The Recruiting Signal
Question. Does the aptitude a veteran sees while training a young operator predict CAF success better than standard recruiting does?
Analogy. You do not learn who someone is from an interview; you learn it weeks into a hard expedition, when the weather turns and the plan fails and you watch how they carry it. A veteran who guides a young operator for months is that expedition partner, with a read the recruiter's hour cannot reach. The study asks whether that long read predicts CAF training completion and service success better than standard recruiting, or whether it is just folklore once checked against real outcomes.
What's at stake. The CAF has a recruiting gap the Auditor General named, and standard recruiting leans on a first interview and an aptitude battery that see a person for an hour. A veteran who has guided a young operator through real problems for months has watched them handle pressure, fail, recover, and debrief, which is a far richer signal of the things service actually demands. The proposal claims this makes the network a better pre-selection channel. If it is right, the network produces recruiting signal the standard process cannot; if it is wrong, the claim is folklore and should not be put in front of a recruiting authority.
The two answers it decides between. Either veteran-observed aptitude, rated over months of real-stakes guiding, predicts CAF training and service success beyond what the standard recruiting measures capture, because it observes performance the interview and battery cannot reach; or the standard measures already capture what predicts success, or veteran observation is too biased and inconsistent to add signal, so the network channel predicts no better. Comparing both against actual CAF outcomes tells them apart.
What a null result would mean. If veteran-observed aptitude does not predict CAF success beyond the standard channels, the network-as-recruiting-signal claim needs revision: the network may develop people without being a better way to spot them. That is a finding about the predictive validity of the signal, not a sign the platform was built wrong.
Why this matters to defence. Recruiting and selection is a named, costly gap (DRDC Objective 2), and a pre-selection signal with real predictive validity beyond the current battery would be directly valuable. The result changes a concrete decision: whether to treat veteran-observed aptitude in the network as a recruiting signal, and a design constraint rides along: whether routing recruitment through a trusted-trainer relationship can survive operator sovereignty, the young person choosing and the platform holding no recruiter role, which is measured, not assumed, because a platform that quietly becomes a recruiter breaks the trust the signal depends on.
How we would run it. For young operators who later apply to the CAF, compare two predictors of their actual CAF outcomes, training completion, early performance, and first-term retention: the aptitude their veteran guide rated over months of guiding, and the standard recruiting measures. The discriminating measure is whether the veteran rating adds predictive validity beyond the standard battery. Because veteran ratings are observer judgments, the design needs more than one veteran rating where possible and a check that they agree, so the signal is not one person's impression. The only-us part is the months-long real-stakes observation a guide gets and a recruiter cannot.
Earliest start. Stage 13: the study needs young operators who later apply to the CAF and their actual CAF outcomes to validate the veteran rating against.