AI Qualitative Analysis: SOUND-BITES Pilot

This dataset can answer which TDF domains predict acceptance, how help-seeking pathways operate, and where implementation pain points occur. Researchers can link domain-coded interview responses to survey and clinical outputs to model barriers and drivers.

Acceptability drivers: which TDF domains (knowledge, beliefs about capabilities, social influences) predict volunteer and client acceptance.

Help-seeking pathways: cross-reference interview reasons with 6-month help-seeking survey responses to model barriers.

Implementation pain points: operational themes such as training, timing, noise, and device handling that can be prioritized for SOPs.

Comparative analysis can reveal differences between clients who received a 'Refer' versus 'Pass', between volunteers by tech familiarity, and across the 8 Meals on Wheels sites. These comparisons can inform targeted implementation changes.

Clients who received a 'Refer' vs 'Pass' (Sound Scouts output): compare expressed intentions and perceived barriers.

Volunteer experience by tech familiarity: training feedback and comfort delivering checks.

Site-level variation across the 8 Meals on Wheels locations to flag local process changes.

Researchers must account for self-selection bias because interviews are optional, and report recruitment and refusal reasons transparently. Researchers should link qualitative notes to audiometric metadata to assess environmental influences on comments and test results.

Interviews are optional, expect self-selection bias; report recruitment and refusal reasons.

Environmental noise in-home testing may influence test results and participant comments, link qualitative notes to audiometric metadata.

Manual transcript checks and participant review (protocol step) are essential before automated coding to preserve rigor.

Evidano ingests bulk audio files and REDCap CSVs securely, with encryption and a policy that customer data is not used to train third-party models. Evidano supports auto-redaction of PII during transcription when consent allows.

Bulk import audio files and REDCap CSVs; data is encrypted and Evidano does not use customer data to train third-party models.

Auto-redact PII during transcription if required (consent-dependent).

Evidano auto-transcribes using intelligent verbatim and supports custom dictionaries, batch speaker-labeling, and a QC queue that mirrors manual participant transcript review. Evidano enables validation of samples and manual correction before coding.

Auto-transcribe with intelligent verbatim, using a custom dictionary (e.g., 'Arclight', 'Sound Scouts', site names).

Batch speaker-labeling and a QC queue for manual checks that mirror the study’s protocol step of participant transcript review.

Evidano imports the TDF codebook and pre-codes transcripts using model-tuned patterns while tracking inter-coder agreement and flagging low-confidence passages for human arbitration. Human coders reconcile codes with Evidano recording consensus and disagreements.

Import the TDF codebook (deductive codes) and let Evidano pre-code transcripts using model-tuned patterns.

Human coders review and reconcile: Evidano tracks inter-coder agreement and flags low-confidence passages for arbitration.

Evidano generates theme frequency tables, co-occurrence networks, and hierarchical code-to-subcode trees and can run automated cross-segment comparisons such as Pass versus Refer. Evidano provides quantitative-style summaries suitable for mixed-methods reporting.

Generate theme frequency tables, co-occurrence networks, and hierarchical code→subcode trees.

Run cross-segment comparisons automatically (e.g., Pass vs Refer, volunteer tech-experience strata) and get effect-size style summaries for mixed-methods reports.

Evidano produces dashboards, exportable tables, and slide-ready visuals, and it exports cleaned transcripts, annotated quotes, and a reproducible audit trail for ethics and peer review. These outputs support conference abstracts and community reports.

One-click dashboards: word clouds, co-occurrence maps, and exportable tables for conference abstracts or community reports.

Export cleaned transcripts, annotated quotes, and a reproducible audit trail for ethics/peer-review.

Yes. Evidano accepts standard codebook formats and maps imported codes to hierarchical analytics for frequency and co-occurrence reporting. Evidano supports CSV and JSON imports and will align imported codes with its analytics structure.

Yes. Evidano accepts standard codebook formats (CSV/JSON) and maps imported codes to its hierarchical analytics for frequency and co‑occurrence reporting.

Evidano tracks new-theme velocity across ordered interviews and flags when the new-theme rate drops below a configurable threshold, providing an indicator to combine with human judgment. The protocol cites saturation at approximately 9–17 interviews as a reference point.

Evidano tracks new-theme velocity across ordered interviews and flags when new-theme rate drops below a configurable threshold, a useful indicator alongside human judgment (the protocol cites saturation ≈9–17 interviews).

Yes. Evidano encrypts data in transit and at rest, supports role-based access controls, and does not use customer data to train public models, making it suitable for research under ethics approvals. Evidano's security features align with HREC requirements referenced in the protocol.

Data is encrypted in transit and at rest, role-based access controls are supported, and Evidano does not use your data to train public models, suitable for research under ethics approvals like Macquarie University HREC ID: 16818.