Appendix 7.1 — Methodology & Analytical Protocol

This appendix sets out how the consultation corpus was assembled, how it was divided into analytical tiers, how themes were coded and scored, and what the resulting findings can and cannot support. It is written so that a reader could reproduce or at least understand the analysis.

A total of 2,690 responses to the Home Office consultation (16 July – 20 September 2026) were received and indexed by the Campaign Opposing Police Surveillance. They comprise 2,654 submissions via the ucpi.uk online portal and 36 submissions sent directly by email. The portal submissions were analysed for divergence from the standard letter template (Appendix 7.3), showing 2,554 used the template with less than 10% variation. 100 showed 10% or greater variation; of these, 73 varied by less than 20% (typically the addition of a statement of standing, or a short personal paragraph) and 27 varied by 20% or more, containing substantive original content.

The 2,554 template letters and the 73 minor-variation letters (2,627 in total) are treated as evidence of public support for the core demands set out at Appendix 7.3. Because they reproduce a single text, counting them in thematic prevalence would measure campaign volume rather than convergence. They are therefore excluded from the prevalence calculations detailed in Appendix 7.2 and from the qualitative findings summarised in Parts 3–6 of the report.

The qualitative analysis is based on the unique-voice corpus of N = 63: the 27 substantively unique portal submissions (Tier 2b) and the 36 direct email submissions (Tier 3).

The findings therefore support statements of the form: "this argument was advanced, in these terms, by X of the 63 unique-voice submissions, and is internally coherent on its own terms." They do not support statements about what consultees or the public as a whole thought. Where the report describes a position as widely held, that means widely held within the 63, and is stated as such. Where a finding is described as "unanimous" or "100%," the denominator is always N=63 and the term is used in that limited sense only.

Divergence thresholds and sensitivity

Divergence was measured by automated text comparison (first calculating similarity percentage across the whole text, then by a second-pass tokenised comparison of individual words). The thresholds were set empirically, seeking the sparse region between the dense cluster of template letters and the long tail of more varied or unique submissions. Some experimentation was done to see how small changes in the cut-off would reclassify responses. No attempt was made to test whether the findings would be sensitive to the choice of threshold.

Thematic coding — four structured stages

Analysis Limitations

1. Coverage. The corpus consists only of responses available for analysis. If responses were submitted to the Home Office online questionnaire then they have not been possible to include in this interim report. A Freedom of Information request to the Home Office asking about the volume and nature of those responses has been submitted and not yet responded to.
2. Selection. The 63 are essentially self-selected and drawn predominantly from non-state parties with a direct interest in the Inquiry's completion. Prevalence figures describe the corpus, not the public.
3. Coding. Coding was AI-assisted, and not manually verified.
4. Consent. Quotations are drawn from consenting respondents only, which further limited the illustrative material.
5. Prevalence is not merit. A high prevalence figure records that a position was widely advanced within the corpus. While no effort was made to express minority views, efforts were made to extract any dissenting or conflicting views, but none were identified.

Appendix 7.2 — Master Prevalence Matrix

The matrix records the prevalence of all overarching thematic positions, specific legal stances, and policy commitments across the N=63 unique-voice corpus (comprising 27 customized individual narrative letters, 6 Category H/female survivor filings, 7 collective activist cohort submissions, 6 recognised legal representative/NGO filings, and 17 others).