Page 33 - Design for Six Sigma a Roadmap for Product Development

Quality Concepts 11

Acceptance sampling, which was developed to solve this problem, is
the inspection of a sample from a lot to decide whether to accept or
reject that lot. Acceptance sampling could consist of a simple sam-
pling in which only one sample in the lot is inspected; or multiple
sampling, in which a sequence of samples are taken and the
accept/reject decision is based on statistical rules.
The acceptance sampling plan was developed by Harold F. Dodge
and Harry G. Romig in 1940. Four sets of tables were published in
1940: single-sampling lot tolerance tables, double-sampling lot toler-
ance tables, single-sampling average outgoing quality limit tables, and
double-sampling average outgoing quality limit tables.

1.3.3 Design of experiment (late 1930s)
Design of experiment (DOE) is a very important quality tool in current
use. DOE is a generic statistical method which guides design and
analysis of experiments in order to find the cause-and-effect relation-
ship between “response” (output) and factors (inputs). This relation-
ship is derived from empirical modeling of experimental data. DOE
can also guide the experimenter to design efficient experiment and
conduct data analysis to get other valuable information such as iden-
tification and ranking of important factors.
DOE was initially developed to study agricultural experiments. In
the 1930s, Sir Ronald Fisher, a professor at the University of London,
was the innovator in the use of statistical methods in experimental
design. He developed and first used analysis of variance (ANOVA) as
the primary method in the analysis in experimental design. DOE was
first used at the Rothamsted Agricultural Experimental Station in
London. The first industrial applications of DOE were in the British
textile industry. After World War II, experimental design methods were
introduced in the chemical and process industries in the United States
and Western Europe.

1.3.4 Tools for manufacturing diagnosis
and problem solving (1950s)
Statistical process control (SPC) is a process monitoring tool. It can
discern whether the process is in a state of normal variation or in a
state of abnormal fluctuation. The latter state often indicates that
there is a problem in the process. However, SPC cannot detect what
the problem is. Therefore, developing tools for process troubleshooting
and problem solving is very important. There are many tools available
today for troubleshooting; however, Kaoru Ishikawa’s seven basic tools
for quality and Dorian Shainin’s statistical engineering deserve spe-
cial attention.