Child Interviews as Evidence: Teddy Bear Cards, Taping, and What is Missing in the File

Uppdrag granskning revealed unscientific child interviews that can lead to LVU (Care of Young Persons Act). The investigation contrasts the methods against the evidence requirement in the new Social Services Act, SBU's knowledge gaps, recorded child interrogations, and what is actually submitted to the administrative court.

Summary: On April 22, 2026, Uppdrag granskning highlighted how investigative child interviews in social services can become life-changing without any external party hearing what was said. In a study from the University of Gothenburg, which UG reports, over 200 social workers responded to questions about how they talk to children. More than 80 percent of those surveyed use methods lacking research support when investigating suspicions of abuse: teddy bear cards, where the child selects cards with different emotional states, and taping, where the child is interpreted based on play with wooden figures. The conversations are documented as memory notes. Police child interviews at Barnahus are recorded. In Helsingborg, the teddy bear cards were submitted to the administrative court as a basis for LVU. The municipality has reported itself under lex Sarah. It is not established that every child interview is wrong. What is established is the gap between the evidence requirement in socialtjänstlagen (2025:400) and what actually ends up in the file.

A child interview is not a subordinate clause to the investigation. It is often the material that the committee, the individual, and the court rely on when the child cannot or does not want to speak in their own words. When the method lacks research support and the recording is missing, the interpretation becomes the evidence. This is the subject of this text – not a replay of the program.

What UG actually showed

Uppdrag granskning's program Child Interviews – in the Closed Rooms of Social Services focuses on investigative conversations: conversations that can concern violence and abuse and form the basis for interventions and decisions, including custody according to LVU. The point in the open article on April 22, 2026, is not that social services talk to children. The point is that almost no one else knows what was said. The documentation consists of memory notes. Sara Landström, professor of forensic psychology, tells UG: “We know very little about these conversations. Unlike police child interviews, these conversations are not recorded.”

The study from the University of Gothenburg, which UG reports, shows two things that should be kept separate. Many social workers lack training in how investigative child interviews should be conducted. And more than 80 percent of those surveyed use methods without research support when suspicions of abuse are investigated. Landström warns of interpretation problems: that things are read into what the child has not said. Anneli Larsson, lecturer in social work, calls the situation “a bit of a wild west” and wants to see national guidelines. UG writes that there were no plans for that at the time. Social Services Minister Camilla Waltersson Grönvall said that what is not evidence-based and research-based should not be used – and that the new social services law should be evaluated.

Five days later, on April 27, 2026, the minister summoned the Director General of the National Board of Health and Welfare. She said she takes what has emerged seriously, that every investigation must be legally secure and knowledge-based, and that it is not acceptable for children's legal security to be determined by which municipality they live in. A meeting is not a regulation. It is the government's reaction to a program. This text proceeds to what can be substantiated independently: legal text, method guide, knowledge gaps, and a documented case.

The Law's Requirements and the Method Guide's Silence

Since July 1, 2025, the social services law (2025:400) applies. Chapter 5, Section 1 is brief: the activities must be of good quality and conducted in accordance with science and proven experience. Everyone performing tasks within social services must contribute to this. Chapter 5, Section 2 places the responsibility on the social committee to systematically follow up, develop, and ensure quality. Chapter 5, Section 3 requires appropriate education and experience for the staff. This is not a political intention. It is current law.

The requirement is not the same as every conversation support needing a large effect study. The National Board of Health and Welfare describes evidence-based practice as a synthesis of research, proven experience, professional expertise, and the individual's situation. This is an important clarification. It does not save a method lacking research support when the method is used to investigate suspicions of abuse and then submitted as evidence to the court. Then the question is not whether the child got to speak. Then the question is whether what was noted can be tested.

The National Board of Health and Welfare's method guide on taping is the primary source of what the authority actually states. The guide clarifies that the presence of a method is not a recommendation. Scientific support should be evident in the knowledge base. Taping is described as a visual representation with painted wooden figures on a taped "stage." It is not manual-based. Training is provided by a private consultant. Under knowledge about the intervention's effects, it is stated that SBU has identified scientific knowledge gaps: one concerning support for foster homes and children placed in foster care, and one concerning families where children are known or suspected to have been subjected to violence and neglect. It is precisely the second gap that sharpens UG's question. The method is used in investigative conversations about abuse. The knowledge authority has indicated that the research is insufficient there.

Teddy bear cards are not recommended as an investigative method in the same guide. Landström tells UG that there is a lack of research specifically on teddy bear cards and that the child may have difficulty understanding what to say. According to UG, the European Association of Psychology and Law has warned against unscientific tools of this kind. It is a researcher warning, not a Swedish regulation. It is still sufficient to distinguish three things: that the child needs support to tell their story, that a card or figure can help in treatment—and that the same tool becomes something else when it is to support a coercive decision.

Helsingborg: the teddy bear cards in the administrative court's file

UG's documented case is not an anonymous illustration. In Helsingborg, a father was accused of serious abuse against his son Adam during a custody dispute. Police and prosecutors investigated and closed the case. The district court and the court of appeal granted the father sole custody. The social services continued their stance: that the father had exposed the son. In the investigative child interviews, teddy bear cards were used. According to the file, Adam was supposed to be angry and sad with the father. Adam, 15 years old when UG aired, says: "That was not the case at all. She tried to get what she wanted, and I wasn't allowed to say what I wanted either."

The conversations formed the basis for isolation and intensive therapy. When the social services applied for LVU, the teddy bear cards were submitted to the administrative court as evidence. This is the core of this text. Not that a child chose a card. But that a method without research support became a document in the court's file, while there was no recording to compare the notes against. The social services in Helsingborg have reported themselves to IVO under lex Sarah and particularly criticize how the child interviews were conducted. Social Director Johan Klingborg declined an interview.

Lex Sarah is the statutory system for reporting, investigating, and remedying misconduct within the organization. A report to IVO is not the same as IVO having already determined a specific fault, and it is not a review of the custody judgment. It is the municipality's own admission that the conversations did not hold. Motverket's previous review of legal certainty and the LVU proposal 2027 addresses the same structure: documentation must be traceable, conflicting information must be disclosed, and the child's position must not be overinterpreted. Helsingborg is the documented example of what happens when these requirements are not met in the conversation that is supposed to support the decision.

Three thresholds: conversation, file, coercion

Distinguish between three thresholds.

The police's child interviews at Barnahus are the visible contrast, not the template for the social services. They are recorded, follow a more structured methodology, and can be reviewed. The social services' investigative conversations have a different purpose: to assess needs and protection, not to prosecute a criminal case. The difference does not explain why the very conversations that can lead to LVU lack equivalent traceability. It sharpens the question. The child can be heard twice, in two systems, with two different evidentiary requirements. Only one of the conversations can be replayed.

What the reader should do with this

Frequently Asked Questions

Does 80 percent mean that four out of five child interviews in Sweden are unscientific?
No. The figure refers to the social workers who responded in the University of Gothenburg's survey, reported by UG: more than 80 percent use methods without research support when investigating suspicions of abuse. It is not a sample reported as a national proportion of all conversations.

Does Socialstyrelsen recommend taping?
No. The method guide explicitly states that the presence of a method is not a recommendation. SBU has identified knowledge gaps, including in cases of suspected violence and neglect.

Is the social service required to record child interviews?
There is no general recording obligation equivalent to police child interviews. SoL requires documentation of significant information and activities according to science and proven experience. The absence of recording is therefore not automatically illegal. It is a legal security gap when the note becomes evidence.

Is Helsingborg proof that LVU always relies on teddy bear cards?
No. It is a documented case where the cards were submitted to the administrative court and the municipality then reported itself under lex Sarah. One case shows how the chain can look. It does not translate to all custody cases.

Has the evidence requirement in the new SoL made the methods impermissible?
The law has required science and proven experience since July 1, 2025. It does not list teddy bear cards or taping. The minister's statement is political guidance, not a prohibition list. The question in an individual case is whether the basis can be tested and whether the method's limitations have been disclosed.

Sources