Question 1 (Answer)
Dear Abby-Normal,
We are working on Online system to view defect counts by using Box & Whisker Plot. During this online system designing, I need to provide the formula to our IT programming. Can I get some confirmation from you regarding the Formula calculation for Quartile 1 & Quartile 3 : Should I follow Method 1 or Method 2 as below? Let says we have 10 numbers 1,2,3,4,5,6,7,8,9,10.
Method 1: to determine Q1 and Q3 using % of cumulative sum: Using this method: Median = 5.5, Q1 = 2.5, Q3 = 8.5, IQR = 6, extreme min = 2.5 - 1.5(IQR) = -6.5, extreme max = 8.5 + 1.5(IQR) = 17.5
Method 2: to divide the data into half, Q1 is the median for first half and Q3 is the median for second half: Using this method: Median = 5.5, Q1 = 3, Q3 = 8, IQR = 5, extreme min = 3 - 1.5(IQR) = -4.5, extreme max = 8 + 1.5(IQR) = 15.5
Signed: Quartile Methodless
Question 2 (Answer)
Hi Dear Abby-Normal,
How are you? Hopefully everything is fine. R&D department is extremely busy now. Just like to ask you a question. I have a full factorial DOE (3 factors with 2 center points, no replicates). In the analysis, it shows that the model is not significant (p>0.05), but one of the main factors (let's say C) is significant (p=0.01). No lack of fit for the model. In this case, since the model is not significant, is C still a significant factor which accounts for the variability in the results?
Signed: Lost in the p’s and q’s.
Question 3 (Answer)
Dear Abby-Normal,
I ran an experiment and found no significant factors and the R-squared value was really low. But I know that the factors are important. Why is this? Does DOE really work?
Signed: DisappOintEd
Question 4 (Answer)
Dear Abby Normal,
I have a measurement study from 10 parts turned in to me that looks too good to be true. It is well known in my company that this lab measurement for percent of ash content has much variability, yet the %R&R is 18.42%, well under the 30% limit. How can this be? Can I trust this measurement system?
| Source |
StdDev (SD) |
Study Var (6*SD) |
%Study Var (%SV) |
| Total Gage R&R |
0.066667 |
0.40000 |
18.42 |
| Repeatability |
0.039441 |
0.23664 |
10.90 |
| Reproducibility |
0.053748 |
0.32249 |
14.85 |
| Operator |
0.051460 |
0.30876 |
14.22 |
| Operator*Sample |
0.015516 |
0.09309 |
4.29 |
| Part-To-Part |
0.355668 |
2.13401 |
98.29 |
| Total Variation |
0.361862 |
2.17117 |
100.00 |
Number of Distinct Categories = 7
Signed,
Repeatedly Confused
Question 5 (Answer)
Dear Abby Normal,
I have run several experiments and analyzed several sets of production data. I know how to interpret the p-values, if p is less than 0.05, I conclude H1, if p is greater than 0.05, I conclude H0. Yet, when someone asks me to explain what a p-value is, I tend to freeze. Is there an easy way to explain p-values.
Signed: Perplexed about p Values
Question 6 (Answer)
Dear Abby Normal,
I came across your website while searching the Internet to solve a problem I’m working on. I’m trying to determine the upper limit for a parameter based on the raw data in the attached EXCEL file. The only requirement is that the upper limit should have a Cpk =1.67 or higher to fulfill a customer requirement. What are the formula(s) needed to carry out this task?
Thank you and best regards,
In the dark
|
