Unfortunately, the Pygmalion Effect is nowhere near as "proven" as this article suggests. My reading of the various efforts to replicate / settle the matter is that the effect is negligibly small if it exists at all.

Consider, for instance, https://thenewstatistics.com/itns/2018/04/03/weve-been-here-before-the-replication-crisis-over-the-pygmalion-effect/