John Key's announced a fairly small-scale breakfast-in-schools programme: they're expanding the current 2-day per week KidsCan programme in decile 1-4 schools with a 5-day version.
Recall that decile 1-4 schools are the same cohort that was included in the Mhurchu et al randomised control trial that found low uptake and little benefit. I summarised the literature here. When NBR reprinted it, it drew a lot of angry comments. Heck, even suggesting on Radio NZ that the Mhurchu et al piece showed fairly small effects can get you called a dickhead on twitter. The government had to do something in this space; this was something. It will be very popular.
The programme is not set to cost very much. I do not expect that it will achieve much, but it will achieve little at less expense than would Hone Harawira's plan that would also be likely to achieve little but at much greater expense. So it has that going for it.
Seamus wonders whether the government will roll out the programme in such a way as to facilitate testing of effects. It would be great if they did. And, it wouldn't be hard either. Randomly assign decile 1-4 schools to phase into the programme at different times and survey kids all the way through so that you could check whether the programme were doing much. Even better if they varied delivery as part of the trial to see whether different implementation techniques were more or less effective in encouraging hungry kids to take part, then put out some best practice guidelines in a few years' time.
If a genie gave me wishes, one of them would be: "Where the government can roll out some new scheme in such a way as to allow for effect evaluation at relatively low cost, it do so by default." I'd also like a pony, and infinite more wishes.