Skip to content
Nuvexa AI
Menu
Research · PRACTICAL TUTORIAL

Separate an observation from an interpretation

Visible design is an observation. User preference is an inference that needs behavior, interviews or other appropriate evidence.

Published by Nuvexa AI · Published · 2 min read

Start here

This example uses the sample details below. Replace them with your own facts before applying the instructions.

  • Landing page has three download links; visitor behavior not supplied.

Start with the facts

Landing page has three download links; visitor behavior not supplied. Keep the source, date and scope of each observation. A gap in the evidence is a question to investigate, not a fact to fill in.

Spot the problem

The wording or action to avoid is: “Visitors prefer this tool because it offers three downloads.” Visible design is an observation. User preference is an inference that needs behavior, interviews or other appropriate evidence.

Make the correction

The page presents three download options. Whether visitors prefer them requires separate evidence.

Check before using the result

Mark each sentence as observation or inference and write what evidence would test the inference.

Worked example and reusable input

Use the stated inputs to reproduce checked local outputs. Replace sample facts when applying a brief to your own work.

TASK
Separate an observation from an interpretation

SAMPLE INPUT — replace with your actual facts
Landing page has three download links; visitor behavior not supplied.

AVOID
Visitors prefer this tool because it offers three downloads.

CORRECTED APPROACH
The page presents three download options. Whether visitors prefer them requires separate evidence.

ACCEPTANCE CHECK
Mark each sentence as observation or inference and write what evidence would test the inference.

Key takeaways

  • Visible design is an observation. User preference is an inference that needs behavior, interviews or other appropriate evidence.
  • Mark each sentence as observation or inference and write what evidence would test the inference.