Beware of statisics, they can be very misleading
- Thread starter JLM
- Start date
Which is why the parameters about what you want to show and exactly what you want to show should be clear.
If I said, "2 out of 3 rabbits in this hutch are brown" then obviously someone else cannot come along and say "2 out of 3 rabbits in the hutch are black" (or white) and remain credible.
What if the Hutch was a large hutch, which was later subdivided, and
more brown rabbits where in some of the divisions than others....
Can someone claim that they're successfully reducing the number of
brown rabbits in hutches A & C, though somehow the numbers have
increased in hutches B & D (formerly all one large hutch), so now they
need funding to study the situation so that they can also reduce the
numbers of brown rabbits in hutches B & D like has been accomplished
in hutches A & C? 8O :lol::lol::lol:
Hi Ron- You missed your calling, you could be teaching statistics and earning the big bucks..............:lol:
Then you'd have to specify which section referred to, right. Define the parameters
Then you'd have to specify which section referred to, right. Define the parameters
Shhhhhhhhh or before you know it some hare brained con artist will be trying to fly that scheme.
Then you'd have to specify which section referred to, right. Define the parameters
Define the parameters? That's my point. What if the parameters change over time,
to tailor the results of the statistics in the direction chosen? Don't get me wrong, I'm
not disagreeing with you, but just pointing this out.
Lets say you're studying something (whatever the something is) and you want the stats to
show that things are increasing (or decreasing...doesn't matter), and you take the
somthing that you're studying and divide that into several catagories, and label
them A-J lets say....and you come out with a set of numbers.
The next time the exercize is repeated but the results aren't to your liking (& no, I
am NOT talking about Global Warming), so some of the info in classification C is
reclessed as G, and some stuff from J is reclassified as B, and D has actually dropped
some but F has increased a bit...so some stuff from F is reclassified as B....
....and the statistics demonstratably prove that something has decreased.
Last edited:
Depends entire on what point I want to make or whose ox I'm trying to gore. If it was for the news to enlighten the general public in a meaningful way, I'd stay away from percentages and give raw numbers, ie Total population, total employable population, number unemployed, but whatever you give out (unless you are writing a book) you are not going to present ALL the facts, but at least with raw numbers what you are giving out shouldn't confuse anyone.
What do the raw numbers tell you about the trend?
If you want to impress on the viewers how the labour market is changing, then the percent change is a completely valid bit of information to convey the changes that are occurring. It's more helpful if they give you something to compare it against, but you didn't mention that, so I have to assume they didn't.
If it were me, I would probably have just went with what the rate is, the change from last month, and the # of newly unemployed for the month of October.
If you want to impress on the viewers how the labour market is changing, then the percent change is a completely valid bit of information to convey the changes that are occurring. It's more helpful if they give you something to compare it against, but you didn't mention that, so I have to assume they didn't.
If it were me, I would probably have just went with what the rate is, the change from last month, and the # of newly unemployed for the month of October.
What they gave was the percent increase over same time last year, and the number of new unemployed for the month of October.
What I took offence to in the OP was the minimizing of the unemployed in Calgary.
Ahh. Well I'm not sure why that's misleading.
The provincial unemployment rate has doubled in one year...so that's nothing to sneeze at. Like Gerry, I don't see how comparing it to other areas with their own economies really makes sense. Calgarians aren't competeing for jobs with Haligonians. So I don't see why they would care for such a comparison if they are out of work.
The provincial unemployment rate has doubled in one year...so that's nothing to sneeze at. Like Gerry, I don't see how comparing it to other areas with their own economies really makes sense. Calgarians aren't competeing for jobs with Haligonians. So I don't see why they would care for such a comparison if they are out of work.
Exactly and that's why stats are crap. They are constantly changed to suit the circumstances they need to suit. Believe in them if you want but they are not real. They are only as real as you want them to be or as someone wants you to believe they are.
Stats are not crap. If you're studying anything, and you make measurements, how do you know that one quantity is really larger than the other? How do you know it wasn't impacted by some unaccounted variable. Next time you measure that same parameter, it could be different.
Stats are crap when the methods and assumptions are crap. When the methods and assumptions are correct, statistics is immensely powerful. Manufacturers use stats to improve their processes. Scientists use stats to test hypotheses and to determine optimal treatments.
The edge Japanese automakers held over domestic for such a long time is due to stats. They require less variability in the parts and manufacture of their vehicles, and thus the product quality is better. This idea was pitched to North American automakers first, but they didn't go for it. Japanese manufacturers (not just cars) embraced this idea, and the result is high quality products.
Stats are definitely not crap.
Stats are crap when the methods and assumptions are crap. When the methods and assumptions are correct, statistics is immensely powerful. Manufacturers use stats to improve their processes. Scientists use stats to test hypotheses and to determine optimal treatments.
The edge Japanese automakers held over domestic for such a long time is due to stats. They require less variability in the parts and manufacture of their vehicles, and thus the product quality is better. This idea was pitched to North American automakers first, but they didn't go for it. Japanese manufacturers (not just cars) embraced this idea, and the result is high quality products.
Stats are definitely not crap.
Mark Twain summed it up nicely:
"There are three types of lies: lies, damned lies and statistics."
"There are three types of lies: lies, damned lies and statistics."
That statement is bullsh*t in context. My son informs me there has been a lot of Newfies working on the rigs in Alberta.