Errors LO12231
Tue, 28 Jan 1997 21:46:50 +1300 (NZDT)

Replying to LO12212 --

As I understand it the problem being discussed in this thread is:

1. A measured increase in error rates when undertaking simple repetitive

2. The observed increase occuring in work teams.

There are a number of points to be made here. They progress the earlier
thread on this subject.

1. The subject of human error is the focus of rigorous research, notably
but not exclusively in the aviation and electric power generation
industries. This research appears to be little known outside settings
where there is risk of catastrophe, and even within these industries it is
still often badly misunderstood (a fact which explains the failure to
further reduce airliner accident rates in the past three years). This
research provides us with the potential for tools to minimise error, and
nearly eliminate the damaging consequences of error. The crucial book on
this subject is James Reason's 'Human Error' (1990). After reading this
move on to his more recent work, together with that of Maurino (A
Canadian) and Neil Johnston (Dublin). Also visit the NASA Human Factors
WWW site.

2. Those who say that it is impossible to eliminate error are correct.
However it IS possible to:

(a) reduce error rates to the irreducible minimum. What this is is
measurable, and determinable within your industry by benchmarking. However
the research has shown that for routine repetitive tasks the error rate is
remarkably consistent across settings at a probability of around 0.0005:

(b) to establish defensive 'safety net' procedures which pick up
and correct most errors before they cause any damage. Some of these which
might be used in team settings have been suggested by earlier contributors
to the thread.

3. Very few work settings have actually reduced error rates to the
irreducible minimum. This is because there are almost always environmental
factors present which multiply the error probability. These factors fall
into a number of categories:

(a) Factors personal to the individual operator, for example having
a hangover, having missed breakfast, being seriously in love, etc.
Despite popular belief these factors are NOT big error multipliers
(typically in the range x3 to x7). Lack of competence is a personal
factor, and the only one which is affected by training. When you read the
full list of factors below you will realise the absurdity of the fact that
more training and more supervision are often the only error reducing
strategies adopted by organisations. The supervision solution also tend to
confuse errors with violations - which are another and far more complex

(b) Ergonomic and physical environment factors, such as lighting,
work station design, heating. The impact of these factors is empirically
identifiable and measureable. They can be significant multipliers - up to
x15. They are obviously susceptible to management influence.

(c) Day to day organisational factors. These include a whole range
of management determined issues, some of which are measurable such as
shift arrangements, productivity pressures; and some of which are not,
such as racial tension between operator and supervisor, strict discipline
(a significant error MULTIPLIER because it increases stress), or poor firm
performance leading to job insecurity. Change - benign or otherwise - is a
short term error multiplier, for example a change to team organisation. If
the change is well founded, then there is likely to be a short term
increase in error rates, followed by a steady decline to lower levels.
These multipliers operate across the range from x2 to x20. They have been
research identified by proxy measures rather than by direct statistical

(d) Deep level organisational pathogens operating at the cultural
level. These are both the most insidious (because culture is by definition
concerned with what is 'taken for granted') and are also often the biggest
multipliers. They are the most difficult to identify and eliminate and
include things like unclear organisational goals ( eg "we do not
compromise on quality but we have to get the product out the door on
Friday no matter what" or - a recent one we encountered for the first time
- "our organisation values straightness and honesty and demands a positive
attitude from its employees" - resulting in those who attempt to be
'honest' and point out problems always being accused of not having a
positive attitude.) These factors tend to be multipliers in the range x10
to x20, but sometimes the only way to find them is by deduction. That is
if we find high error rates and none of the empirically measurable causes,
then we should start digging for deep level cultural factors. That these
factors really do multiply error rates has been demonstrated by the
reductions in error rates that occur when the organisation succesfully
confronts dysfunctional cultural issues.

We have some experience with adapting Safety Audit tools developed in
aviation settings for other industries. Safety audits are not the same as
risk management exercises, the latter now being suspects of having the
effect of increasing error rates in many cases.

Generally speaking all of the above message is easily transferable to
analysis of the occupational health and safety field.

Phillip Capper
Centre for Research in Work, Education and Business
New Zealand


Learning-org -- An Internet Dialog on Learning Organizations For info: <> -or- <>