Lies, Damned Lies and Statistics
Sep. 21st, 2006 09:10 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
cheshirenoirI was good at Statistics in High School, and at Uni.
I could usually "look" at the information and intuitively pick up what kind of data I was looking at and what was relevant.
Which is why when I see abuses of Statistics (Or even just raw data interpreted badly) I get twitchy.
Why am I bringing this up? Well, as many of you know I work in IT. One of my jobs is to keep on top of the IT literature. These days, that means I read a whole bunch of Tech websites. One of these sites, previously a world class authority, has been slowly degenerating over the last 5 years to the point where, when reading their reviews, I will often draw quite different opinions to the reviewer, merely based on their data.
An example of their incompetence, is an insistence that peak performance is more critical than average performance. They always seem surprised when the product with the best average performance outperforms the product with the best peak performance in "real world tests". C'mon! This is year 11 maths!
I do wonder where they are hiring these people from? They have no need for centralised offices, as they are submitting to a completely electronic forum. Theyhavehad the reputation to get the best gear and the best people.
I wonder what went wrong?
I could usually "look" at the information and intuitively pick up what kind of data I was looking at and what was relevant.
Which is why when I see abuses of Statistics (Or even just raw data interpreted badly) I get twitchy.
Why am I bringing this up? Well, as many of you know I work in IT. One of my jobs is to keep on top of the IT literature. These days, that means I read a whole bunch of Tech websites. One of these sites, previously a world class authority, has been slowly degenerating over the last 5 years to the point where, when reading their reviews, I will often draw quite different opinions to the reviewer, merely based on their data.
An example of their incompetence, is an insistence that peak performance is more critical than average performance. They always seem surprised when the product with the best average performance outperforms the product with the best peak performance in "real world tests". C'mon! This is year 11 maths!
I do wonder where they are hiring these people from? They have no need for centralised offices, as they are submitting to a completely electronic forum. They
I wonder what went wrong?
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
no subject
Date: 2006-09-21 01:47 am (UTC)"Peak" by its very nature implies "out of the ordinary" or "something we're not going to be doing all the time".
"Average" implies just that - "this is the standard workload that this device is designed to maintain".
Whilst it is useful to know what a device can do when pushed to its limits, surely you are better off basing your decisions off what it is designed to do on a normal, day to day basis - its average performance.
Of course, having not done any formal studies into statistics, I could be completely off base, but that's just how I see it from an uneducated perspective. :)
no subject
Date: 2006-09-21 01:55 am (UTC)We can compare lists of which websites are bad. :P
no subject
Date: 2006-09-21 01:57 am (UTC)no subject
Date: 2006-09-21 03:06 am (UTC)*raises jaw from floor*
Year 11? Heck, plain common sense will do. Average implies a condition that is more common.
no subject
Date: 2006-09-22 02:48 pm (UTC)I stopped paying attention ages ago when it was pointed out they were Intel sympathisers. :P I had noticed the content had been increasingly watered down too.
At least that site run by some guy with initials A.L.S. still seems to be of readable quality...
no subject
Date: 2006-09-22 11:22 pm (UTC)Dans Data is good for a giggle when he updates.