Childhood trauma is a key modifiable risk factor for psychopathology. Despite significant scientific advances, traumatised children still have poorer long‐term outcomes than nontraumatised children. New research paradigms are, thus, needed. To this end, the review examines three dominant assumptions about measurement, design and analytical strategies. Current research warns against using prospective and retrospective measures of childhood trauma interchangeably; against interpreting cross‐sectional differences in putative mediating mechanisms between adults with or without a history of childhood trauma as evidence of longitudinal changes from pre‐trauma conditions; and against directly applying explanatory models of resilience or vulnerability to psychopathology in traumatised children to forecast individual risk in unseen cases. The warnings equally apply to research on broader measures of adverse childhood experiences (ACEs). Further research examining these assumptions can generate new insights on how to prevent childhood trauma and its detrimental effects. Read the Commentary on this article at doi: 10.1111/jcpp.13195