OBJECTIVE To examine those sources of information which nurses find useful for reducing the uncertainty associated with their clinical decisions. BACKGROUND Nursing research has concentrated almost exclusively on the concept of research implementation. Few, if any, papers examine the use of research knowledge in the context of clinical decision-making. There is a need to establish how useful nurses perceive information sources are, for reducing the uncertainties they face when making clinical decisions. METHODS Cross-case analysis involving qualitative interviews, observation, documentary audit and Q methodological modelling of shared subjectivities amongst nurses. The case sites were three large acute hospitals in the north of England, United Kingdom. One hundred and eight nurses were interviewed, 61 of whom were also observed for a total of 180 hours and 122 nurses were involved in the Q modelling exercise. RESULTS Text-based and electronic sources of research-based information yielded only small amounts of utility for practising clinicians. Despite isolating four significantly different perspectives on what sources were useful for clinical decision-making, it was human sources of information for practice that were overwhelmingly perceived as the most useful in reducing the clinical uncertainties of nurse decision-makers. CONCLUSIONS It is not research knowledge per se that carries little weight in the clinical decisions of nurses, but rather the medium through which it is delivered. Specifically, text-based and electronic resources are not viewed as useful by nurses engaged in making decisions in real time, in real practice, but those individuals who represent a trusted and clinically credible source are. More research needs to be carried out on the qualities of people regarded as clinically important information agents (specifically, those in clinical nurse specialist and associated roles) whose messages for practice appear so useful for clinicians.