Skip to content
The Chat Archive
gpt-3.5-turbo-0613· OpenAI Arena Human Preference 55K

Which correlation calculation is correct for Likert scale answers?

  1. User

    Which correlation calculation is correct for Likert scale answers?

  2. Assistant

    When working with Likert scale data, it is common to use the Spearman's rank correlation coefficient. This is because Likert scale data is typically ordinal, meaning that the responses can be ranked but the intervals between the response options may not be equal. The Spearman's rank correlation coefficient assesses the strength and direction of the relationship between two variables based on their ranks, rather than their actual values.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.