Test Partnership Tips: The Clock Belongs to the Question Set

Test Partnership is a London publisher whose tests turn up across UK legal, professional services and early-careers hiring. Two things about it are unusual enough to change how you prepare, and neither is the thing candidates normally worry about. The clock does not belong to the question. And the test is deliberately built so that you get about seven in ten right whoever you are.

Work out which test you have actually been sent

Start here, because the naming is genuinely confusing right now and it is Test Partnership's own doing. Its critical thinking test is being rebranded from Concepts to Insights, and as of September 2026 the rebrand is half finished on a single page: the body copy calls it "Our Insights Critical Thinking Test", the factsheet that page links to is titled "Concepts Critical Reasoning", and a named customer testimonial on the same page calls it "Concepts Critical Thinking". If your invitation says either, it is the same test.

Some other signals. An assessment link pointing at tptests.comis Test Partnership's own candidate platform. Customers drop the plural, so an email saying "insight numerical" means the Insights suite. And if the invitation names Net the Numbers, Link Swipe, Pipe Puzzle, Number Racer, Word Logic or Shape Spinner, you have the MindmetriQ games rather than the traditional tests, which are a different preparation problem entirely.

The question-set clock, which is where most marks go

This is the single most useful thing to know about the Insights numerical and verbal tests, and almost nobody practises for it because almost no other publisher does it. The instruction screen says it plainly: "This test comprises five question sets. Each question set has a time limit of three minutes and contains three questions. Therefore in total, you will have 15 minutes to answer 15 questions."

Three minutes covers reading the chart andanswering all three questions about it, spent however you like. The header on their own sample reads "Time remaining for these questions", plural. The verbal test works identically with four statements per passage.

The fix is a reading habit rather than a speed habit:

• Read the stimulus once, for structure, before looking at any question. What are the rows and columns? What are the units, and are they thousands or millions? Is there a footnote? That pass serves all three questions, and doing it three times is where the budget goes.
• Budget roughly 90, 45 and 45 seconds across the three questions rather than 60 each. The first one carries the reading; the other two should not.
• If a question is not falling out in 45 seconds, guess and move. Nothing is deducted, and the time you save protects the next question in the same set rather than some abstract later one.

Concepts works the same way with a twist: the budget varies by item type. 120 seconds for a three-statement deduction block, 90 for a two-statement conclusion block, 60 for a single argument or inference item. Insights Inductive is the exception in the range: 60 seconds per figure sequence, hard, with nothing carried forward.

Seventy per cent is the designed outcome, not a near miss

Every test Test Partnership sells is computer adaptive. Most vendors stop there. Test Partnership publishes the number it targets, and it is not the one a psychometrician would predict.

Always answer

There is no negative marking anywhere in this range. No source mentions it, and the positive evidence is the instruction you are given before the verbal test: "Please try to attempt every question before the end of the time limit. If you cannot answer a question in time, use your best guess." A publisher that deducted marks would not say that.

What makes guessing a poor strategyis not a penalty but the option design. The numerical test gives you ten options, and on a real sample item they run "67.2%", "68.2%", "69.2%" and so on up to "76.2%". You cannot eyeball your way to one of those. But a guess still beats a blank, every time.

Bring a calculator, and stop drilling mental arithmetic

The numerical instruction screen is unambiguous: "You will require a calculator and some rough paper for this test. Please ensure that you have these before starting this test." Candidates who prepare by practising mental arithmetic are optimising for a constraint that does not exist. The constraint that does exist is reading the chart fast enough.

The MindmetriQ games invert this. Their manual states that Net the Numbers and Number Racer require basic calculations in quick succession without a calculator, and that only general addition or subtraction is needed. Different product, opposite advice.

Two tests where all of the above is wrong

The Perceptions situational judgement test and the TPAQ personality questionnaire are both untimed, and Test Partnership is emphatic about it in three separate places for the SJT: the website badge reads "Untimed", the factsheet lists "No time limits" as a feature, and the sample sets a per-scenario limit of 30,000 seconds, which is its way of saying there is not one.

So the pacing habits the rest of the range trains are actively unhelpful there. Two specific things about the SJT are worth knowing, because they are unusual:

• Six points, and no neutral one. Very ineffective, Ineffective, Fairly ineffective, Fairly effective, Effective, Very effective. There is nowhere to hedge to, so declining to commit is itself a choice and it loses marks on the actions that are strongly keyed.
• Credit spans two adjacent labels.Test Partnership's own sample data keys each action to a band such as "Fairly effective" to "Effective", not to one exact point. You do not have to match an expert exactly. But if you compress everything into the middle, you can rank all four actions in exactly the right order and still miss every band.

If you are sitting the MindmetriQ games

Six games, and employers pick from them: one per cognitive area for a quick screen, four or five for targeted selection, all six plus personality for a full profile. They run per-item clocks of 15 to 30 seconds, adapt after every item, and are genuinely designed for a phone, unlike the traditional tests.

The most useful single tip is for Net the Numbers, and it comes from the vendor's own instruction: "Remember the highest number may be negative." The grid contains negatives, so the square with the biggest-looking numbers in it is frequently not the best one. Scan for minus signs before you add anything. On Pipe Puzzle, trace the whole route with your eyes before rotating a single segment: turning things to see what happens is how the clock disappears.

Where to practise the awkward parts

The shared question-set clock is the thing worth rehearsing, because it is the one habit that does not transfer from any other provider. Our Test Partnership practice tests implement it exactly, including failing the rest of a set when its budget expires, along with the 70% adaptive targeting and the difficulty-weighted scoring. If your invitation turns out to name Logiks or PAPI rather than Insights or TPAQ, that is a different publisher: see Cubiks. And if you are not sure at all, the provider lookup guide turns a product name into a vendor.