July 03, 2020, 04:00:27 PM
Forum Rules: Read This Before Posting

### Topic: Absolute Standard Deviation  (Read 27187 times)

0 Members and 1 Guest are viewing this topic.

#### CopperSmurf

• Full Member
• Posts: 139
• Mole Snacks: +6/-4
• Gender:
##### Absolute Standard Deviation
« on: February 28, 2008, 03:27:39 PM »
I've tried looking in my textbook, google and my fellow chemistry friends, but there's nothing to define what the "absolute standard deviation" actually is. I need to know exactly what it is and how to calculate it because it is a part of my lab report. I know it has something to do with precision and I think it should be something similar to the standard deviation calculations.

Feeling very very lost and it is such a basic concept.

Thanks in advance to anyone willing to lend me a hand.
« Last Edit: March 01, 2008, 12:18:00 PM by CopperSmurf »

#### JGK

• Chemist
• Full Member
• Posts: 738
• Mole Snacks: +66/-19
• Gender:
##### Re: Absolute Standard Deviation
« Reply #1 on: February 28, 2008, 04:37:26 PM »
Experience is something you don't get until just after you need it.

#### CopperSmurf

• Full Member
• Posts: 139
• Mole Snacks: +6/-4
• Gender:
##### Re: Absolute Standard Deviation
« Reply #2 on: March 01, 2008, 11:34:56 AM »
I've tried those websites and I think I've an idea what to do, but I still don't know what absolute standard deviation actually is or how to calculate it. And I can't calculate it from proportions because I'm not given that. The problem I have is:
x = (23.843±0.008)·(14.34±0.02)/(5.77±0.07)
and it asks for the absolute standard deviation
I thought it would be done by taking:
sum of all (error/value)^2
and then take a square root over it all and I thought it was 0.012 but it's not right.
« Last Edit: March 01, 2008, 12:14:51 PM by CopperSmurf »

#### Arkcon

• Retired Staff
• Sr. Member
• Posts: 7360
• Mole Snacks: +533/-147
##### Re: Absolute Standard Deviation
« Reply #3 on: March 01, 2008, 12:14:10 PM »
Is there a clear definition of absolute standard deviation somewhere?

I for one, don't recognize the term, and Google searches don't help me either.  I know standard deviation, symbolized by the Greek letter Sigma, and the population standard deviation, symbolized by s.  I generally use the latter one.  I know relative standard deviation, which is s/mean, and is usually expressed as a percent.

I don't know if any of those definitions are the same as the one you have for absolute standard deviation -- you're really going to have to scour your textbook, the lab textbook, class notes, someone else's class notes, or ask the instructor.
Hey, I'm not judging.  I just like to shoot straight.  I'm a man of science.

#### JGK

• Chemist
• Full Member
• Posts: 738
• Mole Snacks: +66/-19
• Gender:
##### Re: Absolute Standard Deviation
« Reply #4 on: March 04, 2008, 02:44:46 PM »
I've tried those websites and I think I've an idea what to do, but I still don't know what absolute standard deviation actually is or how to calculate it. And I can't calculate it from proportions because I'm not given that. The problem I have is:
x = (23.843±0.008)·(14.34±0.02)/(5.77±0.07)
and it asks for the absolute standard deviation
I thought it would be done by taking:
sum of all (error/value)^2
and then take a square root over it all and I thought it was 0.012 but it's not right.

I think your first step should be to calculate the range of values available for X (calculate 3 data points):

(23.843 + 0.008).(14.34 + 0.02)/(5.77 - 0.07)

(23.845).(14.34)/(5.77)

(23.843 - 0.008).(14.34 - 0.02)/(5.77 + 0.07)

The first and third values represent the range of possible X values and the middle value is the median value.

Use the Wikipedia link in the previous post and you can calculate the absolute SD by at least 1 method.

Experience is something you don't get until just after you need it.

#### enahs

• 16-92-15-68 32-7-53-92-16
• Retired Staff
• Sr. Member
• Posts: 2179
• Mole Snacks: +206/-44
• Gender:
##### Re: Absolute Standard Deviation
« Reply #5 on: March 04, 2008, 11:01:12 PM »
Absolute Standard Deviation is just the traditional standard deviation. It is the average of the absolute value between the measurements and the average.

To solve this problem, you need propagation of uncertainty:
http://www.rit.edu/~uphysics/uncertainties/Uncertaintiespart2.html

To be clear, what I would define as Absolute Standard Deviation, I would do by deriving the same way the standard deviation function is; but instead of squaring it to take care of the problem of the sum of all values minus the average, averaging to 0, one could take the absolute value. But this does not makes sense for this question.

#### CopperSmurf

• Full Member
• Posts: 139
• Mole Snacks: +6/-4
• Gender:
##### Re: Absolute Standard Deviation
« Reply #6 on: April 04, 2008, 06:20:33 PM »
Thanks everyone for trying to help me out with this.

I eventually got a real answer from my prof after approaching him a few times and he eventually said that he wanted me to do a propagation of error, but he said it in a really weird way...