Measuring Self-Directed Agency
Question. Is the disposition to act on your own judgment a trainable capacity, or a fixed trait?
Analogy. Is the disposition to act on your own judgment like fitness or like height? You can measure both, but height is fixed and fitness rises with training. The study builds the ruler and asks which one this is, whether it moves with practice or sits where it started, and proves it is real fitness and not just IQ in a tracksuit.
What's at stake. Every other claim in this programme assumes you can measure whether a person's capacity for independent judgment went up or down, and right now nobody can. If the disposition is trainable, a low baseline is a problem you can fix, and the institutions that lower it, and the adversary that exploits a lowered one, become measurable. If it is a fixed trait, training is theatre and the intervention is mis-specified.
The two answers it decides between. Either self-directed agency (agentic disposition) is an environment-driven, trainable capacity that moves with practice and is distinct from general intelligence; or it is a stable trait that schooling, posting, and tooling cannot shift. The ruler, validated against wargame performance and known-groups contrasts and held discriminant from IQ, tells them apart.
What a null result would mean. If the measure cannot be moved by training and tracks IQ, the trait-stability view holds and the trainability premise the platform rests on needs revision. That is a finding about the construct, not a sign the platform was built wrong.
Why this matters to defence. Personnel (DRDC Objective 2): do officers' agentic dispositions atrophy in long low-agency postings, and can guided real-stakes practice restore them? That is a DGMPRA-scale question with a concrete personnel decision attached. Future threats (DRDC Objective 6): if school and bureaucracy lower the disposition before any adversary arrives, cognitive warfare lands on pre-softened soil, and the entry baseline measures pre-existing vulnerability.
How we would run it. Build a measuring scale for the disposition on its named parts: running your own loop on a problem, revising your model when the world pushes back, handling not-knowing, picking up tools just in time. Then check the scale is real in three ways: does a person's score predict things it should (their wargame performance, whether their ideas get taken up by the network, whether they stay), does it separate groups it should (founders versus long-tenure bureaucrats), and, the hard gate, does it measure something different from raw IQ, because if it is just IQ we have built nothing new. Measurement rides the consented researcher role and the wargame, not the platform's behind-the-scenes signal. Training only counts as proof if the score moving also moves real outcomes off the platform.
Earliest start. Stage 7: no validated agency-disposition measure exists, so the scale itself is the deliverable, built on the researcher role and the wargame and validated sideways through the nomological net.