Geometric Mean
Six Sigma – iSixSigma › Forums › Old Forums › General › Geometric Mean
 This topic has 4 replies, 3 voices, and was last updated 17 years, 5 months ago by harry.

AuthorPosts

June 2, 2004 at 1:24 pm #35732
In the telecommunications industry, when is it appropriate to use a geometric mean? When measuring response time of a large circuit, should a geometric mean be used? (that is the current method of choice but I don’t necessarily think it is correct.) By the way, response times are checked 100 times every 5 minutes.
0June 2, 2004 at 3:06 pm #101116What are you trying to do? Please provide more information and we can help.
0June 2, 2004 at 4:49 pm #101128A group measures response times for circuits. How long (in milliseconds) does it take to ping, or get a response, from a circuit. It is an indication of the availablity of the circuit. Currently, they use the geometric mean to report on response times. Because response times are additive (New York to Dallas to San Fransisco, for example) and because there is a lot of data (28,800 pings per circuit per day), I don’t think a geometric mean is the appropriate way to measure response times. However, I don’t know enough about geometric means to know if I am correct. To me, a regular average would suffice for reporting hourly or daily response times. Make sense?
0June 2, 2004 at 5:13 pm #101129
Kim NilesParticipant@Kniles Include @Kniles in your post and this person will
be notified via email.Harry:
You might want to read this site at http://www.math.toronto.edu/mathnet/questionCorner/geomean.html where you will see how the geometric mean answers the question, “if all the quantities had the same value, what would that value have to be in order to achieve the same product?”. Since the arithmetic mean answers the question “if all the quantities had the same value, what would that value have to be in order to achieve the same total?”, one would tend to believe your situation is properly using the geometric mean as I understand it.
Good luck either way.
Sincerely,
https://www.isixsigma.com/library/bio/kniles.asp
http://www.KimNiles.com
0June 2, 2004 at 7:26 pm #101142Thanks for the reply. I visited the link and reviewed the notes on the geometric mean. I think the key here, at least where I am coming from, is that response times are additive, they are not multiplicative. The time it takes for a response to get from New York to Dallas to San Francisco is the total of those intervals, not the multiplication of those intervals, hence my objection to using the geometric mean. However, some Ph. D. told this group that the geometric mean was appropriate. I disagree and think it is nothing more than a way for the response times to look better on the month ending report. From a practical standpoint, hardly anyone reading the report understands what a geometric mean is anyway. I think it is the wrong use of a statistic and was wondering if I was correct. Thanks for your thougths.
0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.