TestGorilla Assessment 2026: Tests and Timers

TestGorilla is a skills-testing platform that employers use to screen applicants with short online tests, typically 10 to 13 minutes each, chosen from a library of several hundred. A typical TestGorilla assessment bundles three to five tests, for example critical thinking, numerical reasoning and a data or role-specific test, and each test opens with a few practice questions under their own short countdown. This guide shows what the tests look like from the inside, based on a full three-test battery observed in a 2026 application round, and gives you original practice questions with worked answers.

Key facts

What TestGorilla is

TestGorilla sells pre-employment tests to employers. The employer picks tests from the library, adds optional custom questions (video, essay, multiple choice) and sends you a link. You never "apply on TestGorilla" in the usual sense: the employer owns the process and decides how your scores are used. That is why prep sites and forums report very different experiences. A small company may send a typing test and a personality questionnaire; a graduate programme may send a demanding reasoning battery.

TestGorilla's own library page for the numerical reasoning test describes four question families: numerical computation, numerical estimation, numerical reasoning, and data interpretation from "line graphs, charts, tables, and raw numbers". The critical thinking test is described as covering deductive and inductive reasoning.

What it looks like: a three-test battery

The battery seen in autumn 2026 carried the employer's logo top left and a "POWERED BY TestGorilla" footer. A chat bubble sat bottom right and an accessibility button bottom left. The layout was the same in all three tests: the question and any instruction on the left, the answer options on the right, and tables or charts underneath the question.

Practice items. Each test began with four practice questions under a blue banner ("This is the 1st practice question." … "This is the 4th and last practice question."). The practice block ran on its own countdown of about five minutes. The button on the fourth practice item read "Start test >". Practice items look exactly like the real ones, and in one case a practice chart came back in the scored section with a different question. Treat the practice block as a free warm-up, but do not spend all five minutes there.

Timer and progress. TestGorilla's candidate help centre says a time and progress bar runs along the top and that there are points between tests where you can take a break while the timer pauses. Each scored test had one countdown for the whole test (12:00 for critical thinking, 10:00 each for numerical and data literacy), a counter ("1/16", "1/12") and a progress bar. There was no per-item timer. There was no back button.

Charts and zoom. Charts had a magnifier icon that opened a zoom view, closed with a pink ✕. The zoom view covered the answer options, so you have to close it before you click. Several numerical charts had no value labels at all, so you read bar heights against gridlines.

Pace. The critical thinking test averaged about 45 seconds per item and felt comfortable. The data literacy test was the tightest: little more than a minute remained at the last item.

Question types with original examples

Critical thinking

Six formats appeared, each with its own fixed instruction: syllogisms (pick the conclusion that follows), two-conclusion items (four fixed options: C1 only, C2 only, both, neither), statement relationships (contradicts / supports / restates / is irrelevant to), assumptions (what the passage depends on, or what most strongly invalidates it), strengthen / weaken / has no effect, and rule application (the application correctly or incorrectly follows the rule, and the result is expected or unexpected).

Practice 1 (two conclusions). Premise 1: If the payment server is overloaded, the checkout page loads slowly. Premise 2: The checkout page is loading slowly. Conclusion 1: The payment server is overloaded. Conclusion 2: Customers are abandoning their baskets. Options: C1 follows, C2 does not / C2 follows, C1 does not / Both follow / Neither follows.

Answer: Neither follows. C1 affirms the consequent: a slow page could have other causes, and the premises only say what happens if the server is overloaded. C2 is plausible in real life but nothing in the premises mentions customers.

Practice 2 (relationship). Statement 1: "A good onboarding plan gives new joiners early access to tools, a named contact for questions, and small tasks they can finish in their first week." Statement 2: "New joiners settle in faster when someone is assigned to answer their questions and they can complete real work early on." Options: contradicts / supports / restates / is irrelevant to.

Answer: Supports. Statement 2 adds a claimed outcome (settling in faster) for two elements of Statement 1, which makes Statement 1's plan more credible. It does not merely say the same thing again (that would be "restates"), because Statement 1 never mentions speed of settling in.

Practice 3 (rule application). Rule: "Purchases above $2,000 need written approval from a budget holder before the order is placed. Orders placed without approval are cancelled by Finance." Application: "An analyst placed a $2,600 order, then asked the budget holder for written approval, which was given the next day. Finance did not cancel the order." Options: correctly / incorrectly follows the rule × the result is expected / unexpected.

Answer: Incorrectly follows, unexpected. Approval came after the order, which breaks the rule. The rule predicts cancellation, so "not cancelled" is unexpected under the test's definitions.

Numerical reasoning

The numerical test mixed quick arithmetic with chart work: number series, work-rate problems, reverse percentages, compounding, ratios, averages from a table, and bar or clustered-bar charts with or without labels. Four options per item.

Practice 4. After a 15% price increase, a software licence costs $69 per seat. What did it cost before the increase? $54.00 / $58.65 / $60.00 / $79.35

Answer: $60.00. Divide by 1.15: 69 ÷ 1.15 = 60. The trap $58.65 is 69 × 0.85, which takes 15% off the new price instead of reversing the increase.

Practice 5. Twelve workers, each working 8 hours a day, assemble 48 units in 6 days. How many days do 8 workers, each working 6 hours a day, need to assemble 24 units at the same rate? 3 days / 4.5 days / 6 days / 9 days

Answer: 6 days. Total effort: 12 × 8 × 6 = 576 worker-hours for 48 units, so 12 worker-hours per unit. 24 units need 288 worker-hours. The new team supplies 8 × 6 = 48 worker-hours per day, so 288 ÷ 48 = 6.

Practice 6. What is the next number? 2, 6, 12, 20, 30, ? 36 / 40 / 42 / 48

Answer: 42. The gaps are 4, 6, 8, 10, so the next gap is 12 (equivalently, n × (n + 1): 6 × 7 = 42).

Practice 7. A starting salary of $64,000 rises by 10% after year one and by a further 5% after year two. What is the salary now? $73,600 / $73,920 / $74,240 / $75,000

Answer: $73,920. Raises compound: 64,000 × 1.10 × 1.05 = 73,920. Adding the percentages (15%) gives the trap $73,600.

Data literacy (working with data)

This test is the least "IQ-like". It asks practical analyst questions: which column needs cleaning, which data type stores a value, which chart suits a dataset, which formula answers a business question, and which statements a chart or table actually supports. TestGorilla's library lists a 10-minute "Working With Data" test covering handling data correctly and basic analysis; whether the battery observed used exactly that test is not confirmed.

Practice 8 (select all that apply). Sales by channel: Online $1,240,000; Retail $2,180,000; Wholesale $960,000; Partners $420,000; Total $4,800,000. Which conclusions can be drawn? (a) Online accounts for more than a quarter of total sales. (b) Retail and wholesale together account for more than two-thirds of sales. (c) The partner channel has the fewest customers. (d) Retail sales are more than five times partner sales. (e) Online sales will overtake retail sales next year.

Answer: (a) and (d). Online is 1,240 ÷ 4,800 = 25.8%. Retail plus wholesale is 3,140 ÷ 4,800 = 65.4%, under two-thirds. Retail ÷ partners = 2,180 ÷ 420 = 5.19. Customer counts (c) and next year (e) are not in the data.

Practice 9 (formula choice). A table lists exchange rates as "national currency per US$": Mexican peso 18.00. Which calculation gives the US-dollar value of 900 pesos? 900 × 18 / 900 ÷ 18 / 18 ÷ 900 / (1 × 18) × 900

Answer: 900 ÷ 18 = $50. The rate tells you how many pesos buy one dollar, so you divide pesos by the rate. Multiplying is the direction trap.

Practice 10 (truncated axis). A column chart of monthly orders has a y-axis running from 95,000 to 101,000. January's bar reaches 100,200 and December's 97,700, so December's bar looks about half as tall. Which summary is appropriate? Orders roughly halved over the year / Orders fell by about 2.5% / Orders rose slightly / Orders did not change.

Answer: Orders fell by about 2.5%. 97,700 ÷ 100,200 = 0.975. The axis does not start at zero, so bar heights exaggerate the drop.

Traps and how to avoid them

• Charts without value labels. Read the gridline step first (for example 5 or 2,500), then estimate. Open the zoom only when two bars are within one gridline of each other.
• Ratio order. "Items above X compared to items between Y and Z" means X : Y–Z, not the reverse. Both orders are usually offered.
• Off-by-one in series. "What is the eighth number?" counts the first term as number one. Write the term numbers down.
• Reverse percentages and compounding. A price after a rise is divided by (1 + r). Two successive raises multiply.
• "Restates" vs "supports". Restates means the same claim in other words. Supports means new information that makes the claim more likely.
• Normative or real-world conclusions. In deduction items a conclusion with "should" or a plausible consequence the premises never mention does not follow.
• Data you are not shown. In "select all that apply" items, statements about profit, customers, items sold or future months are distractors when the table only shows revenue for past months. The number of correct statements is not given.
• Zoom covers the options. Close it before you answer, or you lose seconds hunting for the radio buttons.

Monitoring and anti-cheating features

TestGorilla states that its tests can include regular webcam snapshots (taken with your permission), full-screen mode detection, IP address monitoring, randomised questions from a large question pool, and a disabled copy/paste function. Which of these are switched on is the employer's choice, so read the invitation and the start screen: they tell you whether the camera will be used. In practice this means: take the test in one sitting, in a quiet room, with the browser in full screen and no other windows open, and follow whatever rules the start screen sets.

How to prepare

One week out (3–4 hours in total). 1. Two sessions of critical thinking (40 minutes each): learn the six formats and their fixed definitions until you can classify an item in five seconds. Practise conditional logic (modus ponens, modus tollens, and the two fallacies). 2. Two sessions of numerical (40 minutes each): number series, work-rate, reverse percentages, compounding and ratio order. Use a basic calculator only. 3. One session of data literacy (30 minutes): data types (string, integer, float, Boolean, date), cleaning checks (inconsistent categories, mixed date formats, missing values), chart choice (bar for many categories, line for time, scatter for two numeric variables, bubble for three), truncated axes, and per-capita normalisation.

Time targets. Critical thinking about 45 seconds per item (16 items in 12 minutes). Numerical and data literacy about 50 seconds per item (12 items in 10 minutes), with chart items taking longer. Bank time on the quick arithmetic items.

Setup. A laptop or desktop with a large screen (zoom views and small chart labels are hard on a phone), a calculator, pen and paper, a stable connection and a working webcam if the invitation asks for one.

Sources

• TestGorilla, "Talent assessment tools: anti-cheating" blog
• TestGorilla test library, cognitive ability tests (names and durations)
• TestGorilla test library, Numerical Reasoning test (topics, languages, 10 min)
• TestGorilla test library, Working With Data test
• TestGorilla candidate help centre

Scritto da

Adrian, Specialista

MSc Psicologia, 7 anni di test fatti.

Ultimo aggiornamento: Oct 1, 2026