Thanks:  0
Likes:  0
Dislikes:  0

# Thread: Response Time Standard Deviation

1. ## Response Time Standard Deviation

Hello,

I am trying to analyze LoadRunner test results and facing difficulty in interpreting Standard Deviation values.
I understand Standard Deviation denotes variation in data values from the mean, i.e. a high value denotes that the data values are more scattered, but failing to interpret the metric in terms of response time.

For example, the transaction particulars are as follows:
Name: TR_ABC
Min: 1.158 secs
Avg: 4.037 secs
Max: 12.547 secs
90 Percent: 7.421 secs
Std Deviation: 2.31

What can I interpret out of 2.31 (here)?

Kindly suggest.

2. ## Re: Response Time Standard Deviation

If you really need to understand it you're going to need to get a book on Probability and Statistics (I think that's the course that I studied this in.)

For a rough and dirty, you can say that about 90% of your values lie within plus and minus 2 SD's.

(Don't remember the exact percentage but it's about that.)

4. ## Re: Response Time Standard Deviation

Thanks ptrussel_nc!

For exact figures, one SD below and above the mean value accounts to 68% of data values, and 2 SD's that to 95% of data values.
So for particulars shared above, can we say that 68% transactions observed response time between 1.727 (4.037-2.31) and 6.347 (4.037+2.31) ?

5. ## Re: Response Time Standard Deviation

Yes &amp; the wiki article is great btw

6. ## Re: Response Time Standard Deviation

I find the "90 percent" output metric to be most realistic for reporting purposes.

7. ## Re: Response Time Standard Deviation

not always and only in context, I've seen 90th%ile lower than average... strange but true!!!

8. ## Re: Response Time Standard Deviation

[ QUOTE ]
not always and only in context, I've seen 90th%ile lower than average... strange but true!!!

[/ QUOTE ]

I am also using 90 Percentile metric for reporting purpose, but in few reports I have seen that the 90 Percent value is far greater than the Maximum response time value. This sounds really wierd, but is true.

Is it the tool loopwhole or it can be rectified by making some changes in any settings??

9. ## Re: Response Time Standard Deviation

It is statistically impossible for the 90th percentile to be higher than the maximum (100th percentile)

10. ## Re: Response Time Standard Deviation

[ QUOTE ]
It is statistically impossible for the 90th percentile to be higher than the maximum (100th percentile)

[/ QUOTE ]

James, I too understand this; but I had this strange observation and this is the reason I shared this here.

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•
Search Engine Optimisation provided by DragonByte SEO v2.0.36 (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.