The Watson-Glaser Critical Thinking Appraisal, published by Pearson TalentLens, has been the standard critical thinking test in UK legal recruitment for years.
But the test was designed long before candidates had AI tools in their pocket. And many law firms are now having to re-test their candidates in person during later stages in order to verify their results. If you're a law firm reviewing your critical thinking assessment, there is another test which may be more suitable.
Signals Critical Thinking
Signals Critical Thinking is our new critical thinking assessment, developed specifically for legal hiring and designed from the ground up to be more resilient to AI than traditional critical thinking tests. It measures the same core skills as the Watson-Glaser, analysing information, evaluating arguments and drawing conclusions. What's different is how the assessment is built.
Candidates read a short passage set in a realistic legal scenario, with dedicated reading time before the questions begin. Four questions then appear one at a time, each with its own short timer, and taken from a large item bank. The passage stays on screen throughout, but candidates can't preview the next question or go back to change an earlier answer. Once a set of four is complete, the test adapts, selecting the next passage from a growing bank to suit the candidate's estimated ability.
The key to its AI resilience is pace. Signals moves quickly, and each question's time limit is short, leaving very little time to copy it into an AI tool, wait for a response and enter the answer. A candidate who tries is likely to run out of time, while a candidate thinking for themselves has the time they need.
That pace is only possible because of how the test is designed. You couldn't just impose short time limits on a traditional critical thinking test: its questions are long, and candidates need to read a passage and its statements together before they can answer, so cutting the time would only end up measuring reading speed. Signals separates the two. Candidates absorb the passage first, with dedicated time to read, remember and process it, and then answer streamlined questions that each take moments to read. The thinking happens up front, which makes short answer windows fair to candidates while leaving far less room for AI.
To be clear, no online test is completely AI-proof, and Signals isn't either. A determined candidate may still find ways to get some outside help. But it offers considerably more protection than a traditional critical thinking test, where there's ample time to consult AI on every question. That means scores are far less likely to be inflated by AI, and far less likely to creep upwards year after year as AI use spreads. The research behind this approach is set out in our AI resilience whitepaper.
Test Partnership has helped us create a more consistent and objective hiring process. The assessments support fairer decision-making and have been a valuable part of our broader approach to inclusive recruitment.
Watson-Glaser vs Test Partnership's Signals Critical Thinking
Resistance to AI
This is the biggest difference between the two. The Watson-Glaser's format, written content under one overall time limit, was never designed with AI in mind, so protecting it relies on monitoring candidates or trusting them. Signals is designed so that the format itself does the work. It isn't immune to AI, but scores are much more likely to reflect the candidate's own reasoning, and much less prone to the year-on-year inflation firms are seeing with traditional tests, so you can shortlist with more confidence.
Relevance to legal work
The Watson-Glaser uses general passages, while every Signals passage is set in a realistic legal context. No legal knowledge is needed, as every answer is based only on the information provided, so it remains fair to non-law graduates and career changers. But candidates feel like the test is relevant to the role, which makes for a more credible experience and a better reflection of the firm.
Candidate experience
Signals is built mobile-first, so candidates can complete it on their phone or computer, and it takes around 15 minutes, compared with 30 minutes for the Watson-Glaser. That's possible without losing accuracy because of the adaptive-difficulty technology and the shorter time limits per question. Candidates don't spend time working through items that are far too easy or far too hard, so you get a precise result in half the time.
Running it in-house
The Watson-Glaser requires you to pay for BPS qualification training to use and read the results. Test Parntership's critical thinking tests require no training. Scores are reported as percentiles or STENs against your chosen comparison group, results can be viewed and compared in the platform, downloaded as PDF reports or exported to CSV, and our business psychologists are on hand to advise on norm groups, cut-scores and reasonable adjustments.
Measurement quality
Both tests convert performance into scores relative to a comparison group. Signals uses Rasch calibration and computer adaptive testing to select questions suited to each candidate's ability, so it can measure precisely without needing a long fixed test, and item analysis is used to review question quality and differences in performance across groups.
The candidate preparation hub has been a really valuable part of the process. It helps candidates feel better informed and better prepared, which creates a more positive overall experience.
Conclusion and next steps
The Watson-Glaser earned its place in legal recruitment, but it was built for a world without AI. Its text-based format and single time limit leave it exposed to AI-assisted answers, and its qualification requirement and general passages add friction for firms and candidates alike.
Signals Critical Thinking measures the same skills, in legal scenarios candidates recognise as relevant, with AI resilience built into the design rather than bolted on. No online test is fully AI-proof, but for training contract and vacation scheme campaigns, Signals gives you much stronger protection against AI-inflated scores than a traditional test, and a better experience for the candidates you're trying to attract. We work with law firms including Addleshaw Goddard, Eversheds Sutherland, Taylor Wessing, DAC Beachcroft and Fladgate.
You can try a Signals practice testlet, explore Signals Critical Thinking and our assessments for law firms, or book a call with our team to find out more.
Sources: TalentLens Watson-Glaser, TalentLens BPS Level A training, Test Partnership AI resilience whitepaper.
Joshua Hancock
Digital Marketing Manager at Test Partnership. Over seven years' experience as a writer, content strategist, SEO and digital marketer.
