At this stage of the elec­tion cycle we are drown­ing in polls. While each media out­let treats its own com­mis­sioned research as holy writ, they all have their own strengths and weak­nesses, and it takes some­thing like Danyl McLauch­lan’s track­ing poll charts to smooth out the noise and cal­ib­rate them against actu­al elec­tion res­ults. Even with that, there’s a grow­ing prob­lem with tra­di­tion­al polling: it only reaches people with land­lines.

Polling cell­phones is dif­fi­cult, and some poll­sters think that the res­ult­ing bias isn’t such a big issue, using demo­graph­ic weight­ing in an attempt to com­pensate. How­ever, there are oth­er aca­dem­ics and polling com­pan­ies who acknow­ledge that it is troub­ling. For instance, stat­ist­i­cian Dr Andrew Balemi has been quoted by Col­mar Brunton as say­ing “the fact that not all house­holds have land­lines is an increas­ing con­cern with CATI [com­puter-aided inter­view­ing].  How­ever, in the absence of any sub­stan­ti­ated bet­ter tech­nique, CATI inter­view­ing remains the industry stand­ard.” That leaves a grow­ing num­ber of people (about 15% of house­holds) out of the poll res­ults.

That would­n’t be a prob­lem if this omis­sion were evenly dis­trib­uted. How­ever, while there’s debate about how sig­ni­fic­ant the bias is, land­line-free house­holds do tend to be young­er and less well off. I decided to see wheth­er there were also geo­graph­ic vari­ations, by map­ping the pro­por­tion of house­holds without land­lines.

Map of Wellington households without landlines

This is based on the some­what ambigu­ously-worded Access to Tele­com­mu­nic­a­tions Sys­tems ques­tion in the 2013 Census (offi­cially, the “tele­phone” answer is inten­ded to mean a land­line, though it could eas­ily be mis­in­ter­preted). With that pro­viso, some geo­graph­ic pat­terns are obvi­ous. In most of the out­er sub­urbs, few­er than a quarter of house­holds eschew a fixed line. This pro­por­tion increases in the inner sub­urbs, and in the cent­ral city, land­lines are an endangered spe­cies. Giv­en that this is the most densely pop­u­lated part of the city, this is sig­ni­fic­ant, so let’s zoom in to Te Aro for some more detail.

Map of Te Aro households without landlines

As well as shad­ing the meshb­locks with the pro­por­tion, this labels them with the approx­im­ate num­ber of house­holds that lack fixed-line phones (Stats NZ ran­domly rounds the data to base 3 for pri­vacy reas­ons). One single block has nearly 200 of these (that would be the Soho apart­ment build­ing), and look­ing at the num­bers it’s clear these would add up quickly. A large pro­por­tion of these homes would have more than one adult occu­pant, so there’s a whole com­munity of city dwell­ers whose voices are going unheard in the polls.

Some have argued that people without land­lines also tend to be the people who don’t vote, and that’s cer­tainly con­sist­ent with enrol­ment pat­terns sug­gest­ing that young, mobile people are less likely to enrol. There’s likely to be some influ­ence from the fact that Te Aro has a lot of new apart­ments: when mov­ing into a brand new place without an exist­ing land­line it’s easy to keep it that way, where­as people in long-estab­lished homes would have to make a con­scious decision to change. Settled house­holds with older inhab­it­ants in Karori, John­son­ville and Seatoun are much more likely to be polled than the young, dynam­ic pop­u­la­tion of the cent­ral city. Inner-city issues, pri­or­it­ies and aspir­a­tions go unheard, and the voice of the polls is the voice of the sub­urbs.

Of course, there is still one way to guar­an­tee that your voice will count, even if the poll­sters can­’t hear it. Vote.