Cermat is preparing a public database of tasks for admissions. Up to 100 thousand examples

Admissions are over and graduations have begun. Zermat is in charge of both essential tests. Its director Miroslav Krejčí describes in an interview for Seznam Zprávy that most children should get to school in the first round.

At the same time, he explains why an error appeared in one of the tests this year, and that the same thing probably happened in the replacement date. Cermat is therefore preparing a public database of tasks that could prevent unclear or incorrectly entered tasks.

Both the matriculation exams and the entrance exams could also be taken electronically in the future.

The children have completed their registration and entrance exams. Of course, the division into schools is pending. But can you already evaluate how it all worked out?

We have simulations available and even the first data from DiPSy (Digital login system, editor’s note), when it looks like over 90 percent of applicants will get into secondary schools in the first round. It will be a little less in Prague, a little more in other regions. The first simulations also show that up to 75 percent of children could get to their first chosen school, roughly 10 percent to the second priority and 5 percent to the third.

Do you take this as a measure of the success of electronic applications, that a far greater proportion of children get into the first round than at the time of paper applications?

However, we do not know how many children did not make it to the first round in previous years. In the past, the media came up with the fact that after the first round, almost half of the kids didn’t get into high school, which was true at the time. Subsequently, however, appeals began to be addressed and the fact that the vast majority of those who were not accepted were appealed, the media did not write about it afterwards and the ministry never collected this data. We will have such numbers this year, but we have nothing to compare them with.

But at least there will be less chaos with registration tickets like last year. It is so?

This was the main reason why we introduced prioritization. I see the centrally announced second round as the second biggest benefit of all the changes. As recently as last year, parents were calling schools to find out which would enroll him. But each of them wrote it out and closed it differently, the information was nowhere clear, and certainly not in one place. This year, on May 20, at DiPSy, we will give parents a list of all the schools in the country, including their capacities for the second round, and they will be able to choose easily.

When you talked about simulations in the first question, does that mean that the allocation algorithm is already linked to the application system?

On the one hand, simulations were made by CERGE, which showed that roughly 13.5 thousand children would not be classified. We gave them data to estimate the results. So far, we have run the algorithm with random results and it turned out that we did not divide about 11,000 applicants. Apart from Prague, in all regions, after the first round, there will be more free places in schools than unassigned children.

When will the sharp division of pupils occur?

On May 6, we give the results of the unified entrance exam to the directors. Based on them, they will make a ranking, give it back to us, and we will divide the ninth. Then it will be time to inspect the file and the results will be published on May 15.

Do you believe everything will work out?

Sure it will work.

Do you already have in mind what you would like to change in your applications for next year?

The issue of changing an already submitted application was much criticized when the system was launched. However, the restrictions are of a legal nature. It should be said that the administrative procedure does not start until February 20, i.e. with the end of the deadline for submitting applications. Then we could set that a new application can only be submitted if you first cancel the previous one.

We will also discuss whether to abolish hybrid applications. Therefore, only electronic and paper ones could remain. It turned out that the hybrids didn’t help much and sometimes made it more complicated.

There were an awful lot of cases where parents didn’t read what was on the monitor and on the paper and thought they had submitted the application by printing it out but not taking it to school. And then a week after the deadline, they found out that they didn’t have children registered.

And there was probably nothing that could be done about it. Or yes?

Directors always decided on individual cases. It is not for us to interfere in any way. I don’t even want to know how they ever did it.

I understand that you are responsible for the system, not for specific applications…

On the other hand, it is sad to tell a girl that she will not go to secondary school, because her mother submitted her application, but the day before the JPZ, she realizes that she applied herself and not her daughter.

Yes, that happened. Just like the case of an applicant from Prague who applies only to schools in Pardubice. That’s a problem when that wasn’t the intention.

We also dealt with the case of a duplicate application, when divorced parents did not agree. Have such ambiguities been resolved somehow?

A week ago there were 40 applications of this nature. We started to solve it with schools. For the most part, it turned out that the directors had everything right on paper, but did not correct it in DiPSy. At the moment, there are still six candidates with applications in conflict. We have three promised to be resolved. However, we have an indication that they will remain unresolved for three of them, because they are in court.

The system cannot include them in the distribution process. They will remain in the status as if they had not submitted any application. The child will not be placed in any school in the first round.

Looking at the whole new e-login system, how do you take back the criticism that was directed at you?

Everything went as it should for me. I take the fact that we started the system a day later as a reward for the deadlines we worked on. I don’t see any problem with that. In addition, the application deadline was extended.

What would you do differently if you could advise your younger self?

There would be only one piece of advice – make more time for developers in the fall.

And would you involve your son in the work?

If it was needed as much as this year, then yes.

An ambiguous task appeared in one of this year’s tests. Have you found out why this happened?

The change in the keyword, which caused the ambiguity, only took place during the last revision of the Czech language. That is, after all math experts have checked the problem. But I can’t just blame it on them because even the original version could be hacked.

It looks like we will have a similar problem in one more task from the replacement dates, which is already 100% for revision of the Czech language.

Are you thinking of making any changes to the assignment preparation system to avoid similar mistakes, as ambiguous assignments appear almost every year?

As long as the tasks are continuously created and everything remains the same secret, mistakes will continue to occur. We would have to have months more time to do it. We do not put the tests on the Internet at all, so everyone who checks them has to physically come to Cermat.

In my opinion, it will only change when we create a database of tasks that will be public. If we have a hundred thousand examples on the Internet, which will consist of the entrance exams, then we will catch all potential mistakes in two or three years by how the students and teachers will calculate the tasks and report back to us the ambiguities.

And are you creating such a database?

That is why I came to Zermatt, but I have not yet started because there was no time. When I joined, I said we would make a database of 10,000 examples, at this moment I have 30,000 on my desk. Next year I want it up and running.

The matriculation exams started this week, so I would like to ask how far along is the digitization of the matriculation exam?

It is more or less related to the task database. Realistically, however, we will launch the electronicization of matriculation and admissions certificates as a separate project.

It is probably also necessary for schools to be technologically prepared for such a change.

On the contrary, I say that we have to prepare it so that the schools can make do with what they have, because it makes no sense to start electronicization when we say that 300 million must be poured into computers and tablets every year.

This is precisely what the task database is related to. When it is large enough and usable, we don’t need to have one test for all children at the same time, but we can easily divide the graduation into fifths and each group will get a different test, but it will be comparable. Then the schools can do without the technology because they won’t need so many devices at once.

What about the security issue? Is there a risk of data leakage with electronic matriculation exams?

The moment they are electronic, the system will run, but it will not have data. On Monday morning, we’ll just determine which assignments will be on the final exam, and the first people to ever see it will be the students when the test starts.