EVIDENCE BASE
The science, with receipts
Every mechanic in RealityWipe maps to published research. Below is that evidence — rated honestly. Strong means meta-analysis or replicated behavioral trials. Mechanistic means the brain-level mechanism is shown but no behavioral outcome has been measured. We don't upgrade ratings for marketing.
STRONG EVIDENCE
Speaking words aloud makes them substantially more memorable than reading them silently.
The production effect: across many experiments, words read aloud are remembered 10–20 percentage points better than words read silently, because speaking adds motor and auditory encoding dimensions that are available again at retrieval. Recording a line aloud is a memory intervention before playback even starts.
Effect size: +10–20 pp
Sources: MacLeod & Bodner (2017) · Frontiers in Psychology (2014)
STRONG EVIDENCE
If-then plans reliably close the gap between intending a change and doing it.
Gollwitzer & Sheeran's meta-analysis of 94 independent tests found implementation intentions ('when situation X arises, I will do Y') improve goal attainment with a medium-to-large effect over goal intentions alone. The trigger-pinned protocol card is this construct verbatim.
Effect size: d = 0.65 (94 studies)
Sources: Gollwitzer & Sheeran (2006) · US National Cancer Institute (2020)
MODERATE EVIDENCE
Generic positive affirmations backfire for the people who need them most — lines must be believable and tied to action.
Wood, Perunovic & Lee (2009): participants with low self-esteem felt worse after repeating 'I'm a lovable person'. Positive self-statements helped only those who already believed them. This is the empirical basis for banning soft affirmations and generating lines from the user's exact words.
Sources: Wood, Perunovic & Lee (2009) · ScienceDaily (2009)
MODERATE EVIDENCE
Naming the script and talking about it from outside — 'that's the Readiness Lie, not me' — is a low-effort, well-replicated regulation tool.
Kross and colleagues showed that distanced self-talk (using one's name or third person) improves emotion regulation and goal pursuit essentially without cognitive effort, with converging ERP and fMRI evidence and behavioral replications from public speaking to healthier eating.
Sources: Kross et al. (2014) · Scientific Reports (2017) · Furman, Kross & Gearhardt (2020)
MODERATE EVIDENCE
Hypnotherapy's effect sizes are real, and the transferable ingredient is the delivery structure: paragraph-length, present-tense, permissive, imagery-rich narrative under relaxed attention.
Valentine et al. (2019) meta-analysis of 15 studies: hypnosis reduced anxiety with g = 0.79 at end of treatment and 0.99 at follow-up, strongest when combined with another intervention. A 20-year systematic review of meta-analyses finds robust effects for pain, distress and IBS. Suggestion scripts across this literature share a form — flowing narrative paragraphs, present tense, sensory imagery — which the install script mirrors deliberately. The label 'hypnosis' is NOT used in product: hypnotizability varies widely and clinical claims are out of scope.
Effect size: g = 0.79 anxiety (15 studies), 0.99 at follow-up
What this does not show: Evidence is for clinician-delivered or guided hypnosis on clinical outcomes — not for self-recorded scripts in a consumer app. We borrow the script FORM only; no trance induction, no clinical claims.
Sources: Valentine, Milling, Clark & Moriarty (2019) · Frontiers in Psychology (2024)
MODERATE EVIDENCE
Vivid sensory imagery of DOING the goal action motivates more than verbally stating the same content — the script's proof-action movement should be cinematic, not declarative.
Elaborated Intrusion theory work shows mental imagery carries more affect and feels more real than verbal processing of identical content; instructed imagery of goal striving promotes goal behavior more than verbal goal thinking (Renner et al. 2019, 'motivational amplifier'). Functional Imagery Training — training habitual sensory imagery of personal incentives — outperformed Motivational Interviewing for weight loss in RCT work (Solbrig et al.).
What this does not show: FIT trials are clinician-delivered protocols in specific domains (weight, self-harm pilots, anxiety early-phase). We apply the mechanism (imagery-rich phrasing), not the trial protocol.
Sources: Renner, Murphy, Ji, Manly & Holmes (2019) · Solbrig (2019) · Behaviour Research and Therapy (2025)
MECHANISTIC EVIDENCE
Your own voice gets special treatment in the brain — it engages self-referential networks other voices don't.
fMRI and PET studies show hearing one's own voice activates right inferior frontal regions and the self-referential network (medial prefrontal and parietal cortices), with processing that measurably differs from hearing other voices.
What this does not show: Shows unique processing, not a behavior-change outcome.
Sources: Kaplan et al. (2008)
MECHANISTIC EVIDENCE
Emotion-regulation lines played in your own voice engage self-referential and memory circuits differently than another voice — and the effect is strongest for defusion-style lines.
A 2024 fMRI study played self-affirmation and cognitive-defusion sentences in participants' own vs others' voices. Own-voice showed distinct neural engagement; in the precuneus, own-voice affirmation intertwined self-referential processing with episodic memory retrieval, and own-voice uniqueness was expressed more strongly for cognitive defusion than affirmation.
What this does not show: n=21, synthesized voices, no behavioral outcome measured. Mechanistic support for the pitch, not proof of efficacy.
Sources: Brain Sciences (2024)
MECHANISTIC EVIDENCE
Repeating a short phrase drains its meaning — the paragraph-length varying script is the satiation-resistant format, not looped one-liners.
Semantic satiation: repeating a word or short phrase ~15–30 times causes temporary loss of meaning via neural habituation — the lexical form keeps firing while the associated meaning drops out. Design consequence: never loop the same short line within a session, keep recording lines rotating (the nightly patch does this), and carry the depth in the full narrative script rather than in repeated fragments.
What this does not show: Classic lab paradigm uses rapid massed repetition (seconds apart); daily-spaced replay is far less exposed. Treated as a design constraint, not an efficacy claim.
Sources: Jakobovits (paradigm); overview (1962) · Practical Psychology (2023)