How to test written communications using system 1 & system 2 thinking

Communications testing is one of our most popular services. We test lots of written communications from Fact Sheets to Usage Instructions and Packaging Labels, digital and in print. It is part of our “Sense” suite of services. Daniel Kahneman’s ideas about thinking fast and slow - ‘System 1 and System 2’ - help us to design different tests for different written material. 


System 1 thinking is fast, default-based and often optimistic which we need to activate when testing ads and some website behaviour. System 2 is careful and thorough, which we need to activate when testing documents that people need to understand such as instructions like opt-in and opt-out processes and regulatory and disclosure documents.

According to Kahneman, people use System 1 thinking unless something stimulates them to use System 2. In the right circumstances, using System 2 can be a good thing to do because it means we concentrate on what we are doing, but test environments need to be carefully designed to discourage over-thinking.

This is what it means in practice:

1. We design each test individually

Usage instructions work very differently from fact sheets which work very differently from ads.

When we test instructions – such as how to opt in to an online service for example – we need to activate System 1 so that we can identify if consumers use heuristics to complete the task or actually follow the instructions.

When people invest in complex investments or sign contracts of one kind or another, they need to know what they are agreeing to. However, quite often fact sheets and other disclosure documents are written and formatted in a way that encourages System 1 thinking. For these tests we want readers to activate System 2, so we design a test which shows whether readers actively think about the issues, identifying how the layout or the language encourages or deters readers from engaging with the text.

2. We conduct the test in the right place

This is based on where the person will be when they use it in real life. Will be they in transit? Then test in transit. Will be they be at home? Then do the test in home.

3. We choose the right measure

To test instructions, measure what people do.
For informed consent, measure how much people engage with the document.
In some cases, readers and users need to understand the detail and others just the ‘gist’*

4. We time the test appropriately

It is not good research practice to ask consumers to spend 20 minutes forensically diagnosing a TV commercial that takes 30 seconds to watch, but it could be appropriate for an investor to understand a prospectus.


If you are not sure what testing you need, or how testing can help, talk to us! 

*For the technically minded, this comes from a dual-processing theory about memory and cognition called Fuzzy Trace Theory;

To read more thought leadership from Sue Bell subscribe to the Bellbird newsletter

Tags: Linguistics, correspondence, Consumer testing, Written communications

For a quote or a confidential discussion about your research needs