Skip to main content Link Menu Expand (external link) Copy Copied

Software Engineering and “The other CRT”

This is a brief summary of Cognitive Reflection in Software Verification and Testing, presented at the Software Engineering Education and Training (SEET 2023) track of the IEEE International Conference on Software Engineering in Melbourne, Australia and published in the proceedings.

This presentation page is available at and its source is available on GitHub.

Getting Started

First, let’s begin with a few questions in a poll

QR Code for

Behavioral Economics

The Cognitive Reflection Test (Frederick 2005) was motivated by Tversky and Kahneman’s research that lead to a Nobel Prize in Economics. They identified how consumers do not act in their rational best interests because of cognitive biases.

Daniel Kahneman, Nobel Prize winner and best-selling author

Photo courtesy of

In Kahneman’s best-selling book Thinking, Fast and Slow he described our brain’s two systems of decision-making:

  • System 1 is fast and reflexive and particularly susceptible to our emotions and intuitions
  • System 2 is slower, deliberate, and utilizes conscious calculations and reasoning

The Cognitive Reflection Test is a proxy for how likely an individual is to inhibit System 1 and demonstrate Cognitive Reflection (using System 2). It is inversely correlated with:

  • Gambler’s fallacy
  • Sunk cost fallacy

In what ways might “Cognitive Reflection” play a role in software engineering?

Inspecting Code

Does this function produce the described behavior?

isPalindrome function

Does this function produce the described behavior?

sortDescending function


Test outcomes vs. Implementation

Fail Pass
Working 🤨 🙌
Buggy 🧐 🪳

Test accuracy


CRT vs Inspection

  • CRT was not a significant predictor (p=0.329) of students’ affirmation of acceptable code.
  • CRT was a significant predictor of students rejecting defective code (p<0.0001) with the log odds of correctly rejecting the defective code increasing by 2.94 (95% CI 1.56-4.50).
    • When considering only students who proposed cases that caused a defect, we found that CRT was a significant predictor (p<0.001) of students identifying a defective case with the log odds of doing so increasing by 2.37 (95% CI 1.05-3.86)

CRT vs Testing

  • CRT was not correlated with test accuracy (ρ=0.940, p=0.008).
    • Test effectiveness (M=0.66, sd=0.27) was not correlated with CRT (ρ=0.140, p=0.159)
    • Test affirmation (M=0.76, sd=0.17) was not correlated with CRT (ρ= -0.199, p=0.045)


How would you interpret the results?

Full Paper

The full paper will be available soon.