The plural of anecdote is/is not data
One of the biggest problems facing anyone interpreting what is going on in politics is that they simply don’t have access to the most important research political parties and governments undertake, writes Noel Turnbull.
One of the biggest problems facing anyone interpreting what is going on in politics is that they simply don’t have access to the most important research political parties and governments undertake -- qualitative polling.
And it also raises a significant policy question: how to make taxpayer-funded research -- which is often pretty political in nature -- available to the public?
The qualitative research is extraordinarily important. The opposition’s recent 'all talk and no action' campaign against Kevin Rudd didn’t create changed attitudes to Rudd, merely reflected what their focus group research picked up. Similarly, everyone around the Keating government in 1996 knew, from the focus groups, that the public were waiting for the election with baseball bats. The Opposition knew the same thing and could, therefore, run a small target campaign with little risk. John Howard campaigned against Mark Latham, seemingly surprisingly, on the grounds of who you could trust because the public perceived what many of his colleagues knew -- that Latham was a brilliant but mercurial and unpredictable person.
The closest we generally get is the insights from Hugh Mackay’s articles when he shares some of the results from his long-running focus group research. Indeed, he was probably the first to pick up changing community attitudes to Rudd. Sometimes data about internal polling is also leaked but this is never reliable because it has, inevitably, been leaked for a purpose -- such as undermining a party leader.
Those working in the marketing and communications industry would also regularly sit watching focus groups through the mirror and probably, even if the research is about soap powder, pick up some insights into the zeitgeist.
But most of us don’t and we have to rely on other sources.
The prime one is the page after page, commentary after commentary, of analysis of the quantitative polling that the media and others do. This is important, and some of it is very good and reliable, even though the media rarely talk about the standard margins for error, and report as if shifts within the margin have more significance than the data can possibly support. For examples of this, just consult The Australian any time a Newspoll is published.
But you don’t understand what people are really thinking unless you can undertake detailed focus group (qualitative) research that explores more visceral reactions. It is this polling, not quantitative opinion polling, that is generally driving the poll-driven politician.
Qualitative research raises some interesting questions about the other sources of information about what people think.
Much media coverage of political events uses anecdotes to illuminate an event, for instance, going out in the street to get some "representative" vox pops in reaction to the fall of the PM. This is a bit like taking a survey of taxi drivers.
The standard argument against such vox pops quotes the pharmacologist, Frank Kotsonis, that "the plural of anecdote is not data". I regularly use the quote myself. Yet in an earlier version of the quote, from US political scientist Raymond Wolfinger, the quote goes "the plural of anecdote is data".
They are paradoxically both right. Kotsonis in that the media doesn’t give us enough of the anecdotes to make a considered judgement; and, Wolfinger also right because data in the social sciences ultimately does come from lots and lots of observations or anecdotes.
Another strategy is to reverse-engineer the statements politicians make, and seek to ask who they resonate with and why, thus uncovering the data the qualitative research would have revealed. But ultimately this is based on a lot of guess-work as well.
What you can’t do is talk to your friends because they probably think similarly to you -- that’s why they are your friends. Nor talk to your peers in the media office because they share the conventional wisdom with you.
Strangers aren’t much help either. When Jeff Kennett unexpectedly lost government it was fascinating to hear people say things such as "I just didn’t know other people thought like me", indicating our unwillingness to share contrary views with wider audiences.
However, while we are never going to see party political research, there is no reason why we can’t see taxpayer-funded attitudinal research on a freely available basis without having to resort to FOIs. Following various scandals involving government communications in the UK and Canada, two major inquiries -- the Phillis and Gomery Reviews -- pointed out major shortcomings in the two countries and recommended greater accountability and transparency. Making market research available was an issue they both discussed.
As Sally Young says in the book she edited Government Communications in Australia (Cambridge 2007): "without the benefit of (such) a formal inquiry to spur change in Australia, the blurring of partisan and government communications has continued to be an issue of concern."
Pending such an inquiry, one of the key reforms that would add to accountability and transparency here, while also cutting the nexus between partisan and government communications, would be to make the research that would help us understand why governments did various things freely available to us all.
*Ritual declaration of interest: the author is a contributor to Sally Young’s book and has also conducted and used research that blurs the lines.