30 Years Guinness World Records Disney Microsoft Warner Bros. Smithsonian Computed on IBM Quantum Hardware 180M+ Solvers
Technical Documentation

Quantum-Assisted Test Construction: Methodology

This page describes, precisely and without embellishment, the computations Advanced Learning Academy executed on IBM Quantum hardware in support of Real World IQ test construction. Every claim here is either verifiable from IBM job records or clearly labeled as a design decision. It is written to survive review by a technical auditor, a journalist, or a federal procurement team.

01

What was computed, on what hardware

Advanced Learning Academy executed quantum computations on IBM Quantum systems accessed through the IBM Quantum Platform using Qiskit Runtime (cloud channel), with hardware-optimized transpilation. Documented production runs include the following verifiable example:

Backend
ibm_torino
Processor Class
133-qubit Heron
Example Job ID
d1lhbkf29o4s73appk4g
Executed
July 6, 2025
Shots
8,192
Framework
Qiskit Runtime, optimization level 3

These were real hardware executions, not simulator runs. Job records are retained and can be independently checked against IBM Quantum Platform records on request. Test forms assembled since July 2025 incorporate the outputs of this work.

02

The three computations

2.1 Question-selection randomization

Test forms draw question orderings and selections from measurement outcomes of quantum circuits rather than from classical pseudo-random number generators. Measuring superposed qubit states yields randomness that is physically generated rather than algorithmically simulated. This is used in test-form assembly.

2.2 Question-set composition analysis

For a candidate question set, question characteristics (difficulty band and question type) are encoded into a quantum circuit of up to 8 qubits. The circuit applies an interference pattern of Hadamard and controlled-phase gates in a quantum-Fourier-transform structure, and is measured over 2,048 shots on IBM hardware. From the measurement distribution we compute the Shannon entropy H, normalized against the maximum possible entropy for the observed outcome space:

composition_score = 1 - (H / Hmax)  // 0 = fully uniform composition; 1 = fully concentrated

A score near zero indicates that the encoded question characteristics are distributed with high uniformity, meaning no single question type or difficulty band dominates the set. Question sets exceeding our concentration threshold are rebalanced and re-analyzed before use. The specific threshold is an internal design decision, applied consistently across forms.

2.3 Sequence-integrity verification

Fixed binary reference sequences used in test-integrity checks were prepared and measured on hardware as 100-qubit circuits to characterize preparation fidelity under real device noise.

03

What this establishes, and what it does not

Institutional credibility depends on claims that match the work. Here is the boundary, stated plainly.

This work establishes

  • The computations described above were genuinely executed on IBM quantum processors, not simulators. Job records exist.
  • Test-form randomization derives from physical quantum measurement.
  • Question-set composition is checked with an entropy-uniformity statistic before forms are released.

It does not establish, and we do not claim

  • That IBM validated, endorsed, or reviewed Real World IQ or any Advanced Learning Academy product. IBM provided computing hardware access; the analysis and its interpretation are entirely ours.
  • That this analysis measures or eliminates demographic bias in the psychometric sense. Adverse-impact analysis under the EEOC Uniform Guidelines is a separate statistical exercise on human outcome data, addressed in our compliance documentation.
  • That the entropy statistic could not be computed classically. The entropy calculation itself is classical; the quantum contribution is the physical generation of the measurement distribution and the physically generated randomness used in form assembly.
04

The claim language we use

Accurate, in use

  • "Test-construction computations executed on IBM Quantum hardware"
  • "Question randomization powered by physical quantum measurement on IBM quantum processors"
  • "Question-set composition analyzed on IBM Quantum hardware using entropy-uniformity methods"
  • Credibility band form: "Computed on IBM Quantum hardware"

Not accurate, never used

  • "IBM Quantum Validated," or any phrasing implying IBM endorsement
  • "Eliminates cultural bias" or "bias-free, verified by IBM Quantum computing"
  • "Impossible to replicate classically"
  • "World's First IBM Quantum Verified IQ Testing"
Why we publish this. Marketing claims about quantum computing are easy to inflate and easy to debunk. Our position is simpler: the work is real, the records exist, and the claims are sized to the evidence. If you are evaluating this assessment for hiring or procurement and want the underlying job records verified, ask.
05

Verification and further reading

IBM Quantum job records for the runs described here are retained by Advanced Learning Academy and can be checked against IBM Quantum Platform records on request: contact team@advancedlearning.academy.

Related reading: What role does IBM Quantum hardware play? and Did IBM validate the assessment? in the Answer Library, the bias and fairness answer, and the federal compliance posture.

Documentation last updated July 2, 2026. Published by Advanced Learning Academy.