Gage RR- Applicability to call monitoring agents
Six Sigma – iSixSigma › Forums › Old Forums › General › Gage RR- Applicability to call monitoring agents
- This topic has 6 replies, 5 voices, and was last updated 15 years, 1 month ago by
kapil.
-
AuthorPosts
-
May 26, 2007 at 10:19 am #47091
Hi
Wanted to seek an opinion on performing Gage R&R for call monitoring process. Currently we perform a calibration wherein different participants listen to a call and rate them against predefined parameters. How can we perform a Gage R&R for the call monitoring process ?
Thanks
0May 26, 2007 at 11:09 am #156593Dear Kapil..
Since the data coming from your study would be attribute type, we need to assess probability of false acceptance, probability of false rejection and Bias tems. Through which the we can assess the ability of participants rather using the term GR&R which would be more appropiate for continous data type.
If you are interested to on that process…pl mail me for more details.
Best regards,
Ramesh Patnana0May 26, 2007 at 11:26 am #156594Kapil,
start with operational definition of the parameters you have in the form.
Get a sample of 30 calls (random picked).
Make an expert to rate the calls (it will be a standard) using the operational definition.
Get at least two measuring persons to rate the sample calls twice, using the same operational definition.
Summarize cases where one tester rated same call differently (REPEATIBILITY) and cases where different persons rated same calls differently (REPRODUCIBILITY).
Goal is the obtain 100% agreement between call testers and expert.
Work on the parameters you dont have R&R, change the operational definition, and get them out of the form….
Actually I love you idea to do gauge R&R on call ratings. I will do it in our Call Center.
I am looking forward your answer.
Thanks
Silviu0May 27, 2007 at 1:38 pm #156620Hey Silviu !
Thanks a ton ! My problem is that the kind of industry i operate in, there are daily updates and these updates override the previous ones. Hence doing a Repeatibility study is kind of difficult.
Also the call quality sheet that we use (defined by client) carries a rating scale of 0 to 3 and there are no ground rules defined against these specific ratings. Hence chances of subjectivity in rating are high.
My struggle is to ensure accuracy of evaluations while i operate with this kind of rating scale.
Kapil0May 27, 2007 at 2:19 pm #156628
Jim ShelorParticipant@Jim-ShelorInclude @Jim-Shelor in your post and this person will
be notified via email.Kapil,
I had a data storage problem like that once.
I had the IT guys modify the system to download the daily information I needed to a CD before the computer was allowed to overwrite the file.
Then they gave the data CD to me and I could upload the data into a database that contained the data history.
With that database, I could analyze anything I wanted.
Have a good day.
Jim Shelor0May 28, 2007 at 3:27 am #156649
Anton JavierParticipant@Anton-JavierInclude @Anton-Javier in your post and this person will
be notified via email.Hi Kapil, here is what we are doing on call center GRR:
1 30 samples (15 known good and 15 known bad) were identified by representatives from quality, operations, and training. This is now determined as the expert rating or response (whether data or voice)
2 all data or voice “inspectors” rated the 30 samples using 3 runs.
3 results were tabulated using minitab GRR attribute. Voice or data “inspectors” who scored lower than 70 (or those who scored higher than 70 but whose CI is lower than 70) will be due for calibration with the training team.
Hope that this helps, Kapil.
Anton Javier
0May 28, 2007 at 5:12 pm #156675Hi Ramesh
Thanks for sharing your views ! Would surely want to discuss this in more details. Can i have your mail id please?
Thanks
Kapil0 -
AuthorPosts
The forum ‘General’ is closed to new topics and replies.