How to measure creativity

August 28, 2019

As a researcher who builds tools to support creativity, I often heard the question “How do you measure creativity?” The answer is not quite simple. “Creativity of what?”, would be my usual follow-up question. Are you interested in measuring the creativity of an idea, a person, or an idea generation session? What is creativity anyway?

To keep it simple, let’s focus on measuring the creativity of an idea. Knowing how to measure creativity on an idea level will help us measure the creativity of a person and an idea generation session (e.g., a creative person/idea generation session produces a lot of creative ideas). The methods I have seen in creativity research are the following:

  1. Ratings on novelty and value. A creative idea should be something new that nobody has thought of before. It also needs to be valuable. With this method, evaluators are asked to rate an idea (or a creative artifact) on absolute novelty and practicality scales. The evaluators are sometimes experts in the field but can also be novices if no special knowledge is required to assess the idea. For example, Yu et al. asked hired crowd workers to rate the originality and practicality of chairs designed for children on 7-point Likert scales.

    Typically, multiple evaluators are asked to rate an idea. If there is enough agreement among the evaluators, you can report the average value of the ratings on each scale.

    In some cases, especially in a controlled experiment, it is also possible to evaluate the novelty (originality) of an idea by counting how often such an idea shows up in a sample of ideas.

  2. Consensual Assessment Technique. Teresa Amabile introduced this technique together with a definition of consensual creativity:

    “A product or response is creative to the extent that appropriate observers independently agree it is creative. Appropriate observers are those familiar with the domain in which the product was created or the response articulated.”

    This method requires multiple evaluators who are domain experts. Each evaluator independently rates the ideas or artifacts relative to one another on multiple dimensions on top of creativity. If the evaluators agree, run a factor analysis to see which objective dimensions have a high correlation with creativity and use that as a measure of creativity itself. As a result, the derived ratings depend less on the subjective opinions of evaluators (e.g., a cat-lover might rate design with cats high even though it is not creative).

  3. Domain-specific approach. Sometimes, you are more concerned about the quality of an idea than its creativity. In this case, you can measure an idea based on specific criteria in the idea’s domain. For example, you can measure the quality of an advertising design by the percentage of people who see the ad and click on it. While these measurements don’t translate directly into creativity, they can send some signals.

Overall, Consensual Assessement Technique is the most rigorous method listed here. However, it requires evaluators with domain expertise and complicated analysis. It also doesn’t scale well when you have to evaluate a lot of ideas. If you do not have access to resources required for the Consensual Assessment Technique, you can use average absolute novelty and value scores. Just make sure that you get high inter-rater reliability before you make any claim about the idea. If you just want a rough estimation, you can use the domain-specific approach. However, you need to keep in mind that the measurement from the domain-specific approach can be noisy.

Choosing an appropriate measurement for you want to learn is considered an art in itself. Measuring creativity can also be quite tricky. Hopefully, this post helps you out a little. :)