Big night for Australia’s Got Talent

Surely in this day and age, with the cost of electronics being so low (you can buy a damn DVD player for $20 bucks now?), why can’t we have more ‘people meters’ out in the population?

I know 100’s of people and have never met one who either watches the top ratings shows, (this one, Winners and Losers, etc) or they are afraid to admit it.

Why are TV shows celebrated or damned based on such a small number of boxes out there in households that we have no real way of knowing what community or spending patterns they have.

No doubt this should be the right forum to raise this as someone will know.

User ID not verified.

Sorry Bucks, but I can’t let your comment pass without reply.

First, last night the TV ratings (metro + regional) were measured based on an active sample of n=4,783 homes which accounted for n=12,293 people. There are other homes and panellists but for various quality control reasons that was the active sample last night. In ANYONE’S language that is a large sample. For example, it is many orders of magnitude larger than the political polls that successfully predcited the hung parliament.

Second, the DVD player market numbers in the billions worldwide. The People Meter market is less than 1 million worldwide – there simply aren’t the economies of scale to drive the hardware costs down. Plus, a people meter is just one component of a very complex system (though I gather that you favour $20 solutions) – the costs are way more than just the hardware.

Third, I am hardly surprised you’ve never met anyone with a people meter. Around 1 in every 1,500 households is sampled across the metro and regional markets. Those selected are not there to promote the system, but to allow their viewing habits to be recorded – in fact, they are instructed to NOT ‘broadcast’ the fact that they are part of the TV ratings panel, which follows sound research practice. So, you may very well – and probably do – have a home in your neighbourhood who is part of the panel … you just don’t know it.

Now I suspect that your next question will be to ask is 1 in 1,500 homes enough. Have you ever had a blood test? The average person has around 5 litres of blood in their body and the average blood sample is 5 milliltres – 1 in 1,000. But when the actual test is done, they only use a drop of the sample which is a much smaller sampling fraction. I know we’re talking about different ‘populations’ here withe different rates of homogeneity, but the analogy still holds – and blood tests are way more serious than TV ratings.

Finally, I would ask how many more ‘boxes’ would you want out there? Remember that in those 4,857 homes every TV set is metered and the average number of sets per home is a little over 2.0 – so we’re already talking around 10,000 ‘boxes’ out there. If we quadrupled the sample we’d halve the standard error (sample design is around +/- 3% so it would reduce to +/- 1.5%) but we would quadruple the cost. Is that really the best way to spend those incremental dollars – or should we be looking at ‘public-place’ viewing … or IDS viewing? I would humbly suggest the latter.

I hope this answers some of your questions and allays some of your concerns.

User ID not verified.