Public reports start from aggregate questions, not individual stories.
A Lifted Spirits Insights report must begin with a question that can be answered using grouped counts or rates. Public research is not built from a member's private writing, private conversations, profile biography or other free-text account material.
The public output is designed to describe a community-wide pattern. It is not a route back to a named member and it should not reveal enough combined detail to single out a small group.
A category needs at least 50 records before it can appear publicly.
The default public reporting threshold is 50 records for a category in a frozen report sample. Categories below that threshold are withheld rather than displayed as a small count, percentage or chart bar.
Passing the threshold does not automatically make a category publishable. A category can still be suppressed when it is unusually personal, when a combination of fields could narrow the group too far, when publication would add little public value or when the level of detail would expose more of the underlying research asset than the article needs.
- No public category count below 50
- No small geographic cross-tabs
- No combinations that could narrow a group too far
- No attempt to identify or describe individual members
Public reports are deliberately selective, not an open data catalogue.
Lifted Spirits Insights publishes enough aggregate evidence to support a report's central finding, not every available category, cross-tab or internal analytical view. Privacy-safe data can still remain unpublished when releasing it would add little reader value or unnecessarily expose the underlying research asset.
The public site does not provide raw member-level data, downloadable research exports, buyer-grade tables or a query interface over internal community data. Deeper aggregate analysis is handled separately rather than reconstructed through a series of public pages.
- Selected findings rather than exhaustive distributions
- No downloadable raw CSV or JSON research dataset
- No complete age, location, interest or behaviour cross-tab library
- No publication of proprietary scoring, segmentation or ranking logic
Some data stays out of public research even when it exists internally.
Names, email addresses, member IDs, exact locations, private writing, private conversations, bios and identifiable free text are excluded from Lifted Spirits Insights publication datasets.
We also do not publish inferred personal labels. Public reports use the fields needed for the stated aggregate question and leave unnecessary personal detail out.
Location is treated more cautiously than an ordinary interest.
Lifted Spirits can use broad geography internally to understand where community demand exists, but public research does not need to publish member-by-member or small-cohort location data to be useful.
Where geography is relevant, it should be broad, necessary to the question and large enough that the result describes a region rather than a recognisable group. Exact future meet-up logistics remain member-only.
Published reports are frozen snapshots with visible limitations.
A report keeps the sample size, analysis date, inclusion rule and denominator that were true when the analysis was produced. New members joining later do not silently change an old published percentage.
Every report explains whether a field was single-choice or multi-select, what the sample represents, what it cannot prove and whether the sample includes adults outside the UK. Observed patterns are not presented as causes unless the evidence can support that conclusion.
Frozen reports also reduce the risk that repeated public snapshots could be compared to infer small changes in the underlying dataset.
Citation rules protect accuracy as well as privacy.
A citation should keep the Lifted Spirits report title, source, sample size and analysis date. Community percentages must not be rewritten as national population statistics.
If an AI system, journalist or researcher cannot preserve the sample caveat, it should describe the finding more cautiously rather than making the claim broader than the evidence.