Minitab data labels decimal places

Case Z: The Furniture Company (Rev 2_19)

This case has been developed for you to apply quality tools and statistics you have learned in the course to a

realistic scenario. You will use Minitab to perform your calculations and generate graphs. Then you will analyze

the information, explain the results, and evaluate the results to make recommendations.

Background

The Ellerson Furniture Company sells desks and bookshelves unassembled to customer. The customer orders the

kit off of the website and the company promises to ship the product within one week. The customer service desk

receives numerous phone calls from the customers inquiring about the status of orders and during the assembly

process.

The new customer service manager, Ms. Casey, thinks that the role of customer service should be to not only help

the customer but to provide feedback to engineering department as well as the manufacturing and warehouse

operations. The relationship between the different departments is depicted in the flowchart below.

Ms. Casey felt that she needed to further define the problems being reported. She called a meeting of the service

representatives and came up with a list of common complaints that were being received.

1. Missing part 2. Shipment late 3. Missing/wrong instruction 4. Incorrect part 5. Instructions confusing 6. Correct part would not fit 7. Wrong product shipped 8. Shipment damaged

Ms. Casey then revised the on-line form that the representatives used in documenting complaints in order to data

analysis. She incorporated complaint codes corresponding to types of complaints identified. Previously the nature

of the complaint was only provided in narrative form which made analysis difficult. The following information was

collected for each complaint:

(includes pin machining)

(manual process)

Date

Caller name/address/phone #

Kit number

Nature of complaint

Complaint codes

Action taken to correct

At the end of a six week period, a data file was generated for Ms. Casey. She requested the assistance of the

Corporate Continuous Improvement Office in analyzing the data. You have been assigned to the project.

Part A.

There are 256 complaints documented in the file. A summary file of data is provided to you with the following

columns of data:

Date

Kit prefix – code which identifies the operations facility

100 Kansas 200 Idaho 300 Texas

Kit suffix – identifies the specific model of the kit. There are 12 different kits.

Complaint code - number corresponding do above list.

You decide to use Pareto charts to help in determining the most prevalent problems. Initially you run charts

looking at occurrences by type of complaint overall and for each operations facility.

The data can be found in your assigned project file.. Create a chart of all complaints, one for each facility (so you

can see the count data), and a chart with all three in one graph. “Do not combine” categories as you will need to

see all the count data. See the attachment to see how to see how to create the Pareto charts easily using variables.

A.1. Initial Pareto Charts of Counts. (Paste the overall chart and the three in one chart into this section. Do not

past the three individual facility charts in the report.)

A.2. What are three most common types of defects/complaints? What percent of your total defects do these

represent? Discuss how this relates to the Pareto Principle.

A.3. Is there any appreciable difference in the in the numbers of complaints or types of complaints by

assembly/warehouse location? Discuss.

In reviewing the data, you wonder if there is a statistical difference in complaints as a percentage of total orders

shipped from each facility. During the time the data was collected, Kansas shipped 995 orders, Idaho shipped 845

orders and Texas shipped 980 orders. You would expect the distribution of the complaints to be proportional to the

distribution of orders shipped. You summarize the data for analysis below:

Facility # of orders shipped

from facility

# of complaints at

the facility

You decide that you should statistically compare the proportions to see if the analysis can provide you with

information useful in targeting improvement efforts.

A.4. At a 95 percent level of confidence, what conclusions can you draw about the proportions of complaints

at the different facilities? (Paste in session window and bar chart of observed and expected values from

Minitab output. Discuss the hypothesis test: Explain null and alternative hypothesis. State conclusions in

statistical terms and explain why the conclusion was made. State the conclusion in practical terms as you

would state to management.)

Set up three new columns in the complaint data sheet to reflect the data in the table from A.3. See Minitab

Exercise 12 for comparing expected versus observed proportions. You will use proportions specified by

historical counts in Minitab. The orders will represent the historical counts.

You decide that you should present your data to the key personnel from each location. An ad hoc committee was

formed with representatives from engineering as well as the production and warehousing supervisors. After

reviewing your data, they have several responses.

Part B.

The members note that not all complaints cost the company the same amount of money to correct. A customer

service representative can usually talk the customer through the assembly process, so it doesn’t cost the company as

much money as having to send out missing parts. After considerable discussion, they give you cost estimate data to

provide weighting of each of the complaint types.

Complaint Cost Count From Overall Pareto

1. Missing part $40

2. Shipment late $5

3. Missing/wrong instruction $12

4. Incorrect part $40

5. Instructions confusing $6

6. Correct part would not fit $40

7. Wrong product shipped $84

8. Shipment damaged $80

You are able to run a new weighted Overall Pareto quickly.

Open a new worksheet. Label four columns – complaint, count, cost, and weighted cost. Enter the data in

the first three columns. Multiply the count by the cost to get the data for the weighted cost column. (You

did this in the first Minitab exercise.) Then create the weighted Pareto chart.

B1. Pareto Charts by Cost. (Paste the weighted Pareto chart by cost (just overall one) in this section.)

B.2. How did the cost weighting impact the order of the complaint types? Discuss.

Part C

The facility with the most complaints also has the most complex products. They want to know if there is a

relationship between the number of parts in a kit and the number of complaints received and which kits are most

problematic. They provide you with the number of parts in each kit. Since you have access to your data and a

computer, you run a quick x-y (scatter) plot.

Data is found in the worksheet titled “Parts versus complaints” in your assigned project file. See Minitab

Introduction for scatter plot instructions. Put an item number label on each point using Labels>Data

labels. You do not need to run a regression analysis, but when doing the scatter plot it may be easier to see

if there is a relationship if you select the option with the line.

C.1. Paste x-y (Scatter) plot here.

C.2. What useful information can be determined from the x-y plot? Discuss.

Part D.

D.1 Based on your analysis thus far (all previous parts), where do you think improvement efforts should be

focused. Why? (Give exactly four areas of effort citing reasons/data.)

To further focus your efforts, you decide that there is more analysis that could be performed with the data that you

are provided.

D.2. Describe at least 2 ways that you can use Minitab to further analyze the given data (generating

additional graphics) and state why the information would be useful.

Look at the files provided to see what other data is there that you could use.

D.3 Select one of the two analyses described in D2 and run the graph. Paste additional graph/s here.

Describe specifically how the graph/s assists in focusing improvement efforts.

Part E.

The process engineer from the Texas facility said that he had recently received several of the complaint inquiries

regarding a pin that was slightly oversize (correct part did not fit). He said there had been some of the equipment

was getting rather old and no longer holding as close to tolerance as previously. He usually reworked the part,

when it was above the specification limit, to where it was just within specifications. He acknowledged that at the

upper spec, it was difficult to fit in the mating part when the mating part was at its lower specification. The

engineering specifications for the diameter are 0.25 ± 0.0025

E.1. How does this information from the engineer fit with your analysis in the previous sections?

E.2. Minitab Capability Analysis. (Paste Minitab output here.)

The engineer shared his data with the group. You do not know how the data was collected. You decide to run a

capability analysis on his data.

The assigned Pin Diameter worksheet file contains the measurements of the critical diameter of the part.

Open this worksheet in your project.

In Minitab, run the descriptive statistics for the critical diameter. Run a capability analysis for his initial

data. (In Stat>Quality Tools>Capability Analysis>Normal, data will be in single column of subgroup size

1. Enter the standard deviation (using all decimal places) from the descriptive statistics for the historical

standard deviation. (If you have done this correctly, the Cp indices will be the same as the Pp indices.)

E.3. What assumptions are being made about the data used for the analysis? Which assumptions can be

verified and how? Discuss.

E.4 What is the capability of the current process to hold the diameter? (Discuss Cp, Cpu, Cpl, and Cpk and

what each tells about the process if you were explaining to management.)

E.5 Assuming the process remains in its current state, what percentages of diameters would be out of the

upper and lower specifications in the future. Explain (as if to management) these percentages relate to the

indices.

Part F.

A list of action items was generated at the meeting. Each of the facilities was tasked with focusing on specific

areas as was recommended to identify causes and implement solutions. The ad hoc committee agreed to meet again

in three months to review the situation.

You have been tasked with helping the Kansas facility.

F.1. Select one of problems/complaints significant to the Kansas facility for further analysis. Use one of the

techniques/tools learned (tree diagram, why-why diagram, cause and effect diagram) to document your

analysis of potential causes. Paste your analysis table/figure here. State why you selected the problem.

It is not recommended to use the Minitab Cause and Effect diagram. It does not provide for the level of

detail needed to get to actionable causes. The examples they give do not give causes at all, just one or two

word statements.

It is important that you take the analysis down to actionable causes. Example of getting to actionable

causes can be found in the Module 3 Basic Tools II lecture and lecture extra – Tool Crash problem..

F.2. From your technique, what do you feel are the two most likely causes of the problem? Why? What are

some actions that could be taken to address these causes?

Part G.

The engineer from Texas shared his latest data. He is proud that he reduced the variation on the process through

overhaul of the equipment.

To get the data to use in this section, generate 61 observations in a new column of the Pin Diameter

Worksheet using Calc>Random Data>Normal, with mean of 0.25 and standard deviation of 0.0007.

Assume you were given this data by the engineer.

You decide to run new capability data to compare to the previous data.

Run the descriptive statistics on the new data to get the new variance and mean.

Determine the new capability (in the same manner as in E.2.)

G.1. Descriptive Statistics (paste session window output) and new capability analysis chart (paste the chart)

G.2. Discuss the changes in the process as indicated by comparison of the new capability study with the

previous study.

You think the data shows the variation to have been significantly reduced, but decide that in order to make that

statement you must do further statistical analysis.

G.3. With a 95% level of confidence, can you conclude the variance has been reduced? (Paste in Minitab

output. Discuss the hypothesis test: Explain null and alternative hypothesis. State conclusions in statistical

terms and explain why the conclusion was made. Comment on the strength of the conclusion. State the

conclusion in practical terms as you would state to management.)

See Minitab Exercise 10. Enter all digits of the standard deviation from the descriptive statistics. Although

the session output only shows three digits, the calculations are performed with all digits.

Part H.

In preparation for the follow up meeting, you collect 6 weeks of complaint data for comparison to the original 6

weeks of baseline data. There were 246 complaints in the current 6 weeks compared to 256 in the previous 6 week

data.

You also noted that the number of total kits sold in the current time frame (3435) was greater than the number sold

in the original time frame (2820). You wonder if the proportion of complaints has been significantly reduced.

H.1. At a 95% level of confidence, can you conclude that the percent of kits sold that had a complaints

decreased? (Paste in Minitab output. Discuss the hypothesis test: Explain null and alternative hypothesis.

State conclusions in statistical terms and explain why the conclusion was made. State the conclusion in

practical terms as you would state to management.)

You can use summarized data similar to done in Minitab Exercise 11 to compare the proportions.

# of complaints Number of

orders

Ratio complaints to orders*

(sample p in Minitab)

Previous 6 weeks

Current 6 weeks

 Minitab will calculate the proportion for you, but you may want to keep this concept in mind later in the case.

You then look at the complaint detail. The numbers of complaints by defect type are shown in the table.

Complaint Number

1. Missing part 47

2. Shipment late 38

3. Missing/wrong instruction 21

4. Incorrect part 18

5. Instructions confusing 83

6. Correct part would not fit 16

7. Wrong product shipped 11

8. Shipment damaged 12

You decide to run new Pareto charts for both counts and cost.

You will use two new columns in your previous worksheet where you ran the weighted cost chart. Enter the

new count data and calculate the cost data. Run count and cost Pareto charts with the new data.

H.2. Paste “After” Pareto charts here.

H.3. In comparing the previous Pareto charts to the new ones, what has changed relative to the order of the

complaints? What is the significance of the change in order?

H.4. When comparing the previous to the new, why can you not directly compare the frequencies or cost

data before and after to determine if improvements have been made? How can you be adjust the data to

make a direct comparison?

Think about the analysis you did in H.1.

H.5. Use the method you described in H.4 to compare the before and after effect for the most costly defect in

the after Pareto chart. Show the analysis. Describe the conclusion of the analysis as if you were describing it

to management.

Part I.

At the follow-up meeting committee members report successful improvements had been made as a result of the

efforts at each facility. They ask that you prepare some statements to be included in the latest report on continuous

improvement which is submitted monthly to corporate.