People are often treated as unreliable narrators of their own past, and this scepticism runs especially deep around emotionally charged early experiences.
Researchers have long worried that memories of abuse and neglect might shift depending on someone’s mood, mental health or current circumstances, meaning what someone tells a researcher, doctor or social worker one year might not match what they’d say the next.
Our new research, published in Nature Mental Health , suggests that these fears may be overstated. We found that reports of childhood maltreatment remain highly stable over time – at least over a period of a few years.
Childhood maltreatment covers experiences of physical, sexual or emotional abuse or neglect. Most research into how these experiences affect mental and physical health relies on retrospective self-reports. Essentially, it involves asking people to describe what happened to them based on memory.
These reports show consistently stronger links with mental health outcomes than reports from outside observers. But this has raised an uncomfortable question: are these stronger links genuine, or are they artefacts of people’s current mood, colouring how they remember their past?
Whether memories of maltreatment are stable isn’t just an academic curiosity; it has serious practical consequences. If people’s accounts of what happened to them shift over time, a single snapshot survey could misclassify who did and didn’t experience maltreatment, muddying research findings and making it harder to identify what’s actually driving poor outcomes.
The stakes are just as high outside the lab. Clinical, legal and social care decisions can hinge on someone’s account of childhood experiences, often given only once.
To test this, researchers have asked the same people about their childhood experiences at two or more points in time, then compared their answers. We pulled together 49 such studies, spanning almost 40,000 people, to see how consistent people’s memories really were.
We found they were remarkably consistent. Over an average gap of two and a half years, people’s memories of maltreatment barely budged, supporting the case for using a single time-point assessment in both research and clinical practice. That said, we still don’t know whether this stability holds over longer stretches of time, so more research is needed.
That’s not to say memory is perfect. About one in five people did change their response over time. This shouldn’t be read as evidence that someone was lying, though.
Memories can shift for all sorts of reasons, such as how someone comes to interpret what happened to them, ordinary quirks of memory, how comfortable someone feels disclosing sensitive information in a given setting, or simple human error.
This is why any record of maltreatment disclosure, whether in research or in clinical practice, should also capture the context in which it was made. This may well shape how consistently that account holds up later.
Memory also depends on the type of maltreatment
Some experiences also proved harder to remember consistently than others. Reports of neglect changed more often than reports of abuse, perhaps because abuse tends to involve specific, tangible events that anchor themselves in memory, whereas neglect is often about the absence of something – care, attention or resources that simply weren’t there. It’s harder to consistently recall something that didn’t happen than something that did.
Memories were also less stable in large, population-representative studies than among people who volunteered for research or who had been recruited through clinical services. One explanation is that people who sign up for studies on this topic may already have spent time reflecting on their past, and grown practised at answering these types of questions consistently.
People with poor mental health may also think about negative childhood experiences more often, and reflect on them more in treatment, making those memories more accessible and likely to be reported consistently.
Finally, we found that while adults’ memories of childhood maltreatment were very stable over time, young people’s memories of maltreatment were less stable and decreased over longer gaps between assessments.
This probably reflects the fact that children’s memory systems are still developing . Early memories may be laid down less firmly, leaving them more open to reinterpretation as children mature and come to understand their experiences differently.
Our findings carry a hopeful implication. Childhood and adolescence may be a particularly valuable window for offering support after trauma. Treatments such as trauma-focused cognitive behavioural therapy work by helping people to reinterpret traumatic memories and weave them into their broader life story in ways that ease long-term distress.
If young people’s memories are genuinely more malleable, that suggests adolescence could be an especially powerful moment to help them process traumatic memories, before those memories settle into a more fixed shape.
