Beware of statisics, they can be very misleading
- Thread starter JLM
- Start date
Perhaps you should reread the original post. I have two kids who work in Alberta, I know how bad the employment situation is. That wasn't the subject of the post.
Well, sorry JLM, but your point seems to have been hidden.
You seemed pretty clearly to be saying that you feel people are gobbling up sensationalist statistics regarding unemployment. If you were saying something else, I don't seem to be the only one not seeing it.
You seemed pretty clearly to be saying that you feel people are gobbling up sensationalist statistics regarding unemployment. If you were saying something else, I don't seem to be the only one not seeing it.
I think what's missing is that while employment levels in Alberta are terrible, there are other parts of Canada where they are even worse, like here in B.C. where all the core industries are down...........logging, fishing, mining and now construction, but what is misleading about the figures reported is that in B.C. we've always had a level of unemployment so percentage wise our figures don't look as gloomy. The fault lies in reporting percentages instead of raw numbers. As for your quote that you don't seem to be the only one not seeing it. Correct- the other one is Gerry.:lol::lol:
I took his point to be that a 'skyrocketing unemployment rate' isn't very meaningful, if it's starting from almost zero.
If you start with a low number, and you triple it, you're still far better off than someone who started with a high number and a lower increase.
Using figures like 'a 300% increase' is only used to sensationalize what is not, in the end, a big problem in comparison to other regions.
If you start with a low number, and you triple it, you're still far better off than someone who started with a high number and a lower increase.
Using figures like 'a 300% increase' is only used to sensationalize what is not, in the end, a big problem in comparison to other regions.
Exactly. TP. Why couldn't I have thought to put it in those words? Have a good day.
There isn't anything misleading about reporting the percent change over a period. One could equally say that the wording starting from almost zero is misleading. It's vague. Is 4 per cent close to zero, and 8 per cent not close to zero? How about 6 per cent?
If the rate jumps significantly in one region and not in others, then there is something noteworthy to investigate. Maybe the jobs in Alberta are a better indication of the strength of the economy. Maybe Alberta needs to prepare for a drop in provincial revenue. There's all sorts of scenarios one can consider when you have data available.
How would you report those findings JLM? Would you just report the raw numbers, like newly unemployed this month? Don't you think rate of change is important when you're discussing a period of time?
If the rate jumps significantly in one region and not in others, then there is something noteworthy to investigate. Maybe the jobs in Alberta are a better indication of the strength of the economy. Maybe Alberta needs to prepare for a drop in provincial revenue. There's all sorts of scenarios one can consider when you have data available.
How would you report those findings JLM? Would you just report the raw numbers, like newly unemployed this month? Don't you think rate of change is important when you're discussing a period of time?
Statistics can be misleading depending upon the parameters that they're
based upon. Here's an example:
On the weekend I was listening to a radio program, and the topic turned to
homelessness. The guest speaker was talking about an area in the city of
Regina (North Central) and she'd stated that there where 10,000 homeless
people in that neighbourhood.
The Announcer (it was a national talk show) stopped her, as though he was
based in (I believe) Quebec, he'd been to Regina and didn't see 10,000
homeless people in that neighbourhood (which he'd been through).
After some questioning, it turns out that the 10,000 person claim was a
tad misleading in that anyone 35yrs old or older that didn't hold a mortgage
on their home (renting) was considered homeless...and their family would
also cound as homeless...and if four young men shared a home (roommates)
but only one was on the lease agreement, then three of those men would be
considered homeless...and so on and so forth.
It doesn't mean that there aren't homeless people in that neighbourhood, but
the numbers where sure jiggered with for sensationalism.
based upon. Here's an example:
On the weekend I was listening to a radio program, and the topic turned to
homelessness. The guest speaker was talking about an area in the city of
Regina (North Central) and she'd stated that there where 10,000 homeless
people in that neighbourhood.
The Announcer (it was a national talk show) stopped her, as though he was
based in (I believe) Quebec, he'd been to Regina and didn't see 10,000
homeless people in that neighbourhood (which he'd been through).
After some questioning, it turns out that the 10,000 person claim was a
tad misleading in that anyone 35yrs old or older that didn't hold a mortgage
on their home (renting) was considered homeless...and their family would
also cound as homeless...and if four young men shared a home (roommates)
but only one was on the lease agreement, then three of those men would be
considered homeless...and so on and so forth.
It doesn't mean that there aren't homeless people in that neighbourhood, but
the numbers where sure jiggered with for sensationalism.
You got the picture Ron, for Tonington's benefit I am going to give another for instance In 2008 Town A had 100 people unemployed, Town B had 10,000 people unemployed. In 2009 Town A has 1000 people unemployed Town B has 20,000 people unemployed. Now if a reporter wanted to give impression that Town A was in dire straits all he has to do is report that unemployment has increased 1000% in one year, compared to Town B (which is far worse off) which has only increased 100%. I stand by my contention (are you paying attention S.J.?) that statistics can be misleading.
Gerry, JLM is simply pointing out the thing that many of us have in the past - Statistics are stupid and unreal. Plus - in most if not all cases, they are fixed to suit the person/company that needs them in their favour. Sounds like this one is simply a reporter wanting to sound intelligent.
You can't say that town B is worse off in your example. What if Town A has a population of 1500 working aged, and Town B has a working aged population of 50,000? Town A now has a higher percentage of unemployed workers, even though Town B lost an absolute greater number.
You still didn't answer my question. How would you report the job loss trend in Alberta?
Last edited:
Ron is right; when the parameters are well-defined and the error margin is defined, statistics are bloody useful in indicating many sorts of issues. Stats are hardly stupid and unreal.
For instance, statistics show that BC is the worst province in Canada for poor children at the moment. CBC News - British Columbia - B.C.'s child poverty rate still Canada's highest Is it stupid to point out how BC stands? Is the data unreal?:roll:
After collecting the data from satellites and other instruments and assembling this data tells us that the trend in the global climate is to rising temperatures. Is that stupid or unreal?
For instance, statistics show that BC is the worst province in Canada for poor children at the moment. CBC News - British Columbia - B.C.'s child poverty rate still Canada's highest Is it stupid to point out how BC stands? Is the data unreal?:roll:
After collecting the data from satellites and other instruments and assembling this data tells us that the trend in the global climate is to rising temperatures. Is that stupid or unreal?
Last edited:
Thanks Tonnington, you've just proved my point exactly, if you don't have ALL the figures you can't draw any conclusion (I forgot to cite in my example that both towns had the same population)
An interesting one would be Crime Statistics. If over time the parameters
of what is being measured is modified, the numbers can be manipulated
showing an increase or decrease based on who's wanting what funding.
For example, if people stop reporting crimes that the police just do not
even respond to....does that mean that there is less crime? Assuming
the statistics (data) are collected based on reported crimes. If your
BBQ is stolen (or bicycle, or whatever) and the police only give you a
file # without doing anything else, and your insurance deductable is
$700 (more than the value of the items stolen), did a crime take
place?
If automobile insurance companies change the deductibles from $500
to $2500...and thus less minor accidents are handled by insurance
claims and more are handled privately, does that mean there are less
accidents? Assuming the statistics (data) are gathered from the insurance
industry.
of what is being measured is modified, the numbers can be manipulated
showing an increase or decrease based on who's wanting what funding.
For example, if people stop reporting crimes that the police just do not
even respond to....does that mean that there is less crime? Assuming
the statistics (data) are collected based on reported crimes. If your
BBQ is stolen (or bicycle, or whatever) and the police only give you a
file # without doing anything else, and your insurance deductable is
$700 (more than the value of the items stolen), did a crime take
place?
If automobile insurance companies change the deductibles from $500
to $2500...and thus less minor accidents are handled by insurance
claims and more are handled privately, does that mean there are less
accidents? Assuming the statistics (data) are gathered from the insurance
industry.
Last edited: