Originally posted by Thurisaz
12, 500 people in a popn mass of 20+ million people...

Do you know anything about significance testing and population samples?

If not, then you're not qualified to judge the accuracy of the survey results.

How many people do you think are surveyed for election polls, etc? It might surprise you, if you do some research.

Where were this samples taken. The local working class pub, lions club???

Originally posted by tezzajw

Originally posted by Thurisaz
12, 500 people in a popn mass of 20+ million people...

Do you know anything about significance testing and population samples?

If not, then you're not qualified to judge the accuracy of the survey results.

How many people do you think are surveyed for election polls, etc? It might surprise you, if you do some research.

Don't be condescending... I am qualified ty and it is the first thing I consider when interpreting data collected from research.

The findings are sensational... and if held to be a 'general consensus' is seriously inaccurate.

The sample size does not even equate to 1% of the popn.

Thurisaz, using your figures from a couple of posts above and assuming a random sample, here's some numbers that I just crunched. Note that I corrected for a finite population, even though it does make the calculations a little more tedious.

Standard Error = SQRT[ (1 - 12500/20434176)*(0.4)*(0.6) / 12500 ]
SE = 0.00438

A 95% confidence interval = 0.4 +/- 1.96*0.004338
95% CI = (0.395662 , 0.404338)

There you go, some basic statistics shows that there's a 95% chance that very very close to 40% of Aussies believe that migrants don't belong here.



If we wish to be 95% certain that our sample and population proportions differ by less than 0.01, then we can calculate the required sample size, n, as follows...

0.01^2 = 1.96^2 [ (1 - n/20434176) / 4n ]
1/n - 1/20434176 = 4(0.01/1.96)^2
n = 9600

Given that the sample size was 12,500 it is far more than what's required to show that there's a 95% chance that between 39% and 41% of Aussies don't think that migrants should be here.

If you think that the sample size is too small, then please supply me with your calculations to prove so.

Originally posted by Thurisaz
The sample size is too small, it is that simple.

The slab of text that you posted did nothing to refute my calculations.

Do you understand the formulae that I used? You claim that you are qualified to understand statistical sampling techniques, yet you supplied NO mathematics to support your claim that the sample is too small? Why?

The mathematics requires a sample size of 9600, to be 95% confident that the sample and population proportions differ by less than 0.01. I showed you the calculations.

Sure, there may be some minor errors, from a practical point of view, if the sample was not truly random. Any bias, or partially stratified grouping will throw a little error into the overall result - which would require a larger sample than 9600 to be as confident. However, I can take at face value that the survey was random, as it was stated in your quote.

Perhaps, instead of quoting text, you should read, understand and apply statistical techniques that can be found in common 1st Year University statistics text books.

Originally posted by Thurisaz
The main protection agaisnt this kind of error is to use a large enough sample.

I agree. I showed you the mathematics that determines a sample size of 9600 gives a fairly confident result. 9600 is a large sample of people. In this case, 12500 is an even larger sample.

It's a fundamental flaw that many people possess, when they use their 'gut feelings' with regards to statistics and probability. Luckily, we've got mathematics to save us, hey?

I know you don't believe me and you're trusting your 'gut feelings', so I advise you to show my calculations to someone who can understand them and have them independently verified. By all means, don't believe me, let someone else tell you that I'm right.

What do I care, anyway... I'm just a troll. At least I'm 95% certain of it.

Originally posted by purplemonkey
your mathematics means nothing when you cannot prove that the sample is from an accurate cross section of society...

Did you read the quote near the top of the page? The sample was random. Random means that people across all walks of life are surveyed, without any bias in their selection.

If there was any stratified groups included in the sample, then that could be adjusted for in calculating the required sample size.

Explain to me why they would call it a 'random sample' if they didn't sample randomly?

Originally posted by Thurisaz
don't refer to me as a troll thanks.

Please show me where I typed that you were a troll? Take a look at my unedited post and read through it all again. It might help you to understand what I typed, before you reply.

look, I was at work last night and read and quickly responded. I have come back to it today as I have the time to respond in more detail but I wont bother,

By all means, please respond and point out to me where the maths was in error.

Nothing about this poll is random. Starting by not including foreign looking people in the poll for example. That's where racism starts already.

[edit on 1-11-2008 by Benarius]

Edits on grammar

[edit on 1-11-2008 by Benarius]

Originally posted by Benarius
Nothing about this poll is random.
Starting by not including foreign looking people in the poll for example.

Where does it state that in the article? Please show me the data that supports this claim.

Originally posted by Benarius
Read the first sentence!
"Forty percent of AUSTRALIANS beleive.....". Now that should be obvious to you that this justifies my previous statement.

You're kidding, right?

Do you think that only people who live in Australia and were born here, with entire generational heritages of people who were born here, are Australians? What about children who were born from migrant parents, here in Australia, do you define them as being Australians?

You probably need to rethink your assumption. Australians are people who have Australian citizenship, whether by birth or by being naturalised.

Please for the sake of humouring me - who do you define as being an 'Australian' and who do you exclude?

[edit on 1-11-2008 by tezzajw]

Originally posted by tezzajw
What do I care, anyway... I'm just a troll. At least I'm 95% certain of it.

As if you would start a thread and then refer to yourself as a troll?

for goodness sake... take your indirect smoulderings and go for a polite walk, breathe in and perhaps consider how you are coming across.

My original imput here was quite simply to state the sample size was so small it cannot accurately represent the popn of Aust.

The sample size equates to 001%

so with that in mind and I am not saying you are wrong...I am saying for the last time, samples this size generally do not leave a big impression with academics or social policy.

So please don't read it and go off and think Australians' as racist bigots...

If you wish to discuss the research findings, at least try and keep it in perspective.

cheers

Australia racist? I cannot remember seeing a minority represented in an episode of Neighbours, unless you count that Irish guy

Seriously though, this isn't native to Australia. In Northern Ireland, Eastern Europeans have faced arson attacks and random attacks of Neanderthal mentality bigotry.

[edit on 1-11-2008 by infinite]

Originally posted by Thurisaz
As if you would start a thread and then refer to yourself as a troll?

Why don't you also take a look at my signature and my mood... At no point did I state that you were a troll. That was your mistake, not mine.

My original imput here was quite simply to state the sample size was so small it cannot accurately represent the popn of Aust.
The sample size equates to 001%

I take it that you have not yet had my calculations independently verified by someone who knows statistics?

so with that in mind and I am not saying you are wrong...I am saying for the last time, samples this size generally do not leave a big impression with academics or social policy.

Academics created the maths that determines the sufficiency of a sample size to accurately determine a population parameter. Some very smart mathematicians worked extremely hard to develop these techniques. Governments use surveys to determine policy, that's why they hire statisticians to determine if results are statistically significant, or not.

If you wish to discuss the research findings, at least try and keep it in perspective.

The research findings were clear - that 40% of Aussies don't believe others belong here. I kept it in perspective and used statistics to verify the claim of the report.

Originally posted by infinite
Australia racist? I cannot remember seeing a minority represented in an episode of Neighbours, unless you count that Irish guy

Yeah and if you remember the story line, he ended up getting the Asian girl pregnant. I can barely remember another Asian ever being on the show!

She ended up going back to live in New Zealand!

Neighbours, where white, 'Aussies' are supreme. Now that's a biased population to sample from!