Poll of the Day > An example of why percentages aren't reliable (without raw data)

(+) Yeah! (+)

So, last night here in Australia a lot of news channels were yabbering on about some 'national survey' that's results were recently analysed. The survey was about how much do you spend on your dog, as opposed to yourself.

Raw data for this survey has not been shown/said on any of the news sites or channels, and whilst I'm aware 'national surveys' don't mean 'everyone got asked' it's usually nice to know if 'national' actually means a large number, or just a survey that had at least one person from each state.

So, according to the survey about thirty something percent of "all" Australians spend more money on their dogs than on themselves. (and makes a single mention that this only includes expandable non-contracted money spent on TV, entertainment and technology but ANY money spent on/for the dog is counted). In fact, it says of this thirty percent, most are spending more than triple on the dog(s) than themselves.

The survey said that "more than 90% of people said owning a pet was good for their mental health" and that nine out of ten people agreed it made them less lonely (as opposed to 90%?)... Now, the survey is supposedly 'national'. The issue here is that the survey is implying at LEAST 90% of Australians own or have owned a pet. Which, I don't have any data about this myself, but think that's wrong. (And again, I'm aware the survey wasn't everyone, I'm just making a point here).

Lastly, the survey said the average monthly cost spent on the dog(s), is $115. So... Look back on the stat about 30% of Australians spending more on the dogs than themselves, and that most of this 30% are spending three times as much on the dogs.
Now I know averages can be skewed by extra high or extra low numbers, but ignoring that for now. This seems to imply that some Australians are spending less than $38 on themselves monthly. This would be a good time to mention costs in Australia are generally higher than that in America (GST, conversion, etc).

Maybe I'm looking too deep into this, but felt that the way the results were shown is either incredibly skewed by the ones surveyed, or purposely vague enough for a good headline. (Which most were akin to "majority of Australians living ruff to pamper their beloved pets")


Oh, and the TV channels that covered it only interviewed people who get pedicures, perms and the such for their dogs frequently.

inb4 "who cares"
---

tl;dr, but all I learned from my stats class is that you can make numbers look however you want them too.
---

CountessRolab posted...

tl;dr, but all I learned from my stats class is that you can make numbers look however you want them too.

That's my point. XD Stats are (mostly) useless without actual raw numbers.
---

(without raw data)

So did you find a source for the survey they cited?
---

---

RCtheWSBC posted...

(without raw data)

So did you find a source for the survey they cited?

Nope. Not even a name.
None of the articles link to any sources. Some of them quote some La Trobe University person, but they don't say if the survey was done by them or not.
---

CountessRolab posted...

http:/ /i.imgur.com/mo03jbo.jpg

That seems like illegal levels of misleading. XD
---

TheGreatNoodles posted...

RCtheWSBC posted...

(without raw data)

So did you find a source for the survey they cited?

Nope. Not even a name.
None of the articles link to any sources. Some of them quote some La Trobe University person, but they don't say if the survey was done by them or not.

That's a shame.
---

CountessRolab posted...

tl;dr, but all I learned from my stats class is that you can make numbers look however you want them too.

As long as they're liberal who cares.
---

Veedrock- posted...

CountessRolab posted...

tl;dr, but all I learned from my stats class is that you can make numbers look however you want them too.

As long as they're liberal who cares.

lolwut
---