Contrast 2 is therefore uninformative rather than disconfirming. Low power qualifies the interpretation of a null; it is not grounds for moving a pre-registered directional test out of the confirmatory family after the design is fixed. The note also reports what a null buys: at P = 8 and τ = 0.10, a null rules out d ≥ 0.28; at P = 5 and τ = 0.15, only d ≥ 0.64. Phase 0 cannot reach the d ≈ 0.05-0.10 effect sizes reported in prior consciousness-RNG literature, which in any case tests a different measurement channel. Seven recommendations follow, covering pre-registration of P and τ, a firewalled pilot, a power-derived stopping rule, statement of the minimum detectable effect, use of the Bayesian track at small P, a contingency for unreachable P, and retention of Contrast 2 in the primary family with an explicit underpowered label. Type I error is validated under the null at an empirical false-positive rate of 0.0015 against the 0.00167 target, with 20,000 replications per cell. Common statistical framework. Family-wise alpha 0.005 one-sided across the three primary tests; Bonferroni per-test alpha ≈ 0.00167. The secondary discriminator is reported at alpha 0.005 outside the correction. Target minimum detectable effect d = 0.3 on Contrast 1, feasibility-driven rather than theory-derived: no current theory predicts an effect size for a consciousness-dependent T₂ shift, nor its sign. A thermal-dummy control cell is recorded alongside the eight analysed cells and excluded from every confirmatory test, and therefore from every power figure.
Advanced Thermal Simulation Service Diamond Heat Spreader Performance Analysis
View Specs →