BTDT

I was trying my best, but got too busy at work.Cheers, AlastairBTW – the up and down numbers can be done using Monte Carlo. There is no logical way of figuring it out. You have to use a known distribution too. (WestElec#3)

Mike:Seems this might be Leslie Martin, the past Manager of Client Development at the SSA, still in Scottsdale.The website went up in July 2009, so it might be a new thing for her. I wonder where her training material comes from?Cheers, Alastair

Stan:Probably General Domestic Appliances, a GE company in the UK. I hadn’t heard of them either. I found a reference back in 2000. Pretty topical reference for a 2001 post.Cheers, Alastair

BC:This thread is for you. It looks like this person wants a copy of your D2 spreadsheet.http://www.isixsigma.com/forum/showmessage.asp?messageID=159896Cheers, Alastair

Bucki:The confidence interval on the linear fit is the 95% C.I. for the location of the best fitting line. The prediction interval accounts for the extra noise in X. If you take a careful look at the plot, you will find about 95% of the scattered points to be within the prediction interval.You can’t really do much more than that. Have a look at…[Read more]

Jeff:We analyzed the call centre data for a bank a few years back. We found:- On-hold times followed normal queing theory.- Handle times were almost perfectly log-normally distributed. We could not see differences between operators until transformed the data. Once transformed, we could use control charts and ANOVA to spot differences. This also…[Read more]

Rupesh:Beta can not be printed out when conducting a test in Minitab or otherwise. It must be calculated separately using the sample size calculator.When conducting the beta calculation, you need to include a few more parameters that are not specified when conducting hypothesis tests.Cheers, Alastair

Peter:
Contact me at 6sigmaguru(at)gmail(dot)com.I will send you an Excel spreadsheet that does the normal probability plot and Anderson-Darling test.Cheers, Alastair

Eugene:The shape of the pH titration curve for your final mixture will depend very much on the buffering properties of the mixture near its final composition. Ask the chemists.Cheers, Alastair

Z-42:Have a look at this message and do some investigating before you make the plunge. While good, it might not be your best choice in the long run.Cheers, Alastairhttp://www.isixsigma.com/forum/showmessage.asp?messageID=158911

Zack:I have data going back a long way on the number of posts per day to the forum. The idea was to construct a case study to show how to identify a non-normal mean shift in data with seasonality and how the iSixSigma traffic related to the timing of the Google virus warning. The applications could include call centre volume, order processing,…[Read more]

Robert S.:I have a sticky on my screen for my latest motto, “Solvitor ambulando”, to use the Latin gerund. Arrogant? Moi?Cheers, Alastair

shibu:Get in touch with the people at Fuji Water, they have been doing Six Sigma for a while.Cheers, Alastair

SiggySig:Your data is not binary(high/low) it is continuous ($/KWhr)General advice?- Logarithmically transform your price data- Include daily average system wide load for California prices- Ignore ambient temperature- Capture weekly variation using the autoregressive structure of your models- Use dummy variables for holidays and long weekends-…[Read more]

Georgee:GE uses a variety of four and nine-blockers in their pitches. I thing the one you want shows different strategies for improvement sorted along two dimensions.Ease of implementation – easy and hard
Impact on customer – big and smallThis puts all ideas out for discussion, but ranks them into the four quadrants. Management will then decide…[Read more]

JimT:You are welcome. This is one of my bailiwicks. I have seen percentage data misused so often that I propose to never permit its use in a project.Cheers, Alastair

Darth:Yes, you can construct an I/MR using a column of data values expressed as percentages. The software will estimate the standard deviation by using the n to n-1 differences between the subgroup means. The result is that the estimate of standard deviation for constructing the control limits is for the n-1 subgroups. This does not include any…[Read more]

The reasons percentage data are not considered continuous are numerous:- For each subgroup you have no estimate of the within-group variation- Statistical tests for continuous data require an estimate of the within-group variation- The effect of subgroup size is ignored and can lead to manipulation of the resultsThe consequences of ignoring the…[Read more]

Sophie:Have a look at a two-factor ANOVA. For each student you will have one factor at two levels (syllabus type) and a second factor at four levels (learning modality). The third data value will be the score for that student in that group and that modality. For your 15 students you should have 120 data points.Cheers, Alastair

Robert S:Thanks for your post. I was away the day they covered that one.Cheers, Alastair