Sorry, I didn't specify well what struck me as odd: it's not the subject-verb agreement (which is simply a matter of bringing to consciousness something a normal three-year-old already does correctly in speaking). There's a substantial literature in child language acquisition pointing to the fact that passive voice is acquired rather later than subject-verb agreement: normal four-year-olds are known to have difficulty interpreting passive constructions - "The blue car was hit by the red car" will be acted out with the blue car hitting the red car - and it isn't until age 5 that normal children even start to produce spontaneous passives orally. An interesting, if technical, article on a possible semantic basis for the delay is
here. Explicitly teaching second graders to recognize passives is therefore odd; teaching them to write passives - which require one to juggle not merely the morphology (plain verb vs. BE plus past participle) but also subject vs. object syntactic roles AND agent vs. patient/theme semantic roles AND theme vs. rheme information roles) is just... overly optimistic.
