The ABC’s Predictable Port Tack

The ABC is  campaigning to defeat the Baird  government in the NSW elections on March 28, with one of the principal weapon being the insidious Vote Compass project. Supposedly based on ‘scientific’ survey questionnaires, it produces meaningless, if not fraudulent, results that can be presented as ‘news’ in TV and radio bulletins.

Vote Compass has a 30-question survey on the ABC’s website. The latest analysis of responses to the survey produced results the Labor faithful might have ordered. Fifty-five per cent of respondents disagreed with the proposition: “Should New South Wales lease its electricity transmission network to the private sector?’  Thirty-one per cent agreed it should; 14% either didn’t know or had no opinion.

The question of privatising half the transmission network – the so-called ‘poles and wires’ which also includes the important sub-stations – has been identified as the critical issue of the campaign. The Liberal/National Party government plans to use the proceeds to fund extensive road and rail infrastructure. The Labor opposition rejects privatisation, offerring a lesser programme of development from the revenue stream from electricity distributors.

The NSW 7pm TV bulletin on Tuesday, March 11, demonstrated just how the ABC employs Vote Compass to support the Labor/Green opposition. Presenter Juanita Phillips cast viewers’ minds forward to the desired conclusion with her opening (the emphasis is mine):

PHILLIPS: “Premier Mike Baird has staked his leadership on the privatisation of the poles and wires. Now a new report shows the government might be struggling to sell the plan to wary voters.”  (My emphasis).

The ABC’s Vote Compass survey finds that one-third of Liberal voters are opposed to it, and the biggest opposition is in regional areas.”

Reporter Claire Ayres deftly took the pass, and began:

AYRES: “The message isn’t resonating with the majority of voters. Vote Compass asked: ‘should NSW lease its electricity transmission network to the private  sector?’.Fifty-five per cent said ‘no’. The most opposition comes from Labor voters but a quarter of Liberal voters are against the plan”. (The actual percentage was 26%).

Anthony Green, the ABC’s election analyst, is a respected psephologist. Cynically, he was then inter-cut to make a brief point, anodyne in itself, but so obvious that he was being misused to add political weight to the issue of the government’s problems:

GREEN: “And the task for the government in the election campaign is to solidify  those people, make sure they don’t wander away, and vote for somebody  else on that issue.”

Reporter Ayres then explained that “There’ll be no change in delivery (of electricity) to regional areas…” but this didn’t deter her from going on… “but opposition from there is amongst the highest in the state.” That remark allowed the presentation of a graphic from the survey showing the supposed opposition to privatisation: Country 60%; Outer Sydney 55%; Inner Sydney 51%;  Hunter/Illawarra 60%.

Next came the loaded vox pop: two Broken Hill residents and one from Dubbo  against privatisation, one unidentified man in favour.

The report concluded with the invitation:

“To find out where you stand, complete the survey at www.abc.com.au/votecompass.”

I would have thought that most people knew where they stood, but the ABC is keen to use this invalid survey device to bend minds and move hearts a little to the left.

No survey questionnaire has statistical validity if the sample self-selects. It is recognised that responses from volunteers skew results in favour of those with a vested interest in participating. In the case of the ABC, which has progressively betrayed its independence and impartiality over many years, the results are meaningless in terms of reflecting the broad electorate’s sentiments and leanings.

A further objection is that the sample size (the number of responses or so-called ‘users’) has not been disclosed, only response percentages. Nor has the margin of error been revealed. The design of the questionnaire form is also questionable. On questions asking “How much?”, the choices range from “much more” to “much less”. On questions such as the poles-and-wires issue, the options range from “strongly disagree” to “strongly agree”.

For each question, the mid-point on the scale is a “neutral” response. Vote Compass claims to use Likert-type scales, but a Likert-type scale in a well-designed survey has no such mid-point, forcing a respondent to choose in one direction or the other. High percentages of “neutral” or “don’t know”  — two options in Vote Compass — can reduce the number of meaningful responses, thus distorting the results. The news report did not separate “neutrals” and “don’t knows”.

The Vote Compass website publishes what appears to be an explanation of its methodology, but it steers well clear of the sampling issue. Instead it launches into a mathematical explanation of what it considers most important – deriving a correlation between responses to the questions, party policy and party affiliation.

“Once a user has answered Vote Compass’s attitudinal and policy-related questions, his or her position in the political landscape is calculated by averaging the questions on their given dimension…..Party or candidate positions are calculated in the same way, given each party’s or candidate’s answers to the same questions answered by the user.”

A voter ‘using’ Vote Compass is first confronted with a list of 30 questions on public policy issues. Many of these questions are not phrased in neutral terms,  e.g.:

• The government should fund walk-in treatment centres run by  nurses.
• Cuts to government services have gone too far.
• Ownership of Sydney’s Goat Island should be turned over to the local aboriginal community.

The voter is then asked to answer a series of increasingly personal questions. The validity of the survey and its attempted correlations is now further reduced by the option of skipping these sections. So, how much reliance can be placed on identification of political affiliation and support from the following questions?

Regardless of the party you intend to vote for in this election, in  general how likely are you to support each of the following parties?

or

How trustworthy do you find each of the following party leaders?

or

How competent do you find each of the following party leaders?

or

How likely do you think it is for each of the following parties to win in your electorate?

What follows is another (optional) probe to find (among other things) the respondent’s voting intention, sex, age, education level and employment, religion, usual news sources, ethnicity, income level, and self-assessment on a left/right political scale (again, non-Likert).

The end result of this interrogation is a statistical assessment, placing the voter as a dot on a two-dimensional grid. The two top squares are in the ‘Social Progressive’ zone, the lower two, ‘Social Conservative’. The two squares to the left identify ‘Economic Left’ and those to the right, ‘Economic Right’. These terms are not defined, yet algorithms determine the voter’s beliefs.

So where did this Vote Compass idea come from? The ABC News Federal Election Producer, Gillian Bradford described it as a good idea, thus:

“While ABC News must report on the machinations of this political contest; we must also deal with an election as a bid to govern the nation in all its complexity. In our coverage we will be seeking to get that balance right,  acknowledging a need to do more on policy issues and voter engagement”.

That was in May, 2013, before the Federal election that year, when something had to be done to counter the Abbott ascendancy.  Note that concern for “governing the nation in all its complexity” had become an ABC responsibility, although not mentioned in its charter. “Policy issues” is shorthand for “things we care about”. “Voter engagement”  is code for “cranking up the opposition”.

Vote Compass was developed by the University of Toronto, Canada. It was taken up by CBC, Canada’s national broadcaster, like the ABC committed to ‘progressive’ policies, and used in federal and provincial elections.  For the ABC it is co-sponsored by the Universities of Melbourne and Sydney.

The ABC’s concern for the preservation of democracy in Australia is touching. Analyst Anthony Green again:

In the real world of democracy, we know voters come to polling day with only limited knowledge of policy specifics, and also with a set of pre-existing views on different parties, leaders and candidates.

Vote Compass is here to help – and how!

Geoffrey Luck was an ABC Journalist from 1950 until 1976