Rating scales

There are several popular scales used in questionnaire design including Likert and semantic differential scales.

In this article:

General guidelines

Google Surveys lets you choose from 5, 7, 10, or 11 stars. There has been a long-lasting debate between 5-pt vs. 7-pt and odd vs. even (see presentation). As our surveys are easy to copy and easy to deploy, a good practice is to run the same question with a couple alternative scales. This lets you validate and cross-tabulate the results. It also lets you test a couple dimensions at the same time (e.g., both affective and cognitive responses to content).

 

Behavioral intent

Use this scale to measure a user’s stated inclination to engage in a certain behavior (e.g., the likelihood of purchasing a product, paying attention to an ad).

extremely_unlikely_extremely_likely

The above scale is one of the most popular, but there are alternatives. It is good practice to include 2-3 alternates in your survey with the same question wording. This lets you validate your data and distributions. Sometimes the alternates represent some nuance to the dimension under investigation. Using alternates will let you aggregate scores into an index.

For behavioral intention, alternatives include:

unlikely_likely

impossible_possible

Credibility (trustworthiness)

This scale measures source credibility and is popular in advertising research. Examples of use include credibility of a nutrition claim in an ad, trustworthiness of a company, or a website’s reputation.

notcredible_credible

Alternatives include:

nottrustworthy_trustworthy

notbelievable_believable

Frequency of use

There are a couple options for measuring how often a user reports on using a product, feature, or technology. A popular scale in marketing literature is:

frequency_use_of_internet_never_veryfrequent

Alternates include:

notatall_veryfrequent

never_always

Agreement

This scale assesses a person’s general agreement with a statement you provide.

stronglydisagree_stronglyagree

Concern

You might want to gauge a user’s concern with privacy or some other policy (e.g., how do you feel about your privacy with regard to XXX?). This scale gauges the user's concern.

notatallconcerned_extremelyconcerned

Importance

This scale gauges the importance.

unimportant_veryimportant

 

Was this helpful?
How can we improve it?