Hi /sci/
So I'm trying to work out the significance of some data.
To simplify it for myself I thought about it in terms of two house holds eating eggs.
Household A has 10 people, they eat 5 eggs a week; B has 4 people they eat 3 eggs a week.
On average A eats 0.5 eggs per person; B eats 0.75 per person. (I am right in saying B eats 0.25 eggs more than A per person?)
Going further, A eats 62.5% of the eggs, B eats 37.5%
A comprises 71.43% of the total people (A and B combined); B is 28.57%.
I am just stuck here. I want to show who is more likely to eat eggs based on these percentages. Is there a way to do that?
I know I have already shown that B eats 0.25 eggs more than A per person. But I am looking to show how much more likely B is to eat eggs than A.
Thanks for your help /sci/
>>363904
Not /sci/ /wsr/
Posted on /sci/ but saw the catalog, copied and pasted but forgot to change
No because u have to find common ground and then simplify so 4 people eating 4 eatin 3 eggs would actually be 1/3
>>363904
>Is there a way to do that?
No. B eats proportionately more, but why? There's no random variable information that would guide you towards any sort of probabilistic guess, unless you want to make more assumptions than warranted..
>>363914
Ok Thanks for that.
I am looking at crime rates between Sudanese and people born in Australia.
The eggs were just to simplify things a little.
How then does someone make a claim like:
"Police say Sudanese 44 times more likely to break the law?"
http://www.heraldsun.com.au/blogs/andrew-bolt/police-admit-sudanese-44-times-more-likely-to-break-law/news-story/8ca308022ba8fbbc4b89ed50504271c5?nk=c31510aeeb62276e3c15e5ee92475446-1502358507
Thanks again.
>>363935
The way the media uses it, "more likely" just means more often. But unless we're literally modeling every person as a random variable, it doesn't make sense from a technical perspective. Journalists and bloggers aren't scientists.
>>363935
They're looking at conviction or arrest rates. The differences between Sudanese and native Australians are almost certainly statistically significant given the numbers in that article, even if they're cherry picking.
>>363949
http://www.abs.gov.au/ausstats/[email protected]/Lookup/by%20Subject/4517.0~2014~Main%20Features~Country%20of%20birth~7
Here is ABS data of imprisonment rates. You can see that Sudanese have the highest rate (701.6 per 100,000 adult population born in Sudan).
I guess I am looking for data showing what percentage of a population commits crime.
The numbers shown in the earlier post can be reduced to crime PER PERSON. But it's pretty safe to assume that not every person in a population shares in committing crime.
So do the imprisonment rates reflect what percentage of a population commits (or has committed) crime. And could this be extended to say a certain population is MORE LIKELY to commit crime - especially coupled with the per person average...
Thanks