Cognitive DistanceA book by Izaias Cavalcanti Free PDF

Free toolkit · From the book

CDI Starter Kit

Measure understanding, not just completion. Usability tests show whether people finished. The Cognitive Distance Index shows whether they understood what they finished.

01What’s inside3 pages

Everything you need to test one flow.

Page 1 of the CDI Starter Kit: the four dimensions, a five-step method and the four score bands.
Page 1 of 3Four things to measureThe four dimensions, a five-step method for one flow, and the score bands.
Page 2 of the CDI Starter Kit: prediction, confidence, agreement and explanation questions with examples from a bank transfer and a medical booking.
Page 2 of 3The session scriptFill-in question templates for before and after the result, with worked examples.
Page 3 of the CDI Starter Kit: scoring formulas, the calibration gap, and four sources of distance with their fixes.
Page 3 of 3Score it, then find the sourceThe formulas, and a table that traces a high score back to its cause.
02Run it on one flow5 steps

Ask before the result, then after.

The predictions only count if people answer them before the outcome appears. Asking afterwards only tests whether they can read the screen.

  1. Pick a flow that ends in an outcomeA payment, a booking, a submission, an AI answer: anything the person has to understand afterwards.
  2. Write the questionsOne or two prediction questions and two or three “why” questions, from the templates on page 2.
  3. Ask the predictions before the resultPause after the person confirms, and ask what they think will happen.
  4. Ask the rest after the resultThree agreement items, “Is the task done?”, the “why” questions, and “In your own words, what just happened?”
  5. Score, diagnose, change, measure againThe sheet gives the score and the band. Page 3 shows where the distance usually comes from.
03Read the score0 to 100

0 is full alignment. 100 is lost.

0–25
Strong alignmentPeople explain outcomes and predict behavior reliably. Monitor.
26–45
Early warningScattered confusion. Review wording and feedback.
46–65
Needs attentionMisunderstanding recurs. Redesign result screens and labels.
66–100
Serious misalignmentPeople succeed but can’t interpret. Restructure.

Watch for the calibration gap

Confident users whose predictions are no better than chance. People who know they’re lost ask for help. People who are wrong and sure leave thinking the task is done.

A working proposal

The bands and the equal weighting have not been validated yet. Use the CDI to compare two versions of the same flow, not to rank unrelated products, and share your results to help test it.

The CDI comes from Chapter 6, Measuring the Invisible, and Chapter 11, A Practical Framework for Evaluating Cognitive Distance. The whole book is free to read.