Correction: read that * as "root mean square" i.e. square root of mean square 
of the deviations from the mean value.

---- Original message ----
>Date: Sun, 25 Jan 2009 22:09:29 -0600 (CST)
>From: <[email protected]>  
>Subject: [USMA:42593] Odometer Precision and Accuracy  
>To: "U.S. Metric Association" <[email protected]>
>
>
>Stan,
>
>Here is a further clarification of precision and accuracy.
>
>>From a set of measurements of the same quantity under the same circumstances, 
>>each measurement deviates from the mean value of the measurements, and the 
>>square root of the * mean of the deviations is called the standard deviation 
>>from the mean as a measure of the precision of the measurements.
>
>Whether an odometer *displays* decimal values to one, two, or three digits (to 
>an *apparent* accuracy of one meter) does not determine either precision or 
>accuracy of a measurement of distance by an odometer.
>
>A set of a reasonably large number of readings of the same distance by the 
>same odometer is necessary to determine the precision of the odometer (as 
>mounted in its carrier vehicle).
>
>The mean value of the odometer measurements (and its standard deviation) 
>compared with an independent measurement of the same distance (perhaps by a 
>laser device) then can be used as a measure of the comparative accuracy of the 
>odometer mean value of the distance.
>
>This may seem complicated, but these procedures are necessary to realize the 
>full requirements of precision and accuracy   
>with respect to an odometer.  A single reading does not determine either 
>precision or accuracy.
>
>Gene.
>
>---- Original message ----
>>Date: Sun, 25 Jan 2009 18:50:45 -0500
>>From: "STANLEY DOORE" <[email protected]>  
>>Subject: [USMA:42568] Re: An Associated Press article in today's Atlanta 
>>Journal-Constitution  
>>To: "U.S. Metric Association" <[email protected]>
>>
>>   Pat et al:
>>    
>>       Let's clarify the distinction between accuracy
>>   and precision. 
>>       Two decimal place odometer readout is precision
>>   but it says nothing about how accurate that readout
>>   number is.
>>       Stan Doore
>

Reply via email to