The dominoes test, best known for its D-48 version, is a nonverbal reasoning test in which the person completes series of dominoes following a rule they must discover. It is used as an estimate of general intelligence, or g factor. In Mexico, it has been used in admissions and selection processes for decades, but almost all of its data comes from school samples, so it should be interpreted as just one more cognitive piece of information and not as a definitive judgment about the candidate.
Table of Contents
What is the domino test and where did it come from?
According to the historical review that includes a thesis from the Faculty of Psychology of the UNAM, from 1981, The English psychologist Edgar Anstey created the test in 1944 for the British Army as a parallel alternative to Raven's Progressive Matrices. The same account states that the original version was restricted to official use and that the French adaptation by Pierre Pichot, from 1950, is the one known as D-48.
From there, adaptations emerged in several Spanish-speaking countries. The thesis mentions a standardization in Uruguay, a parallel version prepared in Caracas, and an Argentinian adaptation based on educational levels. That's why today there are editions circulating with different instructions, times, and comparison tables, and this is the first thing you should clarify with whoever is offering you the test.
There is also the D-70, another form of the same test. 1982 study in Psychological Reports He compared them with 321 university students: he found similar means, deviations, and relationships with other criteria, although the correlation between the two forms was relatively low, from .58 to .69. In practice, D-48 and D-70 are treated as parallel forms, but it is not advisable to mix their scores or use the tables of one for the other.
What does the dominoes test measure, and what does it not measure?
The test is based on Spearman's two-factor theory, which distinguishes a general factor common to intellectual tasks and task-specific factors. If you want the complete theoretical context, we explain it in the article on the factor g and what it predicts in selection.
Broadly speaking, each problem requires three things, according to the description of the UNAM thesis:
- To accurately perceive the points on the displayed tiles.
- Discover the principle that organizes the whole.
- Decide which values would fill in the blank card.
The problems range from easiest to most difficult and combine different types of rules. We do not reproduce worksheets, problems, answer keys, or solutions here. They are protected by copyright, and if they are circulated among candidates, the results no longer reflect anything about the individual.
What the test doesn't cover is equally important for HR:
| It contributes | It does not contribute |
|---|---|
| An estimate of abstract reasoning with non-verbal material | Technical knowledge of the position |
| A comparable measure between candidates, if applied equally to all. | Soft skills such as communication, teamwork, and leadership |
| A test that relies little on vocabulary | Values, motivation, or alignment with the company culture |
| This is useful information when the job requires learning quickly or solving new problems. | Experience, previous results, or actual performance |
If you're undecided between dominoes and another nonverbal test, the Raven's test It measures a very similar construct and has versions for different difficulty levels. To compare both with other reasoning options, review what a logical reasoning test.
How is it applied and graded?
It can be applied individually or collectively and is intended for people aged 12 and over, according to the UNAM thesis of 1981. It's a test of power, not speed: the idea is for each person to go as far as their ability allows, and not so much to reward whoever goes the fastest.
Time is where the editions differ most. The thesis recommends allowing a wide margin, with some tolerance, precisely because it is a power test, and in its own study, the test was applied for 30 minutes. Other editions specify different times. The sensible approach is to follow the manual for your edition and maintain the same time and instructions for all candidates in a process.
The grading process, in general terms, follows these steps:
- Each correct answer is worth one point; the card must be completed in both halves.
- The total number of correct answers is the direct score.
- The raw score is converted into a percentile using the edition's comparison table.
- The percentile translates into a range, which is what usually appears in the report.
Step 3 is the most demanding. guidelines of the International Test Commission They urge the use of appropriate comparison groups and advise against drawing conclusions based on norms that are either irrelevant to the individuals being evaluated or outdated. A table from another country, from decades ago, or from students at a different level could place the same candidate in a very different range.
What does the evidence say about its reliability and validity?
There are studies, but almost all of them involve students and use academic criteria. This table summarizes those we were able to find in open sources:
| Study | With who | What he found | What doesn't answer |
|---|---|---|---|
| UNAM, 1981 | 200 applicants to a technical high school in Michoacán | KR-20 reliability of .90 and modest correlations with first and second semester grades | If it predicts job performance |
| Domino, 2001 | 126 Mexican-American and 139 Anglo-American seventh-grade students | Reliable test, without slope or intercept bias between groups, and related to school performance | If those results hold true among adults seeking employment |
| Salimi and colleagues, 2011 | 1,008 high school students in Iran | Test-retest scores of .90 at two weeks, .91 by halves, and .56 with Raven's Matrices | Validity for a specific position |
| Furr and colleagues, 1982 | 321 university students | D-48 and D-70 behave as parallel forms, with a correlation between them of .58 to .69 | Which of the two is suitable for selection? |
Three prudent interpretations of this data:
- The consistency is good in these samples. Coefficients of that magnitude indicate that the test consistently ranks people. This is a necessary condition, but it says nothing about whether the result is relevant to the job.
- The relationship with the Raven is moderate, not total. The Iranian study found .56, and the UNAM thesis cites .55 from the Uruguayan standardization. They measure something similar, but they are not interchangeable.
- Local norms change the reading. In the 1981 thesis, the average score was 19.96 correct answers, and the authors concluded that the test was relatively difficult for that population. Therefore, they calculated their own tables instead of using foreign ones.
Does it predict job performance?
In our review of open sources, we found no study linking the dominoes test to job performance. The evidence that does exist is general in nature regarding cognitive aptitude. the meta-analysis by Sackett and colleagues, from 2022, The study estimates a validity of 0.31 for the cognitive aptitude tests and 0.42 for the structured interview. This is a useful reference point for understanding the weight of such a test, but it's not a domino effect.
The practical consequence is clear: if your company has been using the test for years, your own hiring data can tell you more than any generic table. Cross-referencing the results with the performance of those who already work with you is the most direct way to know if it contributes to your roles.
When is it appropriate to use it in a selection process and when is it not?
This might make sense in these cases:
- Positions where learning new procedures or solving problems without a manual is part of the job: analysis, technical maintenance, quality, planning.
- Processes with many candidates without prior experience, such as internship or technical training programs, where there is little other comparable data.
- Populations with very different reading levels, because the material is non-verbal and depends little on vocabulary.
And it makes even less sense in these cases:
- Positions where technical knowledge or experience is crucial. In these cases, a knowledge test or a work sample is more valuable.
- Roles where the emphasis is on dealing with people, such as sales or customer service, unless the profile also requires that reasoning.
- Any process where it will function as an automatic filter. A low result shouldn't exclude anyone without reviewing the rest of the information.
If your test battery already includes a well-documented reasoning test, adding dominoes rarely adds any new information. Before deciding, review how to choose psychometric tests for work depending on the position.
How to integrate it into a battery of psychometric tests?
These steps will help you use it wisely, whether you're applying it today or considering incorporating it:
- Part of the job profile. Describe the job's reasoning requirements and its requirements in terms of knowledge, soft skills, and values. The cognitive test only covers part of the picture.
- Request the documentation for your edition. Author of the adaptation, year, application time, and origin of the tables. ITC guidelines They recommend avoiding tests with insufficient or unclear technical documentation.
- It applies equally to everyone. Same instructions, same time and same conditions, whether on paper or online.
- Combine sources. Add a structured interview and skills tests, such as a competency tests, so as not to decide with a single measure.
- Check your own data. With the support of a psychometrics specialist, compare the results with the performance of those you've already hired. If there's no correlation, you have grounds to withdraw the application.
- Take care of the consent. Inform the candidate what the test is used for and process their data accordingly. Federal Law on the Protection of Personal Data Held by Private Parties and the applicable regulations. This text is not legal advice; consult your legal department.
The final decision is made by people. A good battery of tests helps ensure that decision is based on comparable data, not just a first impression. If you're looking for options with comparable results across candidates, check out the psychometric tests for Mexico that combine cognitive tests, personality and skills.
Frequently Asked Questions
What does the dominoes test measure?
It assesses general intelligence, or g factor, through reasoning with domino sequences. It is a non-verbal test: the person must discover the rule that organizes each sequence and complete it. It does not measure job-specific knowledge, soft skills, or values, so in the selection process it is used in conjunction with the interview, CV, and other tests.
How long does the dominoes test last?
It depends on the edition. A 1981 thesis from UNAM recommends allowing a generous margin, with some tolerance, and in their own study, they applied it with 30 minutes. The correct approach is to follow the manual for your version and give all candidates the same amount of time and the same instructions, so that the results are comparable.
What is the difference between the D-48 and the D-70?
These are two forms of the same test. A 1982 study of 321 university students found similar means and relationships with other criteria, although the correlation between the two was .58 to .69. They can be used as alternative forms, but each should be interpreted using its own comparison tables.
Where can I find the solved dominoes test or the PDF version?
We don't share it here: the material is usually copyrighted, and knowing the answers invalidates the result. If you're in HR, request the test and its manual from a vendor that documents the edition and origin of its tables. If you're a candidate, it's best to get plenty of rest and follow the application's instructions.
Is the dominoes test useful for hiring?
It can provide useful cognitive data for positions requiring rapid learning or the ability to solve novel problems. However, it's not sufficient for a single decision: its evidence comes primarily from school samples, and we couldn't find any open-label studies with work performance data. Combine it with a structured interview and skills tests, and avoid automatic cut-off points.
If you're reviewing which cognitive tests to include in your processes, at Hirint you can combine cognitive tests, competency tests and soft skills in the same assessment.