Statistician queries DCC’s analysis of online survey

Dunedin City Council claims about what "most respondents" said in a survey have been questioned by a statistician.

The council ran an unscientific online survey as an exercise in early engagement ahead of its 2025-34 long-term plan being drafted.

In a meeting agenda, the council said 138 responses were received and most respondents wanted it to "do more" — either by a bit or a lot — in five out of six categories.

The exception was libraries, arts and culture, for which respondents "in majority" asked the council to "do about the same as now".

A graph produced by the council seemed to contradict its interpretation.

The Otago Daily Times was able to find a clear majority for the combined bloc of "do a bit more than now" and "do a lot more" in just two categories.

This became three if "I don’t know or no answer" was ignored in the category of social and community wellbeing.

Asked to explain, the council said it was comparing "do a bit more" and "a lot more" with "do a bit less" and "a lot less".

"Do about the same as now" was regarded as neutral and not grouped with either of the other two blocs.

"We’re entirely comfortable the advice to council reflects the results of our early engagement survey," a spokesman said.

The ODT invited input from the University of Otago’s department of mathematics and statistics.

After looking at the council’s agenda, senior lecturer Dr Tilman Davies said the council seemed to be confusing the idea of the statistical "mode", or the "most frequent" group, with "the majority" in a sample.

This would amount to the same thing if there was a contest between two groups, but a majority (greater than 50%) was more difficult to achieve if there were several groups.

Dr Davies said it was inappropriate to leave out "do about the same" when comparing the combined "do more groups" with anything else.

"If you’re doing ‘about the same’, [this] has to be combined with the ‘do less’ crowd if you want to make claims about exclusively ‘doing more’ as being in the majority."

At a meeting this week, Cr Lee Vandervis highlighted the "do more" bloc did not attain a majority in most categories.

Corporate policy acting manager Nadia Wesley-Smith seemed to reply it had more than either of the other two groups most times.

The ODT was unable to establish whether she believed this amounted to a majority.

Dr Davies said Cr Vandervis was correct.

He also supported a point made by Cr Steve Walker.

This was that the sample size was too small to read anything much into the results.

grant.miller@odt.co.nz