I wonder if the difference in skill is not only to do with listening, but also to do with the ability to process symbols.

It could be that the children who recognize the "da" as a repeating element do so because they assign it some sort of symbol, say x, and process all instances of it as x, whereas the children who do not recognize the "da" as a repeating element do so because they have not assigned a symbol to it and, therefore, expect a different instance each time.

That would translate, when they were older, into reading. A symbol-recognizing child would simply learn that knife reads /'naɪf/, and means sharp cutting object, whereas a non-recognizer would painstakingly read each instance phonetically and thus find it harder to translate into meaning?

I wonder if similar patterns are replicated in symbolic languages and phonetic languages. It would be interesting to compare, say, reading development between Chinese kids and Spanish kids or something.

I'm totally waffling, I know nothing about any of this. Child development is FASCINATING though, thanks for posting this!