Of 28 children's bicycle seats, only four passed everything in the EU test. The rest had a defect in construction or labeling

On September 5, the SITA agency published a report on an EU test in which 68 percent of children's bicycle seats failed. The number checks out, and so does the comparison with the earlier round from 2019, but the test had a second layer that didn't make it into the agency dispatch: alongside the laboratory tests, there was a check of labeling and instructions, and once that was included, 24 of 28 collected seats failed to meet the requirements. Only four passed absolutely everything. The method of selection also matters — the dispatch says it was random, while the original report speaks of risk-based selection.

What was tested and under which standard

The testing took place under the Coordinated Activities on Product Safety program, abbreviated CASP, which is funded by the European Commission and in which national market surveillance authorities jointly purchase and check a selected category of goods. In 2025, the program covered two such categories, chemicals and children's bicycle seats, plus an accompanying activity on mental health and new technologies. The report files are dated July 15, 2026. CASP 2025, súhrnná správa ↗

The authorities collected 28 seats, of which 21 were rear seats for children up to 22 kilograms, six were front seats up to 15 kilograms, and one was a rear seat up to 15 kilograms. Three samples were lost on the way to the laboratory, so the tests were actually carried out on 25 units, while the authorities checked the documentation for all 28. The activity covered seats for children from nine to 22 kilograms who can already sit up on their own, on both bicycles and e-bikes with pedal assistance up to 25 kilometers per hour. CASP 2025, správa o sedačkách ↗

Here is the first discrepancy compared with what reached the Slovak reader. The dispatch states that the tested models were intended for children up to 15 kilograms, but three-quarters of the samples were rear seats for children up to 22 kilograms — that is, for a considerably older and heavier child. SITA ↗

Testing was carried out under the EN 14344 standard in its 2022 version, which replaced the original from 2004 and added requirements on strangulation risk, flammability, and product labeling. The test plan covered general requirements, thermal risks, and mechanical risks. The activity began at the end of June 2025, and the laboratory finished its work in November of the same year. CASP 2025, správa o sedačkách ↗

The second layer of checks and the 86 percent figure

The laboratory tests, which produced the better-known figure, were failed by 17 of 25 tested units, or 68 percent. Alongside them, the authorities also checked warnings, labeling, and instructions in the official language of the country where the seat was sold, and here 21 of 28 samples failed, or three-quarters. The most common shortcomings were a missing figure for the child's maximum height and weight, the manufacturer's or EU contact point's name and address not being visible enough, missing requirements for selling rear seats, and incomplete usage information. CASP 2025, správa o sedačkách ↗

The two figures do not simply add up, because fourteen seats failed both layers at once. When both are counted together, 24 of the 28 collected units failed to meet the requirements, or 86 percent, and four passed without any reservation whatsoever. The three lost samples enter this count only through the documentation check, since the laboratory never saw them, and the report does not state whether one of them might actually be among the four flawless seats. The program's overall summary report lists one more non-compliant sample in the documentation check than the detailed report on the seats, though the two agree on the total figure of 24. CASP 2025, súhrnná správa ↗ CASP 2025, správa o sedačkách ↗

Where the seats failed

The following breakdowns concern the 25 laboratory-tested units, not all 28. The shortcomings mainly concerned restraint systems, mechanical strength and durability, edges and protrusions, side protection, and foot-attachment systems. The highest failure rate was for the requirement on sharp edges and protrusions, which 11 of 25 samples failed, or 44 percent. Five seats failed a large number of provisions at once, each in 13 to 16 points of the test plan. CASP 2025, správa o sedačkách ↗

The report's authors identified as the most serious those defects that an adult rider would not easily notice. These involved straps and restraint systems: failures in the tip-over test, insufficient strength and durability after simulated material aging, poorly functioning adjustment mechanisms, and restraint systems lacking a child-safety lock. CASP 2025, správa o sedačkách ↗

The authorities assigned a risk level to only part of the 24 non-compliant units. Two were rated as serious risk, one as high, three as medium, and four as low; two are recorded as purely formal non-compliance, and for twelve — half of the total — no risk level is reported at all. Twenty measures were adopted in total: five requests for the manufacturer or importer to modify the product, four withdrawals from the market, three sanctions against economic operators, three requests to add warnings, one recall from end users, one removal of a listing by an online marketplace, and three measures classified as other. Two notifications from this campaign were added to the Safety Gate system. CASP 2025, správa o sedačkách ↗

Samples were selected by risk, not at random

This is the second, and more important, discrepancy compared with the agency's coverage. The dispatch states that the authorities randomly selected 12 seats in brick-and-mortar shops and 13 online. The summary report describes it differently: "While the approaches varied slightly, MSAs generally applied a risk-based sampling method." The abbreviation MSA refers precisely to the national market surveillance authorities that purchased the samples. The authorities therefore mostly reached for goods where they expected non-compliance. CASP 2025, súhrnná správa ↗

This is not a mere formality, because it determines what the 68 percent figure actually means. With random selection, it would be an estimate of the share of non-compliant seats on the market; with risk-based selection, it is the result of a check targeted at suspect goods. To this must be added the scope, namely 25 laboratory-tested units from nine countries. The report concludes with a sentence that should accompany every citation of this figure: "The final results do not offer a statistically robust representation of the European Single Market." CASP 2025, súhrnná správa ↗

A stricter standard than in the first round

The Slovak dispatch also included a comparison with 2019, when every one of the 44 samples in the same category failed. What it lacks, however, is one thing: the two rounds were not conducted under the same regulation. In 2019, testing followed the 2004 standard; this time, the stricter 2022 version was used. The result is therefore better, even though the bar was raised in the meantime. CASP 2025, správa o sedačkách ↗

Which of this year's figures is comparable to 2019 cannot be determined from the document. The report does not state whether the earlier result concerned only the laboratory tests or also the labeling check. In the first case, 68 percent stands against it; in the second, 86. Repeat testing, moreover, is deliberate, not accidental: the program purposely returns to categories that had a high incidence of non-compliance in the past, precisely so it can be said whether anything has changed. Between 2019 and 2025, nine more notifications on children's bicycle seats were added to the Safety Gate system, beyond the two from this year's campaign. The authors summarize their conclusion as follows: "Although the results show an improvement compared with the CASP 2019 activity, where all 44 tested samples were non‑compliant, the failure rate remains concerning." CASP 2025, správa o sedačkách ↗

Online and brick-and-mortar retail

Of the 13 seats bought online, 11 failed, or 85 percent, while of the 12 units bought in brick-and-mortar shops, exactly half failed to pass. These are small numbers, and the goods were selected by risk, so the difference cannot be extrapolated to the entire market. CASP 2025, správa o sedačkách ↗

The data on origin can easily be misread. In absolute numbers, most of the non-compliant seats did not come from distant imports but from EU and European Free Trade Association countries, where ten of 16 samples failed. In proportional terms, however, it is the opposite: those ten out of 16 amount to 63 percent, while among the six units originating outside this area, five failed, or 83 percent. For three samples the origin could not be determined, and two of them failed. CASP 2025, správa o sedačkách ↗

The reports came out in English too, and no Slovak authority took part

In 2025, the overall program involved 26 authorities from 18 member states, and not one of them was Slovak. The seats specifically were tested by twelve authorities from nine countries, including the Czech Trade Inspection Authority. The Czech Republic also has a second authority on the list, the Ministry of Industry and Trade, which took part in the accompanying activity. Participation is voluntary, and the Commission asks authorities beforehand which categories they want to check, so non-participation alone is not a shortcoming. CASP 2025, súhrnná správa ↗

The test's recommendations are aimed not only at parents and manufacturers but also at standard-setters. According to the authors, the provision on sharp edges would need to be clarified, since its assessment has become more subjective, and the standard should also address how thicker clothing in cold weather affects the correct tightening of the straps. The ANEC association, which represents consumers in standardization, contributed to this part. CASP 2025, správa o sedačkách ↗

For the reader, however, something else matters more: both the detailed report and the two-page overview for consumers came out in every official language of the Union, English and Slovak included, are freely accessible, and contain concrete advice on what to look out for before buying and before riding. Among other things, they recommend checking whether the seat has shoulder straps, footrests, and foot straps, whether the point of sale states the child's maximum height and weight, and whether the seat has any sharp edges, screws, or nuts that could catch on clothing. CASP 2025, informačný list ↗

A list of the specific products the authorities assessed as dangerous is published on the Safety Gate portal. The agency dispatch referred readers to it as well, and, within the limits of its genre, it did provide the main figures, including the comparison with the first round. What is missing from it is precisely the part of the report that puts that main figure into context. SITA ↗

Illustrative photo: a bicycle with a child seat parked near a shopping center in Tomaszów Mazowiecki, Poland, October 2022. Author WrS.tm.pl, Wikimedia Commons, public domain.