Wednesday, 5 November 2014

I do wish journalists would look at methodology before quoting surveys

The Sydney Morning Herald on 29 October 2014 published an article containing this statement:

The level of trust in the Abbott government has soared in three months as public attention shifts from the budget to a heightened sense of nationalism in the context of national security, a new survey indicates. 
In the middle of the year only 26 per cent of people thought the federal government could be trusted to do the right thing for the Australian people.
At the end of October the figure had climbed to 36 per cent - the highest level of trust in the federal government recorded since 2009 by the Mapping Social Cohesion survey. 

What the journalist failed to note about this survey was that:

The target for the project was to achieve n=1000 completed questionnaire with respondents aged 18 and over, who were born in Australia and whose parents were both born in Australia.

Now the 2011 national census revealed that almost a quarter (24.6 per cent) of Australia's estimated population of 21.5 million people were born overseas, 43.1 per cent of the population (or 9.2 million people) had at least one overseas-born parent and 15 per cent did not have citizenship. Approximately 84.5 per cent of the population at that time would have been 18 years of age and over.

What this indicates is that the Mapping Social Cohesion survey did not include the possibility of canvassing the opinions of large section of the Australian community.

Even living in regional New South Wales as I do, it would exclude a good many voters in the town in which I live.

All of which changes the weight readers might have given to this newspaper article if they had realized the demographic limitations it contained.

No comments: