Show what a published rating actually aggregates

Identify what the score claims to describe

Record the number, maximum possible score, count label, surrounding wording and URL together. A store-service score placed beside a product can be read as a product rating unless its subject is clear. Likewise, a count of ratings, a count of written reviews and a count of purchasers describe different populations. Copy the label actually displayed rather than silently changing it in the working record.

Name the underlying service or database, the entity or product it covers, and the time window. Keep the evidence for each display separately when a homepage badge, product widget and collection summary use different sources. An aggregate has to be traceable to the set it describes; a recognizable review-service logo is not that trace.

Reconstruct the population before the arithmetic

Use an authorized review export, documented service report or other genuine record that identifies the included entries and their ratings. Keep personal information in the controlled source system. In the worksheet, record only the source reference, included count, relevant dates and selection rules. Establish whether the same review appears through more than one import and how any duplicates were handled.

Compare the included set with the published collection and moderation policy. Document product filters, date filters, migrations, removals and other exclusions. Do not remove unfavorable entries to make the summary match the number the business wants to display. Where incentives or other material relationships exist, preserve those facts for a disclosure assessment; a mathematically correct score does not settle whether its presentation is misleading.

The FTC's truth-in-advertising guidance supports truthful, nonmisleading claims backed by evidence. Its endorsement guidance also addresses honesty and unexpected material relationships. Those principles support this evidence check; they do not prescribe this worksheet's calculation or decide the legality of a particular rating display.

Reproduce the calculation on the same basis

For a simple arithmetic mean on one rating scale, add the included rating values and divide by the number of those rated entries. Record the unrounded result, the displayed precision and the rounding rule. The accompanying count must describe the population used or clearly identify a different one. Written reviews without a numerical rating cannot silently be included in a numerical average's denominator.

Do not average already rounded product averages as though each represented the same number of reviews. If a combined mean is appropriate for the stated subject and the groups do not overlap, calculate from the underlying ratings, or from each group's unrounded sum and count on the same scale. Keep products, businesses and rating scales separate when their combination would misdescribe the subject.

A service may supply its own documented scoring method rather than a simple mean. In that case, preserve the service's named method, report and date, and attribute the score accurately. If the method or included population is unavailable, record that limit. Do not label your own reconstructed number as the service's official score.

Correct the representation and assign its refresh

Choose the correction from the mismatch: fix the count if it was copied incorrectly, correct the subject label if it names the wrong population, or replace a stale figure only when a newer supported figure exists. If you cannot substantiate the aggregate, withdraw the unsupported numerical claim pending investigation rather than inventing a replacement. Preserve the underlying genuine reviews and the history of the correction.

Record the source's as-of date separately from the date the website last fetched or published it. A successful widget refresh does not establish that the source data itself is new. Assign an owner to compare the published display with its next documented update, especially after imports or filter changes.

A Prism website-review consultation can be scoped around the public display, its surrounding claims and the evidence gap. Bring the page URLs and a nonpersonal summary of the basis. The consultation sets the work and any reporting; legal interpretation remains with qualified counsel, and no rating correction establishes processing approval.

Aggregate-rating basis

Complete one copy per published score. Treat a score as reconciled only when its subject, population, count, method and date explain the visible display. Mark an unavailable source or calculation as unresolved; never fill the table with invented reviews or customer identities.

Worksheet entries are not submitted by Prism’s worksheet and are not saved by the site. Use record types, availability, anonymized observations, or match/mismatch results. Do not enter government identifiers, customer names or addresses, customer messages, receipt-access links, card or bank details, passwords, or keys. Send sensitive documents only through the provider’s verified secure channel.

These are temporary notes. Leaving or reloading this page may clear them. Worksheet entries are not sent automatically. If you copy notes into the consultation message and submit the form, Prism receives them as part of your request.