# Understanding variance in coaching activities

Six Sigma – iSixSigma Forums Old Forums General Understanding variance in coaching activities

Viewing 35 posts - 1 through 35 (of 35 total)
• Author
Posts
• #40273

Participant

Good Afternoon,
I work in a Call Center, use Minitab 13 (should I say am trying to learn to use Minitab 13)
I have 35 responses to a survey (total pop 58) which represent 7 coache’s teams.
Question was “Do you believe that there has been an improvement in coaching?”
Answers were Yes/No, I converted Yes = 1 and No = 2 in Minitab
Column C1T = Coach Name
Column C2T = Yes/No
Column C3 = 1/2
I am trying to figure out how to test the hypothesis that there is no difference in the agents perception between coaches (Note: number of responses by agent is not equal across coaches, ie there is not 5 responses per coach)
John

0
#124541

Darth
Participant

35 responses spread over 7 people is a bit small to draw any valid conclusions.  The appropriate test might have been Chi Square but you have already concluded that you will have less than five responses in a cell which Chi Square doesn’t like.  Furthermore, the format Mini likes would be seven columns for the coaches and two rows (yes/no) for the responses.  But, keep in mind, Mini will only accept 7 or less columns and that also includes your response column for yes/no.  This is under Stat/Tables. Bottom line is that with 35 responses over 7 coaches, you can’t tell any statistically valid difference.

0
#124542

Participant

Darth,
I had attempted a Chi Square and a few others none of which provided what I was looking for.
However, I did not know about the preferred format for Chi Square so thank you for that.
I suspected that the data points were few but as a newbie to Minitab I keep trying to use it to improve my ability
John

0
#124551

“Ken”
Participant

Hi John,
I caught your post along with Darth’s answer, and thought it would be great to enter myself into the data confusion contest.
Say listen, is there a particular reason why you decided to collect binomial, good vs. poor or 1 vs. 2, data?  It is usually more challenging to work with this data in categorical form, and not very discriminating.
If your up for a challenge, consider trying this survey again with 5 categories, as follows:
1- no improvement2- very little improvement3- some improvement4- very noticeable improvement5- excellent, no complaints
Using 5 categories in the survey provides a bit more discrimination in the responses, and moves the data closer to a “continuous like” form.  Notice to others, I said “continuous like.”  Here’s the twist, given the 5 cats of the data with the 7 coaches you might be able to evaluate the data using a One-way ANOVA to get an approximate view of differences between coaches.  The ANOVA module is reasonably robust to group distribution shape and variance differences among groups.
To the others viewing this post this is an unorthodox approach, but I’ve used this approach for years with great success…  Sometimes we can cheat if we know where to pull the right levers.
Ciao,
Ken

0
#124557

Dayton
Member

Ken,

How about a practical (as close to actual as you can be without divulging proprietary information) example of successful application of what you propose – inclusive of issue, data and analytics.   Thanks.
Vinny

0
#124558

“Ken”
Participant

Vinny,
Please speak English.  If you would like the mock file I prepared for John, then provide an email.  I gladly send it to you.
Ken

0
#124560

Dayton
Member

How was I not speaking English?

I asked whether or not youd provide an example of successful use of your unorthodox (in your words, not mine) approach.

I understand what you said you did and dont need guidance, mocked-up data, or a template to work through it, but you said you had used it for years with great success.

I simply asked for an example and, I thought, politely (in English) at that.    But, I admit to being a native Texan and then having spent many formative years in California so theres no telling how I actually come across.  Sometimes intermingling Yall with Dude is a little unsettling to others.

The most Id have done, if provided an example, is look it over and think about it and possibly ask why you chose that approach over another.   Not being mean or smarty pants  but like you and others, curious enough to ask.
Vinny

0
#124561

“Ken”
Participant

Hey Vin,
Sorry for the misunderstanding, it must of been the Euro-Texan twang distorted by the Inet that threw me!
Say listen, I understand your interest in assessing the validity of my approach, but this screen format is far too basic to convey anything more than a few verbal thoughts, and the occassional comic relief picts.  Additionally, it’s far too much work for me to prepare anything generic from my past efforts given the fact you’re not paying me a cent for my time.  So, what I’m willing to do is provide you some experiences on my past work using this novel technique via the forum, (which is really not too novel), and you can decide whether I’m FoB.
In a second approach, I would send you the Mintab file I prepared for John.  I prepared this file as a communication tool using his survey application which you already know about.  In this way we save a little time getting up to speed.  I prepared this file as a communication tool, and not to train anyone.  If you can knotch your ego down a bit this is the best alternative for you to understand my suggestion to John.
As a third approach, I could suggest what would be best called a “thought experiment” to evaluate the my analytical suggestion from a different perspective other than application or academic.  I’ve not done Gedunken work via the web before, so it would be a healthy challenge for me to convey thoughts using this media.  (this is my preference…)
SoVin, before I do anything–you tell me how you want to proceed with either of the three suggestions above.
Ken

0
#124565

HF Chris
Participant

Ken,
There is nothing unorthodox about this method. It is simply a likert scale. http://www.socialresearchmethods.net/kb/scallik.htm . It is useful if used correctly.
Chris

0
#124567

Dayton
Member

Hmmm what to do, what to do???..

Even though my ego is permanently notched and relatively unmovable, Ill attempt to overlook the fact that I have a vast and well-honed intellect and see what youve got.

So, lets do in an order of descending interest, options three, two and then one.   But lets, for my clarification, begin three with a mind mapping exercise that graphically depicts your logic and its analytical and interpretive paths.

Im familiar with Einstein’s gedanken (versus gedunkan, but danken or dunken its the thought that counts  get it? its the thought that counts??? anyway) experiment with photons and if your applied Six Sigma gedanken experiment demonstrating the relativity of seemingly misapplied statistical treatments is even partially as spectacular were in for a treat.

So lets gedanken first and if need be, go to option number two in which well attempt to otherwise get me up to speed quickly using the modeling template and data you sent John.

Number two failing, lets then revert to number one in which you can attempt to convey the logic of your approach via the admittedly mono-dimensional, limited and sterile media  the isixsigma forum.

Vinny

0
#124574

“Ken”
Participant

Chris,
Yes, your are correct about the scale I used.  But, it’s not the scale I refer to that is unorthodox–it’s the analysis I used to evaluate ordinal data.
Ken

0
#124576

“Ken”
Participant

Ok Vin,
I’m sure you’re having great fun with the shadow boxing.  I’m will gladly to spend more time with you tomorrow on this when my mind clear a bit after some rest.  BTW, you don’t get to run the show…  It’s one selection, and one only.  I don’t have much other time to play.  Sorry!
So, it a thinking exercise you would like?  Great!  I’ll try to keep it simple.  Not because of you…  We all know you are the smartest guy in the world.  Ops!  Except for your buds!  Anyway, let’s take a simple application.  Think of three models of SUV’s.  Most Americans like SUV’s, right!  Their wonderful for using up the last remaining fuel reserves of the world.  Sorry–got a little political there.  Anyway, think of three models of SUV’s that you will evaluate for gas mileage.  Can you see them?  Imagine driving each one for enough time to measure the gas consumption rate about 5 times.  All of the SUV’s are claimed to have similar weights, engine sizes, etc.  Just to keep this simple, the gas consumption rate is computed to the nearest tenth of a mile.  Does all of this make reasonable sense?
If so, once completing the data collection.  You would like to compare the three SUV’s to determine if there are any statistical differences between gas consumption rates.  Which statistical analysis is best to use for this comparison?  Why?
See, I told you this was a simple one…
Will continue tomorrow when my head clears.
Regards,
Ken

0
#124586

“Ken”
Participant

Vin,I’ve had a second thought about giving you a choice as suggested in my previous post. I don’t think you would appreciate my effort. So, I’ll leave you with my original offer… I can send you a Minitab v13 project file workup of John’s original problem. Since my last post, I’ve added a complete OLR analysis of the same data. The results support the ANOVA analysis as I would suspect, but OLR does not have the distribution and variance restrictions of ANOVA. Anyway, if you want the file for review, then I’ll need your email address. Otherwise, I guess you’ll just have to stay in suspense…BTW, some of my past applications of my suggested approach follow:1. Comparing injection molded polycarbonate discoloration across 5 different molders supporting process improvement work.2. Relative coverage of port bond in a IV Bag application supporting DOE improvement work.3. Investigating the quality of customer service by representative to improve the customer care process.4. Evaluating the extent of embeded particulate in a large sample of poly bags-comparison of groups US vs. Japan.5. Targeting ceramic package coloration in reference to the customer standard, (DOE process characterization)Most of the above measures ended up being categorical in nature, and were analyzed using the my suggested approach. While OLR could be used, most practitioners are unfamilar with Logistic Regression. This makes the uptake and understanding more difficult. Ciao,Ken

0
#124598

Participant

Ken,
Thank you very much for your reply and offer. I would greatly appreciate taking you up on your offer, as I am very new to Minitab. My e-mail is john dot fraser at canadapost dot ca
The reasons that I chose a Yes/No type question were as follows
Answering the survey was not mandatory and was done via e-mail. We were attempting to make it as easy as possible for the population to respond to the questions about the effectiveness of a Workshop that we piloted here, modeled after GE Workout. It was a pretty simple (first time for us so we did not want to add too much complexity) Lean type problem we were addressing
We wanted a very “clear” answer due to small number of agents in the population. A Likert scale might allow some fence sitting (selecting response 3) and might also have required some definitions regarding where to place response.
Lastly, I did not (a habit I must begin to develop) consider using Minitab for the survey output. However, recently my boss and I were mentioning that we have to find reasons to use Minitab to improve our skills, hence this exercise.
Thanks again,
John
PS, If you would be kind enough to look at post 76372 about Project Selection Criteria, I would like to get your thoughts, my boss (a Finance guy) believes it might be possible to develop a tool that reduces selection to numbers. I think that a human element will always need to make the final decision. Your comments would be great.

0
#124602

Dayton
Member

Ken buddy,

No, no, no.   You gave me a choice.

Im afraid I must insist on getting a good gedanken thinking exercise in logic.   And further since its my choice I want the Einstein gedanken as applied to photons used as a corollary to proving the relativity of seemingly misapplied statistical treatments used in profoundly unorthodox approaches when moving data to a continuous-like form.

While many would have found your references to me shadow boxing and thinking of myself as a smarty pants as being somewhat judgmental and borderline antagonistic on your part, I understand and accept that you need some time to clear your head as SUV gedankens can be mentally fatiguing and stressful for even the most experienced gedanken thinkers.

I also apologize for seemingly wanting to run the show on this, you are obviously the change agent and change agents always step forward when others step backwards, always look the beast calmly in its eyes, meeting its gaze and smiling slightly before the kill, right?   So, lets stick to your original offer and not flinch and claim increasing pressures of business imperatives are cramping your style and driving your attention elsewhere.    Lets go with option 3, then 2, and then, and only if need be, to option 1.   That way were not sequentially and in lock-step moving 3, 2, 1, we are using the options as default points only if each of the higher level triaged options fail to resonate with me, the gedankened.

And never fear, I deeply appreciate your effort.

Vinny

0
#124605

Darth
Participant

Come on Vinny, play nice with Kenny.  I think my first response was the most appropriate; 35 responses across 7 people won’t provide much value for distinguishing between them.
While a Likert scale might help, you still have the issue of 35 responses across 7 people so where is the gain?  Secondly, a 5 point scale still doesn’t give much discrimination.  A scale of 7-10 is preferred.  But, unless you get response across the range, you will still have pretty much a discrete set of data.  With a 5 point scale you will likely cover only 2-3 of the possible responses.  At least that has been my experience.  But again, only 35 response for 7 people.  Only real solution is to capture more data.  The data collection is obviously not important to the decision makers since it is voluntary and limited in scope.  If in fact it is voluntary, you now open up the can of worms of non-response bias.  Do those not answering the survey feel that same as those that do answer?  If not, is the measly 35 truly representative of the population of respondents?

0
#124607

Darth
Participant

Oh, by the way Vinny, who elected you the smartest of the bunch?  While it is probably the truth, being from Florida, I demand a recount.  We now know that Stan is the richest but half-witted, Carnell is the most traveled alcoholic, Phil is the funniest and I am a jerk but why are you the smartest????  I mean tie dyed t-shirts and melon headed girlfriends who resemble big footed bears……..

0
#124609

Dayton
Member

Hey lets not unravel as a forum.  No one said you werent fairly smart, being a PhD from a big private school with a football team and all, but come on.. cruising SoBe in a vett in this day and age of blood borne pathogens running rampant????    How smart could you really be?   And thats probably why youre usually irritable, many STDs directly impact neurological functions.

And I already did the recount.  I won again.
Vinny

0
#124616

BTDT
Participant

Yeah, but I have my own cartoon.BTDT

0
#124618

“Ken”
Participant

Vin,
I’ve enjoyed the tit-for-tat, and how well you’ve used my German mispelling as folly.  I figured that you would catch that one, and work it much as Stan and Darth do!  Considering you’ve been on the forum scene for such a short time, and haven’t managed to rub Stan or the others the wrong way I really wonder whose really behind the mask…  Perhaps I should address you by your more common screen name Stan.
In any event, the offer still holds.  Provide an email, and I can provide an example of my earlier suggested technique fully notated.  I believe I have Stan’s email.  Should I just send it to you using that instead?  I don’t really have any more time to play.
I never fear anything coming from my CRT or LCD screens, and I truly question whether you really appreciate anything I or anyone else has to say on this forum.
Ken

0
#124628

Dayton
Member

I can see that Ive offended you and possibly hurt your feelings or at the very least you appear to be miffed.   That was not my intent  no miffing intended.    I did intend to spar a few rounds with you as mutually we wound our way to a common understanding of what you were actually saying and my slight concerns regarding what I believed you were saying  but that was only because you seemed to know your way around the ring.
I know who at least the primary Stan is but have not met him personally  and I am quite convinced that, barring any multiple-personality disorder that Im not aware of, I am not him/them.
Like you, I only come onto this forum to learn and discuss a topic near and dear to my heart, Six Sigma.   Actually, I have been posting for over a year on this forum and have managed at one time or another to annoy, anger or at least frustrate most of the participants  including Stan, Darth, Mike Carnell, Phil, BTDT, Robert Butler, etc., etc., all of whom I have a acquired great deal of respect for..
But the main thing to remember and truly believe is that I enjoy discussing these technical topics and the occasional departure into the inane, but there is no lack of respect or appreciation of others intended.   Well unless the occasional idiot surfaces, but that certainly doesnt apply to you, who has worked in the medical products industry successfully applying arcane German logic to removing the potential for charred PVC specs impregnating sheet film used in the fabrication of flexible IV bags.
Besides, you are a tricky devil yourself with the intentional misspelling of that German word, anticipating that Id rise to the bait and attempt to have some fun with it.
Yes, Id like you to send me the data that you have several times promised to send me.   Please send it to my email address honoring my favorite dead Italian mathematician Evangelista Torricelli at [email protected]  (and none of that porn stuff, Im not all that high in my organization).
You would be disappointed, behind the mask is another famous dead Italian mathematician Vincenzo Viviani (Vinny) but no one nearly as accomplished as Stan.
Vinny

0
#124635

Darth
Participant

Vinny, the cat’s out of the bag so no sense in trying to fool the Forum any longer.  You need to admit that you and Stan are one and the same.  In fact, according to numerous prior posts, Stan and I are the same, hence the name Starth.  Since it is well established that Stan is really an amalgamation of five separate people ranging from Helga to Silicon Valley Stan to Rocket Scientist Stan this becomes really confusing.  I think it is just a warping of Cygerspace that is creating all these diverse and dysfunctional personalities.  In truth, Stan is really Reigle who hails from the planet V.  In fact, if you really analyze some of V’s posts, they are certainly similar to Vinny…ah Vinny…V…see the connection.

0
#124658

HF Chris
Participant

Does this mean you are all a split personality of someone we have not met? Or are ya’ll the ego, super-ego, and id…and who is controlling whom?
Chris

0
#124664

Markert
Participant

Chris, I’m sure that others will have the same inquiring mind that you have, so I have taken my secret New-VGB Bold Font decoder ring and peeled another couple layers off the onion.
By the way Darth, excellent work the last couple days, including this insightful history of a year in the life of iSixSigma.
Here it goes:
Vinny, the cat’s (Muffin, Vinny’s animal crainial adornment and family pet) out of the bag (Darth’s ex-wife, in fact just about every man’s ex-wife) so no sense in trying to fool the Forum any longer.  You need to admit that you and Stan (a borg-like collaboration of at least 4 persons unearthed last year while waiting for the big 1.5 sigma shift showdown, commonly refered to as the big debate) are one and the same.  In fact, according to numerous prior posts, Stan and I are the same, hence the name Starth. (Starth is a well documented name, do a keyword search for additional references) Since it is well established that Stan is really an amalgamation of five (it was four last year consiting of Olde Stan, Helga, also known as Girl-Doc-not-Doc-IQ160+Helga, commonly refered to as Helga, and the two west coast Stans) separate people ranging from Helga to Silicon Valley Stan to Rocket Scientist Stan this becomes really confusing. (Perhaps Stan now has a new Adult Stan in their group, or has hired an unpaid intern for summer weekend duty or while Olde Stan is changing one of the many diapers from BabyStan. Or perhaps Helga’s Babe of a Sister, now has honorary membership in the Sarcastistican Stan SuperCommunity) I think it is just a warping of Cygerspace (Michael Cyger of iSixSigma fame) that is creating all these diverse and dysfunctional personalities. (No kidding, what will it cost in therapy for our little group? Would Dr. Carl LaFong be available for this) In truth, Stan is really Reigle (Do a keyword search) who hails from the planet V. (V. Laxmanamaanananaannam, former forum contributor who vanished last summer after a couple of weeks and a trip in his spaceship to way over yonder after a well documented introduction to his new work function) In fact, if you really analyze (not to confuse with analize, often written by some of our forum members in error, ouch) some of V’s (Laxman) posts, they are certainly similar to Vinny (a name taken for some dead mathematican/statistician who determined the precise time it took for the earliest form of one hour photos to be ready, or for some other reason) ..ah Vinny…V…see the connection. (Be the connection).
So Chris, as you can see, it does take some forum insight to get the true message in Darth’s post. Is it better now?
By the way, are you “no socks Chris” from a month or so ago?

0
#124666

HF Chris
Participant

Phil,
I am the sock Chris (I ended up with my dirty sock in my mouth trying to correct Stan).  Speaking of that, does this also make Stan the annoy character? We still haven’t answered who is the id in the group.
Chris

0
#124668

Mikel
Member

Phil / Darth,
Excellent work as usual. More coherent than usual as well.

0
#124669

Darth
Participant

Chris, I think you got it.  Here is a poster I got out of Stan’s storage shed.  He used to hang it on his college dorm wall.

0
#124670

Mikel
Member

Darth, it’s time for remedial cut and paste sessions.

0
#124671

Markert
Participant

Dirty sock or not, you have turned out to be a valued member of the forum with an interesting perspective and attitude.
Glad to see you now have socks, cuz some here were commenting about those yellow, crack, fungus toenails; but perhaps GiantMedicalProducts Company has a product that Vinny may be able to get free samples on to help the problem.
Id stay away from the id.

0
#124672

Darth
Participant

This will prove to the naysayers that you are rich and successful.  I hope you and Mrs. Stan don’t mind me sharing this precious moment when I came to visit Chez de Stan.

0
#124674

scoobster
Member

don’t u have anything better to do?  how old are you any ways….12?  stop ripping off your latest client and do some work.

0
#124675

Markert
Participant

Yea Darth. Get them all out of bed and to the hotel and get them working. It isn’t enough that you work them 10-12 hours a day, they need to be working at least 16 hours each day, spending no time at home with their families and missing the evening news. Sleep, we don’t need no stinking sleep. Get them all on the phone and get a meeting room lined so mr.-lower-case-text-messaging-abbreviation here gets his wish.

0
#124703

Darth
Participant

Phil, Phil, wake up!!  I just finished waking up all the people I know in CA so we could talk Six Sigma.  Last night I was calling clients until midnight, regardless of their time zone.  Since I am up already, I can watch some cartoons and eat my Fruit Loops…..or was that prune juice and Extra Bran cereal.  I forget that I am not 12 anymore…in fact, I forget what day it is.  Phil, can little Stan and Vinny come out and play???  Pleasssssssse.  I’m feeling sleepyyyyyzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz.

0
#124705

EdG
Participant

Amazing how bright a star shines just before it goes out…

0
#124707

Dayton
Member

Dont listen to him Darth.

Besides hes probably really looking at Mars anyway, my science teacher told us that during this next month Mars will be the closest it will be to the earth at any time in the next 50,000 years or so.   In this next month Mars will appear as large and bright as the full moon  only redder.

I think Stans probably grounded again and I have this part timer at Burger King, but maybe after work.
Vinny

0
Viewing 35 posts - 1 through 35 (of 35 total)

The forum ‘General’ is closed to new topics and replies.