Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

4 Introduction

Table 1.1 Examples of processes

Input Activity Output

Manuscript, - Word processing Letters, documents


audio tape
Clock cards, - Computing Pay slips, credit transfers
time records
Raw data Data processing Reports, statements, invoices,
VAT returns
Enquiries Sales activity Orders
Orders Manufacture or Delivery
selection from stock
Seeds, fertilizers - Cultivation Crop
Materials or - Manufacturing Finished goods or
components 'value·added' materials

For example we have


Food-+ Digestion-+ Energy
(Is this not, in any case, known as the digestive process?)
Here, food provides the input, digestion the activity. Both need to be
controlled to ensure the right output. One might even consider monitoring
energy input and charting weight gain or loss as examples of the use of SPC
techniques!
Some less flippant examples include agricultural, clerical, service and manu-
facturing processes, such as those in Table 1.1, where the inputs may comprise
materials, documents or verbal statements, and the activities may be carried
out by individual action or the use of equipment.

1.4.2 The traditional role of quality control has been to check the output of the
process, identifying substandard output for rectification, repair, re-work, down-
grading (for sale to undiscerning customers as 'seconds'?) or outright rejection.
Such control (which is certainly not the activity intended by the statistical
quality control pioneers like Shewhart in the USA and Pearson and Dudding
in Britain) is analogous to steering a ship by looking at the wake. The identi-
fication of half-a-dozen aspects common to all processes and systems can focus
attention on problem prevention rather than trouble detection. Various authors
list four, five or all six of these aspects under different names, e.g. men or people,
machinery or equipment, methods or procedures. We adopt the acronym
IMPOSE, i.e. a means of applying the system strategy so that the intended
outputs are realised. Figure 1.1 places these aspects in the system context, and
they should be regarded as sources of data, areas for control and improvement
and for the application of problem-solving techniques.
Statistics 5

Redesign/update of
process/product/servic e

Aspects of process Action


(IMPOSE) on process
Inputs
Machinery Sources
People } Information
of
Operations Process from process
Services
Data
Env1ronment

Process audit
Output
information

Fig. 1.1 Process control.

Whether in production, design, administration, service, the results of control


(rather than detection of faults) are smoother operations, reduced cost, higher
outputs, improved quality and customer satisfaction and better human relations.
The latter result from the sense of a job done well, and from the satisfaction
gained by people with a day-to-day knowledge of problems in the system actually
having that knowledge tapped and acted upon.

1.5 Statistics

1.5.1 Having briefly discussed 'control' and 'process', we must now define
Statistics. In this particular context it is any quantitative technique which
contributes to
monitoring and control
understanding the system
problem identification
problem diagnosis
problem solving
improving the system
rational and objective decisions.
These techniques may be grouped as follows.
Logical data collection: aimed at solving problems and understanding systems.
6 Introduction

Well-designed record sheets: making it easy to collect the required data, using
tally sheets, control charts, fault location diagrams, questionnaires, etc.
Statistical summaries: both numerical and graphical; tables, charts and
diagrams, measures of magnitude and variation, rates and percentages.
Establishing priorities: pareto analysis, ranking, consensus judgement.
Generating and organizing ideas: including brainstorming, cause and effect
analysis, establishing quality circles or problem-solving techniques. Failure
mode and effects analysis (FMEA) could be usefully employed here in a broad
sense.
Looking at relationships: two-way tables and scatter diagrams for investigating
numerical cause and effect patterns.
Monitoring and control: observing patterns over time, using run and control
charts.
These activities are served by the basic statistical tools of SPC/TQM, and
some are briefly reviewed in early chapters of this book. The bibliography lists
sources of more detailed information.
The main purpose, however, is to provide a grounding in the statistical
concepts on which these and other tehchniques depend, with a minimum of
mathematical theory. These concepts will then be applied to SPC topics, dealing
especially with the following areas.
First principles ofcontrol charts: with applications to individual values, moving
averages, non-normal data and processes with special problems or features.
Cusum methods: not only as an alternative to control charts, but as a general
tool for data presentation and analysis.
Capability analysis: evaluating the ability of a process, machine or measure-
ment system to produce results acceptable to the user or customer. Included
here is 'gauge capability', or the evaluation ofrepeatability and reproducibility.
Methods based on normal and non-normal distributions are presented.

1.5.2 We bear in mind, however, that SPC, in either product, process or total
quality applications, must not be dominated by techniques. To use an often-
quoted aphorism, SPC is 20% statistics, 10% common sense and 70% engineering.
For non-engineering applications, one would substitute a detailed expertise in
the area of interest - medicine, computers, plastics technology. Or, as W.
Edwards Deming puts it, 'statistics is a basic but comprehensive way of thinking
and acting which combines information, opinion and common-sense for the
better design of products, systems, strategies and decision-making'. The accent
is on communication, teamwork and problem-solving.

1.5.3 Successful implementation of SPC, or the effective use of statistical


techniques for effective system management, thus depends on a blend of manage-
ment skills, technology, human relations and statistics. Figure 1.2 summarizes
the content of this chapter with regard to the role of the statistical methods,
Statistical} Efficient Management (Contron of any
Process = System (Process) using quantitative information
Control effectively presented and analysed (statistics)

Continuous improvement Areas of application Examples of statistical tools

Data gathering sampling, checksheets, tallies, fault diagrams, control and


run charts
Measure performance
Data presentation stem/leaf, box/whisker, histogram, bar chart, run chart,
scatter/blob diagram, other graphs and tables, control
l
Highlight problems charts.

Setting priorities Pareto analysis, ranking, risk analysis, FMEA


l
Identify causes
Finding causes Cause-effect diagrams, brainstorming, collation/correction,
FMEA, flow sheets
l
Develop solutions
Performance measures Averages, measures of variation, rates, proportions, capability
indices
l
Measure improvement Monitoring Control charts (various), cusums

-Relating to process Stratification, collation, annotation, planned observation,


experiments

(FMEA =Failure mode and effects analysis)

Fig. l.l Statistical methods in system and process management.


8 Introduction

irrespective of the nature of the process, the product or the size or type of
organization. Often the most difficult step is that of finding a suitable measure
of performance, especially in applications that are not industrial nor driven by
a customer specification.
The statistical methods are often limited to a so-called set of 'seven statistical
tools'. Any set of seven is necessarily restrictive, so Fig. 1.2 uses seven potential
areas of application, with illustrations of some of the appropriate tools. There
is considerable overlap between these areas of application and the tools - control
charts provide convenient data collection forms, effective means of monitoring,
and a useful form of data presentation. They also provide a ready format for
further analysis, including the calculation of performance measures.
The rest of this book is concerned with developing an understanding of
statistical principles so that the various tools can be properly and usefully
applied to process management and improvement.

You might also like