Beautiful Digital Experiences Hypothesis Registry v0.1

Status vocabulary

• Initialized: claim is specified but not adequately tested.
• Provisional support: evidence bears on the predicted mechanism, with important transfer limits.
• Mixed: supporting and contradicting evidence or boundary reversals are known.
• Narrowed: the original form was too broad and a scoped revision is active.
• Rejected: current claim form is contradicted; history is preserved.

Required hypotheses

HY-BEAUTY-001

• Claim: Aesthetic quality improves perceived usability more reliably than measured task performance.
• Mechanism: halo/affect-as-information changes expectations and post-use appraisal; it does not necessarily change interaction structure.
• Prediction: appearance manipulation produces a larger and more consistent effect on perceived-usability measures than on completion, error, or time.
• Support: EV-BDE-001, EV-BDE-002, EV-BDE-003.
• Contradiction: some task and goal conditions show performance effects, including slower completion, so “no performance effect” is not the claim.
• Boundary conditions: task stakes, goal orientation, exposure, age, domain, usability severity, and brand/content cues.
• Alternatives: experiment-demand effects; perceived craftsmanship signals actual engineering quality; positive emotion changes persistence.
• Test status: provisional support; confidence: medium.
• Falsifier: robust, larger objective-performance effects across preregistered, repeated, multi-domain studies.

HY-BEAUTY-002

• Claim: People prefer category-recognizable designs containing limited controlled novelty.
• Mechanism: prototypicality supports prediction and fluency; novelty supports interest, identity, and informational reward.
• Prediction: category-fit and beauty form a conditional inverted-U across novelty, with the optimum varying by audience and task.
• Support: EV-BDE-005, EV-BDE-006, EV-BDE-007, EV-BDE-008, EV-BDE-009.
• Contradiction: EV-BDE-008 favored high prototypicality and low complexity in rapid screenshots; it did not demonstrate a novelty optimum.
• Boundary conditions: expertise, epistemic motivation, repeated use, product risk, category maturity, and cultural model.
• Alternatives: novelty is merely production-value or prestige signaling.
• Test status: initialized; confidence: low.
• Falsifier: monotonic preference for prototypicality or novelty after controlling execution quality and familiarity.

HY-BEAUTY-003

• Claim: Beautiful interfaces use consistency of relationships rather than identical repetition.
• Mechanism: invariant spatial, typographic, and semantic relations support prediction while controlled surface variation preserves information and rhythm.
• Prediction: relationally consistent variants outperform identical repetition on coherence and interest without harming task performance.
• Support: indirect only: EV-BDE-009, EV-BDE-010 and existing Composition Science relational-legibility work.
• Contradiction: none directly tested; component sameness may improve learning in dense applications.
• Boundary conditions: frequency of use, system scale, safety, localization.
• Alternatives: quality comes from simple component consistency, not relations.
• Test status: initialized; confidence: low.

HY-BEAUTY-004

• Claim: Art direction contributes more to perceived website quality than decorative interface styling.
• Mechanism: imagery supplies high-area, semantically rich signals of specificity, material quality, atmosphere, and budget.
• Prediction: changing imagery quality and coherence while holding layout constant produces larger effects than changing ornamental UI styling while imagery is fixed.
• Support: no direct causal evidence registered.
• Contradiction: application-like products with little imagery can be highly admired.
• Boundary conditions: image-led versus task-led category, viewport, content volume.
• Alternatives: imagery operates as brand or production-budget signal.
• Test status: initialized; confidence: very low.

HY-BEAUTY-005

• Claim: Perceived polish is cumulatively produced by many resolved details.
• Mechanism: each mismatch is weak evidence of incomplete control; repeated matches increase inferred craftsmanship and reduce prediction errors.
• Prediction: incremental detail repair yields monotonic or thresholded gains in craftsmanship ratings with macro-composition held constant.
• Support: EV-BDE-010 identifies craftsmanship as a separable perceived facet.
• Contradiction: a dominant art-directed feature may overwhelm small defects at first exposure.
• Boundary conditions: exposure duration, defect salience, expertise, task pressure.
• Alternatives: one global latent impression causes detail ratings.
• Test status: initialized; confidence: low.

HY-BEAUTY-006

• Claim: A design can violate surface conventions and remain usable when deeper perceptual and semantic structure is preserved.
• Mechanism: grouping, hierarchy, labels, mappings, feedback, and state continuity carry action meaning below conventional styling.
• Prediction: convention-breaking variants with preserved deep structure retain comprehension and recovery while increasing expressive-aesthetic ratings.
• Support: EV-BDE-009 distinguishes expressive aesthetics; EV-BDE-007 makes fluency expectation-dependent.
• Contradiction: unfamiliar controls can impose learning and accessibility costs.
• Boundary conditions: novice/expert status, risk, frequency, assistive technology.
• Alternatives: users transfer conventions that experimenters failed to identify.
• Test status: initialized; confidence: low.

HY-BEAUTY-007

• Claim: Beauty judgments stabilize differently when based on complete experiences rather than static screenshots.
• Mechanism: interaction reveals temporal coherence, content resilience, friction, state completeness, and performance.
• Prediction: screenshot–experience agreement falls as task/state diversity grows; repeated-session ratings show lower halo and greater sensitivity to craft and friction.
• Support: EV-BDE-004 and the scope boundary in EV-BDE-008.
• Contradiction: rapid appeal ratings can be internally stable across short exposure durations.
• Boundary conditions: site type, interaction depth, media weight, exposure interval.
• Alternatives: rating change is mere exposure rather than better evidence.
• Test status: provisional support; confidence: low-to-medium.

HY-BEAUTY-008

• Claim: Designers and general users differ in admired features but overlap on coherence, intentionality, legibility, and execution quality.
• Mechanism: expertise changes learned categories, attention, and novelty reward; common perceptual constraints preserve partial agreement.
• Prediction: experts weight originality and detail more; both groups penalize severe incoherence and illegibility.
• Support: EV-BDE-009 and EV-BDE-010 provide candidate shared dimensions but do not test audience groups.
• Contradiction: no direct registered evidence.
• Boundary conditions: design education, domain expertise, culture, status signaling.
• Alternatives: apparent expertise effects are vocabulary or response-style effects.
• Test status: initialized; confidence: very low.

HY-BEAUTY-009

• Claim: Visual trends succeed partly because they satisfy recurring psychological or cultural needs, not only imitation.
• Mechanism: trends package needs for novelty, nostalgia, status, tactility, clarity, optimism, or technical demonstration into recognizable forms.
• Prediction: adoption explanations and preference effects cluster by need, and useful mechanisms survive after surface motifs decline.
• Support: EV-BDE-005, EV-BDE-006, EV-BDE-007 provide possible psychological routes.
• Contradiction: diffusion can occur through platform defaults and mimetic pressure without preference.
• Boundary conditions: professional networks, tooling, economic cycle, medium.
• Alternatives: availability, vendor promotion, and portfolio incentives.
• Test status: initialized; confidence: very low.

HY-BEAUTY-010

• Claim: Durable digital designs combine stable compositional principles with replaceable stylistic layers.
• Mechanism: perceptual and semantic relationships age more slowly than culturally coded surface motifs.
• Prediction: blinded historical evaluations retain coherence/craft ratings better than trend-forwardness ratings; surface restyling can preserve task structure.
• Support: indirect Composition Science evidence only.
• Contradiction: interaction conventions and compositional norms also age.
• Boundary conditions: platform change, content category, brand heritage.
• Alternatives: durability is mainly repeated exposure or prestige.
• Test status: initialized; confidence: low.

HY-BEAUTY-011

• Claim: Minimalism succeeds aesthetically only when reduction increases emphasis, coherence, or perceived material quality.
• Mechanism: removal decreases competition and increases signal contrast, but beyond a point it removes identity, cues, and informational reward.
• Prediction: reduced variants improve ratings only while hierarchy and specificity remain; sterile and ambiguous variants lose warmth, interest, or usability.
• Support: EV-BDE-006, EV-BDE-008, EV-BDE-010.
• Contradiction: low-complexity screenshots were preferred in EV-BDE-008, but minimalism was not independently manipulated.
• Boundary conditions: information density, art direction, brand familiarity, task.
• Alternatives: “minimal” ratings proxy fashion or luxury.
• Test status: initialized; confidence: low.

HY-BEAUTY-012

• Claim: Accessibility constraints can increase beauty when they improve hierarchy, clarity, contrast, and feedback.
• Mechanism: perceptual robustness and explicit state communication can also create coherence and craft; constraint can focus expression.
• Prediction: well-art-directed accessible revisions improve inclusive performance without reducing, and sometimes increasing, aesthetic profiles.
• Support: EV-BDE-012 defines relevant constraints; no causal beauty evidence.
• Contradiction: some remediations may reduce a specific mood or brand expression.
• Boundary conditions: remediation quality, disability, display, context, original design.
• Alternatives: gains come from general redesign effort, not accessibility.
• Test status: initialized; confidence: very low.

HY-BEAUTY-013

• Claim: Beauty depends partly on fulfillment or skillful revision of category expectations.
• Mechanism: internal category models establish priors for form, trust, information density, and interaction.
• Prediction: identical visual languages receive different beauty, trust, and appropriateness ratings when transferred across categories.
• Support: EV-BDE-005, EV-BDE-007, EV-BDE-008; EV-BDE-011 shows visual cues enter credibility judgments.
• Contradiction: strong brand or expressive authorship may establish a new category model.
• Boundary conditions: category knowledge, risk, culture, brand, novelty tolerance.
• Alternatives: ratings follow familiarity alone or inferred production budget.
• Test status: provisional mechanism support; confidence: low.

HY-BEAUTY-014

• Claim: Repeated exposure can raise liking through familiarity but also reveal superficial novelty and unresolved details.
• Mechanism: fluency increases with exposure while task evidence updates the initial halo.
• Prediction: simple unfamiliar variants may rise, novelty-led variants may fall, and well-resolved variants remain stable; changes correlate differently with task evidence.
• Support: EV-BDE-004, EV-BDE-005, EV-BDE-006.
• Contradiction: registered longitudinal evidence mainly shows halo attenuation, not all proposed trajectories.
• Boundary conditions: exposure spacing, task variation, initial usability, mastery.
• Alternatives: regression to the mean and memory effects.
• Test status: mixed/initialized; confidence: low.

HY-BEAUTY-015

• Claim: Visual quality cannot be evaluated accurately without realistic content, responsive and interaction states, and repeated-use scenarios.
• Mechanism: ideal screenshots omit stressors that reveal compositional resilience, state completeness, accessibility, and temporal craft.
• Prediction: rankings based on ideal heroes diverge from rankings after long content, errors, localization, mobile reflow, keyboard use, and repeated tasks.
• Support: EV-BDE-004, EV-BDE-008 boundary, EV-BDE-012.
• Contradiction: first-impression appeal is itself a valid, stable outcome for some contexts.
• Boundary conditions: claim type; static campaign pages may need less interaction evidence than tools.
• Alternatives: divergence reflects content quality rather than visual-system resilience.
• Test status: narrowed: accurate complete-experience quality requires these states; first-impression beauty does not. Confidence: medium on construct logic, low empirically.

Additional hypotheses exposed by Cycle 1

HY-BEAUTY-016

• Claim: Appropriateness moderates the conversion of coherence and expression into trust and desire.
• Mechanism: the same aesthetic cue can signal competence in one genre and irresponsibility in another.
• Prediction: category-transfer manipulations create beauty–trust dissociations.
• Status: initialized; confidence: very low.

HY-BEAUTY-017

• Claim: First-impression beauty and durable beauty are related but distinct constructs.
• Mechanism: first impressions weight global visual statistics and priors; durable judgments incorporate temporal coherence, friction, content, and state completion.
• Prediction: they load on correlated but separable factors across repeated sessions.
• Status: initialized; confidence: low.

Revision history