Reading a Multicheck result: what counts as a good score

The certificate arrives as a PDF, two pages, a lot of numbers. And somewhere on it stands a 62, and nobody tells you whether that is good. Most people read the number either as a school mark or as the percentage of tasks they got right. Both readings are wrong, and both fail in the same direction: an average result looks poor, or a weak one looks sufficient.

This article explains what is on the two pages of the certificate, how the numbers come about, and what the question of a good score actually turns on. Kognita does not sell preparation courses. Nothing here is written to make your result look worse than it is.

First, what the certificate is for

If you have moved to Switzerland recently, the Multicheck will have arrived in your life without much explanation. It is a commercial aptitude test run by gateway.one and taken at one of over 35 test centres. Families book and pay for it themselves. Depending on the version it costs CHF 60 for Attest, CHF 85 for Detailhandel und Service, or CHF 100 for the other six versions, and it takes between 1.5 and 3.5 hours.

It is taken in the 8. Klasse, the second-to-last year of compulsory lower secondary school (Sekundarstufe I), at roughly age 14. That year is when everything happens, because this is when applications for an apprenticeship starting the following August are written. A fourteen-year-old in Switzerland is not choosing options for later. They are applying for a job.

The certificate exists to support that application, and to understand why it matters you need to know what a Lehre is. An apprenticeship in Switzerland is not a consolation route for pupils who could not manage anything else. It is the main way out of compulsory school: a three- or four-year apprenticeship leading to the EFZ, or a two-year one leading to the EBA, combining paid work in a Lehrbetrieb, the company that trains the apprentice, with school. It leads on to the Berufsmaturität, a vocational baccalaureate, and from there to the Swiss universities of applied sciences; with the supplementary Passerelle exam it opens the regular universities as well. The proportions are visible in the national figures: in 2023 the overall Matura rate was 43 per cent, made up of 23 per cent for the academic gymnasiale Matura and 15.5 per cent for the Berufsmaturität. The academic route through the Gymnasium is the smaller one.

One more piece of orientation. Education in Switzerland is a cantonal matter. There are 26 cantons, no national school system and no national school test, so which school types exist, what they are called and which assessments a pupil meets depend on where the family lives. The Multicheck is one of the few things that works the same everywhere, because it is a private product rather than a school one. The school-run assessment you will hear about in the same breath, Stellwerk, is not: whether it happens at all depends on your canton, and it costs CHF 10.00 rather than CHF 100.

The two pages of the certificate

The certificate has two pages, they show different numbers, and that is exactly what confuses people. The first page reports norm scores, drawn as horizontal bars. The second page lists the raw scores from which the norm scores were derived. According to the provider, the calculation is based on the raw values reported as the share solved correctly.

If you put the two pages side by side and the numbers do not match, you have done nothing wrong. They do not match because they answer two different questions.

Percentile rank is not percentage correct

This is the single most useful thing on this page, so it is worth slowing down for.

The provider says it itself: the norm scores are, in technical terms, percentile ranks. A percentile rank does not tell you how much you solved. It tells you how many other people you left behind. A score of 62 means that 62 out of 100 young people in the norm sample reach that value or less.

The practical difference is large. A raw score of 62 per cent of tasks solved can correspond to a percentile rank of 40 or of 85, depending on how hard that area was for everyone else. Someone who reads the norm score as a percentage is not out by a few points. They are out by an order of magnitude, and usually in the direction that makes them feel worse than the numbers warrant.

The reason the two are so easily confused is that both stop at 100. A useful habit is to read the norm score as a position in a queue rather than as an amount of anything. Position 62 in a queue of 100 people says nothing about what you carried; it says who is in front of you.

The overall fit with the occupation, a number with a sign

At the top of the first page of the newer evaluations sits the overall fit with the occupation, given both as a figure and as a graphic. It pulls together the normed test performance from all the competence areas and relates it to the requirements profile of the chosen occupation. Areas that matter more for that occupation carry more weight.

This number follows a different scale from the bars underneath it. The provider states that the value follows a standard normal distribution, that the unit is standard deviations, that the average sits at zero, and that roughly 68 per cent of all results fall between minus one and plus one.

The last two rows are calculated from the provider’s own figures rather than measured by us: if 68 per cent sit in the middle and the distribution is symmetrical, roughly 16 per cent are left for each tail.

A minus sign here is not a failure mark. It is the sign that belongs to every result in the lower half of a scale whose midpoint someone decided to call zero.

What “good” actually turns on

Asked what score is needed, the provider answers plainly: the statement about suitability comes from relating the individual result to a requirements profile in which a minimum level of fulfilment of an occupation’s demands is defined. There is no number that is good everywhere.

Two sentences from the same answer matter more than any threshold. First: a result below the desired range is not automatically a reason for a rejected application. Second: a result inside the desired range is no guarantee of an offer. The decision always rests with the training company.

That sounds unsatisfying, and it is the only honest answer. Anyone who gives you a fixed points boundary either does not know the requirements profile of your chosen occupation, or is selling you something.

Area scores, overall result, overall fit

Several aggregated values sit next to each other on the certificate, and they get mixed up regularly. Depending on the version, the area scores pull together school knowledge, potential and occupation-specific skills, or in newer versions technical, methodological, personal and social competences. The raw scores of the underlying areas are added up and then converted into a percentile rank by reference to the norm.

One detail that raises questions: in some versions, performance in the second foreign language, meaning Italian or French, does not feed into the school knowledge area score. The reason is comparability. Young people who chose different languages should be able to receive the same area score. The performance is still reported on the certificate; it simply does not count at that one point.

In practice: an area score is always a percentile rank, whereas the overall fit is a value in standard deviations. Read both on the same scale and you will misread one of them.

Why your school reports measure something else

The Multicheck does not know which school type you attend, and that is deliberate. The provider states that there is no alignment of the evaluation or its interpretation to different school levels. You are measured against the demands of the apprenticeships, not against the learning objectives of your class.

A school mark takes in participation, effort and homework over months. The Multicheck records school knowledge and cognitive ability in a single morning under standardised conditions. The provider points to studies at the universities of Bern and Basel and to a research project with Prof. Dr. Martin Kersting of the University of Giessen; there, a clear relationship between marks and test result showed up particularly in mathematics and English, while in areas such as logical reasoning the test adds information the marks do not carry.

What the certificate does not say

It says nothing about you as a person. It says nothing about whether you will get an apprenticeship place. The provider explicitly advises training companies not to use the Multicheck as their only selection criterion, but to look at it alongside a Schnupperlehre, the short taster placement in the company that is standard practice in Switzerland, a structured interview and the school reports.

It says nothing about how you will develop, either. A test score is a snapshot taken under time pressure on one particular day. How that fits into the wider picture, and what role aptitude tests play in the search for an apprenticeship at all, is set out in the overview of aptitude tests for an apprenticeship.

And it says nothing about how good somebody will be during the apprenticeship. The provider describes the purpose of the analysis as an information base for the best possible fit between an advertised apprenticeship and the individual circumstances of the applicant. A fit is a probability, not a verdict.

The most common misreadings at a glance

Almost every misunderstanding about the certificate comes down to a handful of sentences. They sound plausible and they lead the wrong way.

Hold those five in your head and you are already reading your own certificate more carefully than most of the people who will be handed it.

When the certificate arrives and who sees it

The certificate is made available for download in the gateway.one account within two working days of the test. May and June are the exception: because of the annual quality check, distribution there happens later, which is communicated before you book. An email announces that it is ready, and an SMS if you ask for one.

Everyone who takes part automatically receives a profile in the provider’s free careers portal. That profile stays anonymous until the young person activates it. Anyone who activates it can apply directly to companies through it, and companies can search for profiles. The certificate does not travel to a training company by itself; you decide who receives it, and when.

One thing that matters for planning an application is which version is being asked for. For an apprenticeship starting in August, training companies usually ask for a certificate from the version that has been running since the test period opened in the middle of May of the previous year. An older certificate is therefore not automatically usable, even when nothing about it is worse.

If you want to place your own scores

Anyone trying to understand what sits behind the task formats arrives quickly at the figural series and matrices that appear in almost every aptitude test. How those tasks are built, and how to recognise the rule behind them, is explained in understanding matrix reasoning.

The free test from Kognita works on the same basic idea as the Multicheck: it compares your result with a reference distribution instead of with a pass mark. It is not a substitute for an assessment by a qualified professional and it is not a substitute for the Multicheck, but in 25 minutes it shows you how a norm-based result comes about in the first place. If you have a Multicheck certificate in front of you that you cannot place, the Berufsberatung of your canton, the public careers guidance service, will help you free of charge.