Everything in the analysis comes from teachers describing their own situation in a single survey pass [1], work run with the American Psychological Association's Task Force on Violence Against Teachers and School Personnel and published in Educational Researcher [2]. Support here is a perception measure: whether a teacher feels comfortable raising a discipline problem with the principal, and whether they believe administrators will back them after an incident [6]. Stress, perceived safety and intent to transfer or quit come from the same person on the same instrument [4][21]. A teacher already halfway out the door will tend to rate the same principal more harshly than a colleague who is not, so the plain protective correlation has an obvious alternative reading.
The moderation result holds up better against that objection. If sour mood inflated both halves, you would not expect the association to thin specifically in the buildings where violence is high [8], which is what the team led by Andrew Perry reports [3]. That shape says the value of principal support is conditional on something the principal does not control. What the write-up does not supply is how much thinner: no coefficient for the interaction, no statement of how much protection survives at high victimization [22]. Absent that, "largely cancelled" and "modestly attenuated" are both consistent with what is published, and they imply different budgets.
The denominators are worth sitting with. Forty-four percent of 5,550 is roughly 2,440 teachers reporting student victimization [7][15]. The 80 percent figure is a ceiling rather than a point estimate, and taken at face value it would mean at most about 4,440 teachers reporting verbal or threatening aggression [17]. Prior work the team cites puts front-line teachers at greatest risk of exactly that nonphysical kind, including threats, obscene remarks and slurs [18].
The outcome measured is intent, not departure, and the paper supplies its own conversion: about one in three teachers who say they want out actually go within the year [5]. Read forward, a policy that moves 300 teachers off the intent-to-leave column should be budgeted as roughly 100 classrooms kept staffed [16]. That is a useful discipline on a literature that mostly measures the cheaper quantity.
Elementary grades carry the heavier victimization load [10], which is where a support-only retention strategy should fail first. Perry's 2024 evaluation in School Psychology found that zero tolerance and physical hardening did not work, while restorative, prevention-based approaches did better [11], so the thing that would restore the value of a supportive principal is the thing many districts have funded least. Perry's own framing is narrower than the press cycle around teacher shortages tends to be: support is "beneficial and necessary, but not sufficient" [9], and the researchers pair it with clear discipline policy and prevention work rather than offering it as a substitute [13]. The 2020-21 fielding year is a caution on the base rates [1]; the shape of the interaction is likelier to travel than the percentages are.
