This preview shows page 1. Sign up to view the full content.
Unformatted text preview: xamine the actual spot where the problem occurred.
b. Decide on the quality characteristics that the evaluation of quality will be based on.
(When these are converted into quantitative data they are called characteristic
c. Formulate clearly the objectives for which the data is being collected.
d. Gather accurate data.
e. Analyse this data using statistical techniques.
f. Obtain accurate information from this analysis. Discussion
The following questions ask you to think about how the ideas in the text could be applied in
your company. Some of the ideas may not be relevant to you. Concentrate on what is
relevant. Keep notes of your conclusions – you will need them to prepare your action plan
afterwards. Where appropriate ask yourself the RADAR questions.
Note: Always include in your discussion any figures referred to in the text, if you feel these
are relevant to your company.
a. Parag. 1: To what extent would you say that the approach to problem solving in your
company is based on experience, and to what extent on quantified facts? Would you
like to see the balance changed?
b. Parag. 2 suggests six steps for establishing the facts. Apply the RADAR questions to
these. Action plan
Take the ideas you have found useful in the text, and in your discussion, and present them in
a well-structured action plan for introducing improvements in your company. You might like
to follow the 6-Point Structure. Alternatively you may choose to prepare one action plan
when you have discussed several texts. A Roadmap to Quality 19 Unit 9 - Problem solving 05-87581_unit 9.qxd 09/09/2005 11:49 Page 20 9.7 Managing dispersion
1. Once data has been collected, it has to be interpreted. Averages are the most common
way of interpreting data, but they often fail to give a true picture of what the data
means. Measuring how the data is dispersed gives a more complete picture. Dispersion
refers to how the different items of data are spread out or scattered in relation to how
they are supposed to be, i.e. in relation to the standard or target values. For example, a
residential street with 10 houses with an average price of $240,000 and where each
price differs only a little from the average, would be very different from a street with the
same average house price, but with 2 houses valued at $1 million and the other 8 each
costing around $50,000.
2. The first thing to do when the data seems to indicate a problem is to clarify whether it is
the average or the dispersion that indicates that there is a problem. Otherwise it is
impossible to solve the problem. Problems that are indicated by averages of the data
can be solved relatively easily. Just review the processing conditions and any other
factors that affect the results. When problems are indicated by dispersion, base your
countermeasures on whether:
a. The range of dispersion (the distance of the maximum and minimum data points)
from the standard or target values (also referred to as the technical standard) is
acceptable, but the average is skewed (distorted or biased).
b. The range of dispersion is too wide.
c. There are outliers (An outlier is an item of data, or a value, that falls well outside the
dispersion range of the rest of the data).
3. Dispersion may be due to chance or to abnormalities. There will always be some
dispersion even when the materials and work methods are those prescribed by the
standards. It cannot be avoided. This type of chance dispersion stays within a certain
range. The values tend to form a bell curve, with the average in the centre. This pattern
is known as normal distribution.
4. Dispersion caused by abnormalities may result from the following factors:
a. Employees do not follow the operational standards.
b. There are changes in materials.
c. Inexperienced employees replace experienced employees.
These factors skew the average and cause outliers.
5. Dispersion in the quality of a product results from the dispersion of something in the
manufacturing process that is strongly related to quality. This dispersion provides a good
opportunity to find out the causes of such problems. It indicates that the cause of the
problem is strongly related to the results. Such a cause can be identified by searching
for any divergent factors (factors that are different from what they should be) and
examining their correlation to the dispersed results. Unit 9 - Problem solving 20 A Roadmap to Quality 05-87581_unit 9.qxd 09/09/2005 11:54 Page 21 6. The three charts in Figure 9.7a, Quality Dispersion, give examples of problems indicated
by the average of the data, by the dispersion of the data, and by outliers.
Figure 9.7a Quality dispersion Causal relationship Technical
standards Problem concerning
through the movement
of average values
standards Observation Dispersive
characteristics Confirmation of
dispersive characteristics Occurs to factors
contribution ratios Identification of causes
through the analysis of
dispersive factors that
correspond to dispersive
View Full Document
- Spring '12