Statistics

Longdoggy · 06-25-2014, 12:57 PM

I see quite a lot of statistical claims on the forum which makes me question... Is this accurate?
I've always learned that 92% of all statistics are made up on the spot.
Food for thought on a Wednesday afternoon.

EMITTR · 06-25-2014, 01:16 PM

I challenge your 92% and er uh provide my Z table correction of 29% as a correct statistical proof. Based on my Z table analysis of course.

Longdoggy · 06-25-2014, 01:18 PM

Quote:

Originally Posted by EMITTR

I challenge your 92% and er uh provide my Z table correction of 29% as a correct statistical proof. Based on my Z table analysis of course.

Hmmm, I think you are right sir.

Myrder · 06-25-2014, 01:23 PM

lacking any sources cited, most of the things said on this forum are fluff. This excludes people who do first hand research and provide their information/methods/data.

Welcome to NBA Jam · 06-25-2014, 01:23 PM

Well if we want to get more accurate, we could look at the misclassification rate for binary or categorical claims, and the absolute percent error for continuous claims...

EMITTR · 06-25-2014, 02:36 PM

Quote:

Originally Posted by Longdoggy

Hmmm, I think you are right sir.

LOL!!! Glad you have a sense of humor!

EMITTR · 06-25-2014, 02:37 PM

Quote:

Originally Posted by Some guy somewhere

Well if we want to get more accurate, we could look at the misclassification rate for binary or categorical claims, and the absolute percent error for continuous claims...

Hey...I don't see that in my Z-table!

Longdoggy · 06-25-2014, 02:40 PM

I could not resist, I knew this thread would get some of us stat nerds going.

GB-335xiCoupe · 07-14-2014, 12:02 AM

Quote:

Originally Posted by Longdoggy

I could not resist, I knew this thread would get some of us stat nerds going.

Stats can always be skewed, depending on your mathematical calculations and (especially for surveys) the exact wording of the questions can massively bias the "results" to a specific viewpoint. It's like a product claiming it is "50% more [insert your desired feature] than the leading brand". First off, what leading brand? And if this product is so much better, why isn't it the leading brand.

Here's the one I saw that I'm thinking, WTF does THAT mean?
"Use up to 4X less" and then in micro print "than the leading brand".

What? So say I normally would use 4 of these items. Does that mean with this brand, I will use 16 less of them? So when I use it, I actually gain 12, instead of consuming 4? See what I mean? Nonsense.

06-25-2014, 12:57 PM	#1
Longdoggy Master Bates 80 Rep 495 Posts Drives: 2020 BMW M340 xDrive Join Date: Jun 2014 Location: New York iTrader: (0)	I see quite a lot of statistical claims on the forum which makes me question... Is this accurate? I've always learned that 92% of all statistics are made up on the spot. Food for thought on a Wednesday afternoon. __________________ CURRENT: 2020 G20 M340 xDrive — Carbahn Tune RETIRED: 2011 E90 335i xDrive — Cobb AP3, PE Exhaust, Alpina TCU Flash, VRSF Charge Pipe 2017 Estoril Blue X1 M Sport — JB4
	Appreciate 0 Tweet Quote

06-25-2014, 01:16 PM	#2
EMITTR Lieutenant 95 Rep 442 Posts Drives: 11 335is Coupe BSM Join Date: Apr 2013 Location: US-CA iTrader: (2) Garage List 2011 E92 BMW 335is ... [9.50]	I challenge your 92% and er uh provide my Z table correction of 29% as a correct statistical proof. Based on my Z table analysis of course. __________________ '11 BSM DCT 7 Spd. 335is \| RPi \| aFe Magnum Flow Direct Fit \| Interad DTM 10 Forged \| Smoked Reflectors \| Huper Optik Tint \| Arkym Rear Diffuser \| Cyba CF Quad Tips \| iND CF CRT Rear Spoiler \| COBB Protune \| Turbosmart BPV \| AutoCarbon CF Grills \| Silver CF e-Brake \| BMW CF Mirror Shrouds \| M Perf. Steering Wheel v2 \| KW DDC Coilovers w/WLAN \| Interior Xenon LEDs \| Custom 335is Accents \| F/R Fenders Rolled & Pulled \| New Pipes Baffle Delete
	Appreciate 0 Quote

06-25-2014, 01:23 PM	#4
Myrder Major 162 Rep 1,264 Posts Drives: 2010 E92 335i MSport 6MT LMB Join Date: Apr 2013 Location: WildWest iTrader: (1)	lacking any sources cited, most of the things said on this forum are fluff. This excludes people who do first hand research and provide their information/methods/data. __________________ 2010\|335i\|LMB\|E92\|6MT\|MSport\|Logic7\|335is Clutch\|AE Performance\|BMS\|Walbro\|VRSF 7"\| 149.7mph NFZ AZ 1/2mi 1992\|Pontiac Firebird\|Mild 355ci\|T56\| -I will look on your treasures, gypsy. Is this understood?-
	Appreciate 0 Quote

06-25-2014, 01:23 PM	#5
Welcome to NBA Jam Is it the shoes!? 3884 Rep 5,112 Posts Drives: (Sold) 2011 E92 335i ZMP 6MT Join Date: Mar 2014 Location: USA iTrader: (1)	Well if we want to get more accurate, we could look at the misclassification rate for binary or categorical claims, and the absolute percent error for continuous claims...
	Appreciate 0 Quote

06-25-2014, 02:40 PM	#8
Longdoggy Master Bates 80 Rep 495 Posts Drives: 2020 BMW M340 xDrive Join Date: Jun 2014 Location: New York iTrader: (0)	I could not resist, I knew this thread would get some of us stat nerds going. __________________ CURRENT: 2020 G20 M340 xDrive — Carbahn Tune RETIRED: 2011 E90 335i xDrive — Cobb AP3, PE Exhaust, Alpina TCU Flash, VRSF Charge Pipe 2017 Estoril Blue X1 M Sport — JB4
	Appreciate 0 Quote