Juvenile suspects are routinely expected to possess an accurate recall of written or oral Miranda warnings. This study addresses the Miranda‐related comprehension recall and reasoning of legally involved juveniles. It is the first juvenile research to compare systematically two levels of complexity for Miranda warnings with the three modalities (oral, written, or combined) of administration. Unexpectedly, easily read written warnings marginally outperformed the combined modality. In order to examine Miranda reasoning, three juvenile groups were operationalized: impaired, questionable, and likely adequate. Predictably, the impaired and questionable groups possessed significantly lower verbal abilities than the likely‐adequate reasoning group. In addition, the likely‐adequate group exhibited the strongest appreciation of the adversarial context in which Miranda waiver decisions are rendered. The discussion addresses the marked disparities in Miranda recall from a total recall versus component‐by‐component understanding of Miranda rights. It also considers more generally how crucially important Miranda misconceptions might be remedied. Copyright © 2016 John Wiley & Sons, Ltd.