Friday 22 February 2013

Regional heterogeneity

You can hide a lot in an aggregate.

David Farrar rightly notes that the recent CERA / Nielsen survey weighted survey responses by district populations across Christchurch City, Greater Christchurch, Selwyn and Waimakiriri. So some reports suggesting the report masked things by weighting each of Christchurch, Selwyn and Waimak equally were wrong.

But there's a potentially much bigger masking of heterogeneity.

After the earthquakes, Christchurch turned quickly into three cities, as Peter Hyde put it:
When we got power back on a week after the February earthquake, I sent out an electronic plea for more direct support and attention for the worst-affected suburbs of Christchurch.
This was in response to the "three cities" I saw developing - Rescue City in the photogenicly-ruined CBD, Shower City in the areas which had their services largely intact, and Refugee City where tens of thousands huddled amongst broken houses, rockfall and liquefaction.
When we bugged out from South Brighton for a house in Wigram on the Friday morning after the quake, it was like moving to another world. There was power. There were supermarkets. People were watering their lawns despite the sewerage system being in disarray. It was as though nothing had ever happened, barring a few chimneys. A lot of the East remains a rather thorough mess two years later.

I remember answering the CERA survey. I don't think I counted as a satisfied customer.

If you look at Appendix 2, every respondent was mailed the survey along with a username and unique survey code. So each response was tied to an address, unless they chose to blind the back-end so they wouldn't be able to tie respondents to addresses. They also had a question in there asking what address you were at prior to 4 September if you'd moved since the quake. That question says
"Please note: this information will only be used to see if there are differences between different areas. Your individual information will not be looked at separately."
Maybe they only ever wanted to aggregate up to Christchurch City, Greater Christchurch, Selwyn District, Waimakariri District level. But the data should be there for doing a within-Christchurch disaggregation. I'd be very surprised if there were not exceptionally strong heterogeneity between Shower City and Refugee City. Maybe nobody ever ran the borough-level analysis. Or maybe not - I really don't know.

If Lianne Dalziel is on her game, she'll already have an OIA request in for the data aggregated by borough. And if it shows that residents in the East - those most affected by the quakes - have reported utter dissatisfaction with quality of life, quality of government response, and pretty much everything else, I'd be pretty surprised if she weren't very vocal about it.

Happy two year anniversary, #eqnz.

1 comment:

  1. A few of us in the west didn't take the oppurtunity of a free paint and decorate.