Problem Set 3

BUS 310
Group Members: Stephen Jefferey, Andy Katz, Gregory Barajas, Zach Green, Riley Hazen, and
Steven Blasco
14.21:

The difference between common cause and special variation is that special cause of variation
represents large fluctuations or patterns in data that are not part of a process, often caused by
unusual events that represent problems to correct or opportunities to exploit. While common
causes of variation represent the inherent variability that exists in a process, usually small causes
of variability that happen randomly or by chance, often referred to as chance causes of variation

14.25:

The difference between attribute and variable control charts are that attribute are used for
categorical or discrete variable while variable control charts are there to monitor and analyze a
process when you have numerical data

14.26:

XBar and R charts are used together because they are a pair of control charts that are used when
there are multiple subgroups for the data

Problem 4

a). Overall Sales (with and Without Promotion)

b). Before the promotion, the chart is in control because it follows the rules, does not go outside
of either UCL or LCL and does not have 8 consecutive days positive or negative. .

c). Since the promotion, it is hard to tell whether the data is in control or not because there are
only 7 days of data to look at, but based on the 7 points the data is in control. Looking at the p
chart based on days before and after the chart, you can see that the promotion is working because
of good sales making the p chart out of control going past the UCL
d). It is necessary to generate two p charts for the two scenarios because if you do two, one being
before the promotion, it will give you data based on all regular variables. It is also important to
choose to do one after the promotion to see how much the promotion is working and then from
the promotion you can create another control grouping of data to see how the promotion data
compare against itself

e). It would be insufficient to compare days 1-30 to days 31-37 because there has been a new
variable added to help raise the sales, the data is not the same. Also, you would not be able to see
the comparison between the data because they are essentially different data sets you would be
comparing.

f). No, in this situation being out of control is good because you changed a variable to make the
sales go up, making the chart out of control but in a good way for the business. It is expected that
with changing the variables there would be a change in control. Thus, we do not need to correct a
problem, but have a regular sales promotion to keep progressing sales

Problem 4:
A.

1-30 P-Chart
0.4

0.35

0.3

0.25
Proportion

0.2

0.15

0.1

0.05

0
0 5 10 15 20 25 30 35
Time
 1-37 P-Chart
0.6

0.5

0.4
Proportion

0.3

0.2

0.1

0
0 5 10 15 20 25 30 35 40
Time

B. It is in control. Segments are evenly distributed above and below the P-bar.
C. There are 7 segments that are above the UCL and 1 segment on it, likewise, there is one
segment below the lower control limit. There is an uneven distribution below the P-Bar.
D. It shows the effect of the promotion on sales.
E. There is not enough data to compare 31-37 to 1-30. Additional time needed.
F. It is desirable because it shows the promotion is making an impact on the data. We need
more time (data points) to fully understand the effect of this promotion.
Problem 5:
Chapter 8:
1. The confidence interval for low is assumed to be invalid due to the sample size being
under 30, which one can assume that the skewness and kurtosis is not between 1 and -1,
meaning that there is an outlier present. Thus, the sample size and probability are invalid.
The confidence interval for medium is assumed to be valid due to the sample size being
greater than 30, meaning the data is normal and that skewness and kurtosis is between 1
and -1. Finally, the confidence interval for high can be assumed to be invalid due to the
sample size being under 30 and the probability is invalid. One can also assume that the
data is not normal, and that skewness or kurtosis is not between 1 and -1.
Confidence Interval Estimate for the Mean
Low
Data
Sample Standard Deviation 10238,6202
Sample Mean 37087,03448
Sample Size 29
Confidence Level 95%

Intermediate Calculations
Standard Error of the Mean 1901,26404
Degrees of Freedom 28
t Value 2,0484
Interval Half Width 3894,5628

Confidence Interval
Interval Lower Limit 33192,47
Interval Upper Limit 40981,60

I am 95% confident that the wage for the percentage of union employees is between $33192.47 and $40981.6.
We can be 95% confident that population mean of wages for Low union companies falls
between $33,192.47 and $40981.59.

Confidence Interval Estimate for the Mean

Medium
Data
Sample Standard Deviation 10385,19016
Sample Mean 34605,60526
Sample Size 57
Confidence Level 95%

Intermediate Calculations
Standard Error of the Mean 1375,55204
Degrees of Freedom 56
t Value 2,0032
Interval Half Width 2755,5619

Confidence Interval
Interval Lower Limit 31850,04
Interval Upper Limit 37361,17

I am 95% confident that the wage for the medium percentage level of union employees
We can be 95% confident that population mean of wages for Medium union companies
falls between $31,850.04 and $37361.17.
Confidence Interval Estimate for the Mean
High
Data
Sample Standard Deviation 12982,63228
Sample Mean 37152,53571
Sample Size 14
Confidence Level 95%

Intermediate Calculations
Standard Error of the Mean 3469,754425
Degrees of Freedom 13
t Value 2,1604
Interval Half Width 7495,9487

Confidence Interval
Interval Lower Limit 29656,59
Interval Upper Limit 44648,48

I am 95% confident that the wage for the high percentage level of Union employees is between $29656.59 and $44648.48

We can be 95% confident that population mean of wages for High union companies falls
between $29,656.59 and $44,648.48.

Descriptive Summary

Low Medium High

Mean 37087,03448 34605,60526 37152,53571
Median 35026,5 33487,5 33231,75
Mode #N/A #N/A #N/A
Minimum 20404,5 18034,5 22617
Maximum 64479 71513,5 61792,5
Range 44074,5 53479 39175,5
Variance 104829343,5702 107852174,5869 168548740,8255
Standard Deviation 10238,6202 10385,1902 12982,6323
Coeff. of Variation 27,61% 30,01% 34,94%
Skewness 0,8155 1,2678 0,9284
Kurtosis 0,6059 2,4795 -0,2483
Count 29 57 14
Standard Error 1901,2640 1375,5520 3469,7544

LOW: The confidence interval for low is assumed to be invalid due to the sample size
being under 30. This confidence interval estimate seems to be valid because the data
appears to be Normal OK since both the skewness and kurtosis are between -1 and 1.

MEDIUM: This confidence interval estimate appears to be not valid because the data
seems to be not Normal since both the skewness and kurtosis are below -1 and above 1,
but the sample size is greater than 30, so the confidence interval appears to be valid.
 HIGH: This confidence interval estimate seems to be valid because the data appears to be
Normal OK since both the skewness and kurtosis are between -1 and 1, but the sample
size is smaller then 30, so the confidence interval appears to be invalid.

2. A. We can be 95% confident that the percentage of sidepreneurs that also get their
degrees falls between 34% and 38%.

B. The interval width for 95% is .05 or 5% and the interval width for 90% is .04 or 4%.,
therefore the difference in interval width is .01 or 1%. The 90% confidence interval gives
a more precise estimate because its width of 4% is smaller.

Problem 8.50:
A 100% confidence interval is not possible unless the whole population is sampled or an
absurdly large estimate interval is provided.
Problem 8.52:
A higher confidence interval will tend to widen the confidence level given a fixed sample size.
The confidence interval will shrink as the sample size increases, all else being equal.
Problem 8.53:
The sample size formula for p is: n=(z/ME)^2 * p *(1-p) Is the largest when p = 5