Cappfinity situational judgement and strengths
Typically 15-20 scenarios with multiple sub-parts · Untimed (60-90 minutes average)
What it tests. Behavioural alignment with the Agile Mindset framework: Commercial Judgement, Determination, Excellence, Imagination, Integrity and Teamwork.
Worked example. A senior associate needs a 100-page due diligence review by 6pm; at 4:30pm a partner needs urgent research for a call in 30 minutes. The optimal action accepts the partner's task and emails the associate immediately with a progress update and a realistic new deadline (ranked above asking another trainee, above silently missing the deadline, with multi-tasking both under time pressure ranked least effective).
Common traps. The superhuman-trainee delusion (ranking all-nighters or simultaneous completion highest) and second-guessing what a perfect lawyer would say, which creates inconsistency the algorithm flags.
How to handle it. Lean toward solutions that protect the client, keep your supervisor informed and own mistakes. Because it is untimed, read every scenario twice.
Watson Glaser: Inference
7-8 questions · Part of the 30-minute block
What it tests. Distinguishing a conclusion that must be true from a merely plausible guess.
Worked example. Given a passage stating many companies are now legally required to use a double bottom line, the statement 'companies are entirely free to choose' is False. A statement about its effect on profit margins, on which the text gives no data, is Insufficient Data.
Common traps. Using real-world knowledge, and confusing Probably True with True. An inference is only True if it is impossible for it to be false based on the text.
How to handle it. Apply the four-walls rule: the passage is the entire universe. If a claim cannot be proven explicitly inside it, it cannot be definitively True or False.
Watson Glaser: Recognition of Assumptions
7-8 questions · Part of the 30-minute block
What it tests. Identifying unstated premises a speaker takes for granted.
Worked example. For 'we must implement AI document review so trainees can focus on high-value work', the assumption that the software accurately identifies relevant clauses is Made; the assumption that trainees enjoy manual review is Not Made.
Common traps. Selecting 'Made' simply because you agree with the overall point. Separate the logical architecture of the sentence from its sentiment.
How to handle it. Use the negation test: insert 'not' into the assumption. If that breaks the original argument, the assumption is Made.
Watson Glaser: Deduction
7-8 questions · Part of the 30-minute block
What it tests. Syllogistic reasoning: whether a conclusion must follow.
Worked example. All Magic Circle firms have London offices; some London-office firms run agile working. 'All Magic Circle firms run agile working' Does Not Follow. 'Some entities with London offices are Magic Circle firms' Follows.
Common traps. The transitive fallacy (if A relates to B and B to C, then A to C) and letting real-world knowledge overwrite a sound syllogism.
How to handle it. Visualise the premises as overlapping circles. If you can draw a case where the conclusion is false while the premises hold, it Does Not Follow.
Watson Glaser: Interpretation
7-8 questions · Part of the 30-minute block
What it tests. Deciding if a conclusion is justified beyond reasonable doubt by the data.
Worked example. Data shows mentorship programmes cut attrition 18% and raised billable hours 12%. 'Mentorship improves retention and productivity' Follows. 'Associates leave primarily because they feel unsupported by partners' Does Not Follow, as the text gives no data on why they left.
Common traps. Over-generalisation: extending a specific trend to a broader population or inferring causation from correlation.
How to handle it. Be suspicious of conclusions using absolute words (always, never, all) when the source uses qualifiers (often, some, tended to, many).
Watson Glaser: Evaluation of Arguments
7-8 questions · Part of the 30-minute block
What it tests. Distinguishing a directly relevant, important argument from a weak, emotional or tangential one.
Worked example. On mandatory sustainability reporting, an argument that transparent data lets investors direct capital to sustainable businesses is Strong; an argument that owners dislike paperwork and find it stressful is Weak.
Common traps. Labelling an argument Strong because you agree with its viewpoint, even when it is logically shallow.
How to handle it. Apply the relevance test: if the argument is true, does it actually solve or significantly affect the core issue? If it addresses only a minor byproduct or personal preference, it is Weak.