---
name: Rating
slug: rating
category: component
status: published
created: 2026-08-21T00:00:00.000Z
modified: 2026-08-21T00:00:00.000Z
definition: A row of stars or similar marks that both shows a score and, when
  interactive, takes one.
aliases:
  - name: star rating
    source: community
  - name: rate
    source: ant-design
  - name: score input
    source: community
  - name: rating indicator
    source: hig
tags:
  - forms
relations:
  contrastWith:
    - meter
    - reaction-picker
  variantOf: []
  partOf: []
  seeAlso: []
implementations:
  - system: hig
    name: Rating indicators
    url: https://developer.apple.com/design/human-interface-guidelines/rating-indicators
sources:
  - title: Rating indicators, Apple Human Interface Guidelines
    url: https://developer.apple.com/design/human-interface-guidelines/rating-indicators
demo: inline
exhibit: false
useWhen: a score expressed as filled marks out of five
---

A rating is two components wearing one costume. Read-only, it reports a score
that already exists: the average a product earned, the difficulty of a recipe,
the strength of a match. Interactive, it is an input, and the only input whose
entire range is drawn at once as a row of identical marks. Both forms use the
same picture, which is why the word covers both, and why a specimen has to make
clear which one it is: a rating you can click asks for an opinion, and a rating
you cannot is telling you someone else's.

Five marks is the convention, stars are the default drawing, and neither is a
rule. Ten-point scales, three-point coffee-bean scales, and thumbs up or down
are all ratings. What matters is that the scale is short enough to read without
counting and that the marks are ordered, so a partly filled row means something
at a glance. Averages need a half step or a fractional fill to be honest, since
a row that rounds 4.4 up to five stars is lying about the number printed beside
it. That number, and the count of votes behind it, usually belongs next to the
marks: stars alone cannot say whether the score came from six people or six
thousand.

The interactive form is a single choice out of a few, so it is a radio group,
not five checkboxes and not five toggles. Give every mark a label carrying the
whole value it sets ("4 stars", never "star 4"), let arrow keys move through the
scale, and provide a way back to nothing for a person who rated by accident.
Hovering the row to preview a score is a nice touch and never the only feedback,
because touch has no hover. A display-only rating should reach a screen reader as
its sentence ("4.2 out of 5") rather than as five graphics.

Watch the neighbouring words. A **review** is the written thing a rating is
usually attached to, and the two are often confused in schemas and URLs. Apple
calls the read-only form a **rating indicator**, which is a useful split: the
indicator reports, the control collects. Ant Design ships the control as
**Rate**. A rating is also not a measurement, however similar the filled bar
looks: a score is an opinion someone recorded, so nothing about it fills up,
runs out, or completes.