A skills test measures what a candidate currently knows about a specific area of the job: Excel, programming, sales, digital marketing, or how a store operates. In recruitment, it's useful when that expertise is needed from day one, and it works if it's designed based on real-world tasks, scored using pre-established criteria, and administered with measures to prevent cheating. Like any test, it complements the CV and the interview; the final decision rests with a person.
Table of Contents
What is a knowledge test in selection?
The U.S. Office of Personnel Management (OPM) It describes job knowledge tests as questions designed to assess technical or professional expertise in specific areas, and clarifies that they measure what the person knows at the time of the test, without attempting to estimate their learning potential. That is the difference with a aptitude testOne asks what you know, the other what you could learn.
Common examples in Mexican companies:
- Office automation: Excel, Word or PowerPoint for administrative and analytical positions.
- Techniques: SQL, Python, or project management.
- Commercials: sales process, prospecting, handling objections; Google Ads, social media or SEO in marketing.
- Sectoral: inventories, visual merchandising, supply chain or regulatory compliance in retail.
- Regulations or internal processes: What an employee needs to know to operate without constant supervision.
The first four areas have ready-made tests in Hirint's library, which includes more than 70 tests; what is specific to each company, such as its internal processes, is covered with customized tests.
When is it appropriate to use it and when is it not?
According to the same OPM guide, These tests are suitable when the candidate needs to arrive with knowledge that is acquired over a considerable amount of time, and are not appropriate when the company is going to provide training after hiring in those same areas.
| Situation | Knowledge test? | Best alternative or complement |
|---|---|---|
| The position requires mastering a tool from day one | Yeah | Short work sample |
| The company trains all new hires on the tool | No | Aptitude test and interview |
| Specialized technical position (programming, data analysis, regulations) | Yeah | Technical interview with a real case |
| Entry-level positions with no experience required | Cautiously, only the essentials | Skills, soft skills and interview |
| Knowledge is demonstrated by a mandatory license or certification. | It has already been evaluated. | Verify the credential |
What types of questions can you use?
The most common format is multiple choice, although written answers and fill-in-the-blank questions are also used. Each format is best suited to measure one thing:
| Format | Measure better | How to score | Main risk |
|---|---|---|---|
| Multiple choice | Mastery of concepts and rules | Answer key | Either memorize it or look up the answer |
| True or false | Basic concepts | Answer key | To guess by chance |
| Case study | Application to a job situation | Rubric with criteria | That two evaluators give different ratings |
| Open question | Reasoning and argumentation | Rubric with levels | Correction time |
| Exercise or file submission | Actual execution (formula, query, report) | Checklist | Let someone else solve it |
A good test usually mixes closed questions to cover a lot of content in a short time with one or two cases that show how it is applied.
How to design a knowledge test step by step?
- Part of the job duties. The OPM indicates that these tests are usually built from an analysis of the job tasks. If you already have one job profile Updated, use it as a base.
- Create a table of specifications. List the topics, the weight given to each, and the expected level (basic, intermediate, advanced). This way, the test won't be filled with questions about things that are easy to ask.
- Create a question bank larger than the test. Write several questions per topic and level so you can rotate them among candidates and update any that become ineffective. If you're creating it from scratch, the guide for create assessment tests for candidates Go through the entire process.
- Calibrate the level. Ask two or three people who do the job well to answer it: if they get a question wrong, review it; if they all get it right without thinking, maybe it doesn't discriminate.
- Set the time. It should be long enough for someone who is familiar with the subject to finish without rushing. Too short a time turns the test into a speed test.
- Check the wording. Only one correct answer, with no double negatives or clues in the options, and using the vocabulary used in the company.
- Update it. Tools, rules, and processes change; schedule a periodic bank review.
How to score it and set a cut-off point?
- Closed questions: Answer key and one point for each correct answer. If there are topics with more weight, weight them according to the specifications table.
- Open questions and cases: Write a rubric before applying it, with clearly defined criteria and levels. If two people can grade separately, compare their notes and clarify any discrepancies in the rubric.
- Results by topic, not just total: A candidate may master formulas but struggle with pivot tables, and that guides the training if they are hired.
- Reasoned cut-off point: Gather those who know the position and ask them what a minimally competent person should answer. Avoid setting the cutoff based on habit or the group average.
- Analysis after the first application: Check which questions almost everyone gets right and which no one gets right, and whether those who score higher overall get them right more often.
How to prevent candidates from copying?
Online, the challenge is that the person responding is the candidate and not someone searching for the answers. International Test Commission guidelines for computer and internet testing They distinguish four application modes: open, controlled, supervised, and managed, depending on the level of human supervision. They also recommend question banks large enough to create parallel forms and generate different tests for each candidate.
Practical measures:
- Rotate questions and options from the bank, so that two candidates do not see the same test.
- Ask about application, not definition. A case study of the position is more difficult to find than a textbook definition.
- Use time per question when the content is easy to access.
- Verify identity and advise of rules before starting, with consent and data processing according to applicable regulations.
- Confirmed in the interview. Asking them to explain how they solved a test case detects answers that are not their own.
At Hirint, tests can be applied with proctoringThe platform verifies identity using a series of photographs of the person and their ID, and detects if the candidate switches windows or tabs while answering. This allows the company to know who took the test and if they left it to search for answers, without having to monitor them in real time.
What does the evidence say about its validity?
The OPM It summarizes several advantages: high content validity, correlation with performance, added value over cognitive ability tests, and a high perception of fairness among candidates. It adds that custom-made tests have shown somewhat higher validity than standardized tests, although they are more expensive to develop and maintain, and can produce larger differences between groups than other predictors, thus requiring a review of the results.
The meta-analysis of Sackett, Zhang, Berry and Lievens (2022) It estimates a validity of 0.40 for job knowledge tests, higher than for cognitive ability tests (0.31). Practical reading, well-designed and linked to the job, is one of the most useful indicators, provided it is not the only one.
In Hirint, for example, the library has more than 70 tests, and many are ready-to-use knowledge tests: Excel at basic, intermediate, and advanced levels (15 minutes each) or in an interactive format on Excel Online (20 minutes); Word (10 minutes), PowerPoint (15 minutes), SQL and Python (20 minutes each); project management (10 minutes); five sales tests and three digital marketing tests (15 minutes each), and a series of retail tests ranging from general knowledge to inventory, visual merchandising, and supply chain. When the position requires something specific, the custom tests They allow you to create open-ended, multiple-choice, single-choice, true or false questions, with document uploads, images, video, or open-ended questions analyzed with AI, with a report of the result of each answer.
If you want to organize which skills to measure in your job openings and how to combine them with others Psychometric tests for companies in Mexico, You can check it with the Hirint team.
Frequently Asked Questions
What is a subject-specific knowledge test?
This type of assessment evaluates a specific area of the job, not general knowledge. Examples include: Excel for an analyst, SQL for a data profile, Google Ads for a digital marketing profile, closing techniques for a salesperson, or inventory management for a store manager. The more closely it relates to real-world tasks, the more useful the results.
How is it different from an aptitude test?
A knowledge test measures what a person already knows; an aptitude test assesses their potential to learn a particular task. If you're going to train the new employee on the tool, aptitude is more important; if they need to use it from day one, knowledge is more important.
What happens if a candidate does not approve it?
It depends on how their role was defined. If knowledge is essential from day one, a low score could disqualify them; this should be explained in the job posting. If knowledge is desirable, the score for each subject helps assess how much training they would need.
Is an online knowledge test valid?
It can be effective if you control who answers and how. Rotating question banks, applied case studies, time limits per question, identity verification, and interview checks all help. Without these measures, the results are less informative than they appear.