Inside the data

It is not often that you get to see the data being collected by the Australian Bureau of Statistics first hand. However, this year our family was selected (randomly) to be part of a survey (conducted monthly) on various house things. It is still going on but the first part was a 90 minute home visit interview that spent much of its time quizzing our children on various aspects of their lives. The initial results of that survey have just come out. Here is an analysis. There is little surprising. 5-14 year olds are using the Internet and more so where it is available. They use it at home and at school and the time spent is growing but not yet at TV viewing levels. Lots of the use is nominally for “education.”

Looking at the results, it is clear our kids where outliers. I suspected as much as they were interviewed. They are high up on the internet use (especially at home) and undertake a variety of activities. They also use educational sites — StarFall, Mathletics and BrainPop — regularly. But I suspect that web-based games and social games are the main activity. My 8 year old son, in particular, likes new stuff and is enjoying being the youngest person on Google Wave.

But at each step of the process I knew they were not asked enough. They asked all manner of questions regarding use but very little in the way of controls. For instance, how many computers were in the house? Who owned a mobile phone? Operating systems used? Parental use of computers are work and home. And all manner of other household information that would surely be of relevance including how much exercise did parents do during the week? It was painful because I knew that this was going to allow some description but not nearly enough for serious research.

The message is to the survey designers at the ABS is this. If you are going to take up that amount of time collecting data, go for it. Don’t hold back on questions and come up with a dataset that will really tackle some important questions related to policy. At the moment, we didn’t get anything here that would help us with the whole NBN value question let alone what educational and health consequences there might be from all of this.

3 thoughts on “Inside the data”

  1. Doesn’t it go back to why the data is being collected and what the policy/research objectives are? If these aren’t well thought out and the data gathering is being done merely to collect some interesting facts – then of course the survey and results will be very general.


  2. But andrew, you understand the problems created by  not asking enough.  He average survey respondent only understands the problem for her created by asking too much –  ie time and hassle.  The ABS understands both (at least, it ought to), but knows which one really hurts – ie the respondents’  problem.  No respondent, no survey.

    Real life surveys have to balance data richness with respondent burden and consequent response rates.


Comments are closed.