A recent report on media diversity in Australia relied on statistics that the Human Rights Commission has now walked away from. Yet it’s a different story from the Reserve Bank of Australia, which refused to give detailed answers about its dubious handling of those government statistics, writes Tim Palmer.
“TV news too white.” This was broadly the headline run by most media outlets when it covered a recent report about Australia’s media.
It was based on a “landmark study” by Media Diversity Australia and the report stated that the science proved just how underrepresented ethnic minorities are in television news and current affairs on-air personalities.
The key figure was that Anglo-Celts made up 75% of the faces appearing on TV news and current affairs yet only 58% of the population was Anglo-Celtic.
However, that 58% statistic is wrong. Moreover, a number of other “landmark” reports from the Australian Human Rights Commission (AHRC) pointing out similar issues in the ethnic makeup of the nation’s political and business leadership are also based on botched data. The AHRC has conceded its data was flawed, unlike the Reserve Bank of Australia, which we will come to.
It is pretty clear that there is a representation issue in Australia’s TV media. You only have to spend half an hour scanning the nightly news (particularly on the commercial channels) to understand that the faces there don’t fairly match the faces on the street outside.
Bad science; bad journalism
However, the issue is that bad science, and the bad journalism that doesn’t question bad science, always matters.
How, for example, did the report decide that 58% of Australia was Anglo-Celtic? Media Diversity Australia took that figure from a social report Leading for Change 2018 by the HRC. That study showed a sweeping lack of diversity at the top in academia, business and politics. But in its attempt to measure the level of under-representation, it used incorrect baseline statistics.
The Leading for Change 2018 report relied on a figure generated by the Race and Cultural Identity Employee Resource Group within the Reserve Bank of Australia.
That group was tasked by the AHRC to crunch the numbers from the 2016 Census and divide the population into four groups: Anglo-Celtic, European, non-European and Indigenous.
But they only used one statistic: the “Ancestry question”.
This was the first failure. The Australian Bureau of Statistics has published numerous cautions about using that data solely for this purpose, advising that it should be used with other variables such as parental birthplace, language and religious affiliation.
In fact, the ABS did away with the ancestry question for 15 years stating:
“The aim of the question was to measure the ethnic composition of the population as a whole. Evaluation showed that it was not useful for this purpose as there was a high level of subjectivity and confusion about what the question meant…”
Then the Reserve Bank took another statistically disastrous step.
Respondents are allowed to nominate up to two ancestries when answering the Ancestry question. In 2016 about 8 million Australians listed two ancestries. But if you count all the ancestry responses you are double counting those 8 million people and will therefore end up with more answers than there are people.
The Reserve Bank solution was extraordinary: where someone nominated two ancestries, it would only count the more diverse one.
So if you put down English and Sri Lankan as your two ancestries the Reserve Bank categorised you solely as Sri Lankan. For the Human Rights Commission purposes, you’d end up in the “non-European” basket.
A statistical blunder
At a stroke the Reserve Bank erased the Anglo-Celtic background of more than 2½ million Australians. And that is how they got to the figure of just 58% Anglo-Celtic.
Glenn Capuano from demographic consultancy .id the population experts, and previously a senior Census officer at the ABS, says his own census return exemplified the mistake.
“I listed two ancestries in 2016; English and Italian. I consider myself 99% English, but the RBA must have classified me as non-Anglo-Celtic European.”
The Reserve Bank says it manipulated the ancestry data in consultation with the AHRC, which has now conceded its use and the way the data was further manipulated was a mistake.
As a result of my questions, the AHRC has decided to change its methodology and will no longer refer to the Leading for Change reports in any presentations and speeches.
In answer to a series of questions it said, “The Commission acknowledges the limitations you have highlighted.”
It’s a very different response at the Reserve Bank, which refused to give any detailed answers to key questions about its dubious handling of government statistics which were published under its aegis.
The Reserve Bank won’t even answer other specific questions such as why it defined all non-Maori New Zealanders as non-Europeans.
If the Reserve can’t be accountable for the numbers it generates, we have a serious problem.
Why does this matter?
When I was executive producer of Media Watch, we did a story on a phoney statistic that had become a journalistic staple. The canard was that “every four days a farmer commits suicide”. I’ll spare you the details of why the figure was a fallacy but you can read the Media Watch report here.
That false figure drove mental health policy for more than a decade. Every time money flowed to help farmers based on that false figure, it was not going to another area that may have been in more genuine need.
In this current case there are many who don’t believe there’s an issue in diversity in our media. I think they are wrong. But when a report emerges that is based in part on false data, critics can use it to try to justify their stance against change.
Channels Nine and Seven were indignant at the Media Diversity Australia analysis. The report was labelled as shallow and unhelpful. And almost all the critics took issue with the way the study had tallied the ethnic breakdown of the journalists on TV.
There are difficult issues there to be sure. The report claims to have used reporters’ biographies online (and how many people state their ethnic background there?), their surnames and “visual appearance” among other factors, to work out ethnic background.
The idea of someone sitting down with a colour chart to scan footage (“he looks a bit Asian around the eyes”) has a distinctly Apartheid era South Africa feel to it and it’s certainly not very scientific. But it’s hard to come up with an alternative method for collecting this data. (The headcount also left out NITV, leaving its Indigenous staff somewhat chagrined).
But the other side of the problem lies with the scores of journalists, including Media Watch, who reported on the study without testing the figures and don’t plan to annotate their online story even though they now know it’s wrong.
Media must try harder
The real report card on Australia’s media should read: “Must try harder.”
As to the lack of official data? I feel quite comfortable with the fact that Australia has no solid measure of who is who ethnically. The mix is too complex and self-identification always makes for difficult measurement. Countries that have a precise handle on the ethnicity of their populations tend to use that information, whether it’s in the gerrymandering of Malaysia or the treatment of Uyghurs.
Finally, none of this means there isn’t a problem with the representation of ethnic minorities in the media, business or politics. Media Diversity Australia’s aims are important and the study was well intended. Its raw figures still stand and demand action from TV executives.
But the report’s baseline data was a bust and in the end junk data doesn’t help the cause.