Ordinal data
Answers with a known order but unknown spacing, which is what almost every rating question produces.
You know that “Strongly agree” is above “Agree”. You do not know that the step between them is the same size as the step from “Agree” to “Neither”. That is the difference between ordinal and interval data, and it is why averaging rating answers is a convention rather than a theorem.
In practice, means of Likert items are used widely and often defensibly, especially across several items measuring the same thing. The safer habit is to report the distribution as well: the share choosing each option, or the median, tells a reader something a mean can hide, particularly when opinion is split at both ends rather than clustered in the middle.
Explain how you summarized the responses. A mean and the share who agreed describe different aspects of the data, so inspect the full distribution before choosing a summary.
Where it goes wrong
Reporting a mean without its distribution can hide polarization or a concentration at one end of the scale.