The Field Guide › Paper
Beaubien and Baker put a number on a claim the CRM literature had mostly argued from case studies and theory: that how CRM is delivered matters as much as whether it is delivered at all. They surveyed 30,752 pilots across 24 of the 30 largest US carriers, working through their unions, with careful screening for age, employment plausibility, missing data and straight-lining, and cross-checking respondent demographics against the sampled population to rule out response bias; 10,166 responses survived the filtering, in what the authors describe as the largest training evaluation conducted to date. Their central comparison is between the Advanced Qualification Program (AQP), which mandates that CRM be integrated throughout technical training rather than delivered as a bolt-on module, and the older Part 121 regime, in which it more often was not. The result favoured integration on every measured dimension: usefulness of feedback, relevance to current line operations, clarity of objectives, and preparedness to fly the line, with the largest single effect in the analysis being the difference between CRM taught as a separate course and CRM taught throughout the curriculum. Almost regardless of which programme they trained under, the overwhelming majority of pilots, 86 percent, agreed that CRM was an important topic and should be integrated throughout training, so the disagreement in the data is not about whether CRM matters but about how well any given programme actually delivers on that shared belief. The finding that complicates any easy AQP-is-solved conclusion is that even under a programme that mandates integration, 11.5 percent of AQP-trained pilots reported receiving no CRM training beyond basic line-oriented flight training, a policy-to-practice gap the authors call puzzling and can only speculate about: confusion over what counts as CRM, or genuine variability in how faithfully individual carriers actually implement the mandate. For a corpus about psychological safety this paper supplies the large-scale, quantitative counterpart to the qualitative and historical case, made elsewhere in this map, that a safety practice has to be lived rather than merely delivered: at the scale of an entire industry, bolting a good idea onto existing training produces measurably worse uptake than weaving it through the work itself, and a mandate on paper does not guarantee the practice on the ground. Its limits are real and are stated candidly by the authors themselves: it measures pilots' satisfaction and perceived usefulness rather than behaviour or safety outcomes, and it was published as a conference paper rather than a peer-reviewed journal article, so its methodological transparency is high but its evidentiary weight is that of a large, well-conducted reaction survey rather than an outcome study.