Hidden Faultlines In Your Organization

1
Hidden Faultlines In Your Organization

• Find them FAST FIX them Forever
• Dr. Ted Spickler
• Quality and Business Services
• 412-777-2054 ted.spickler.b_at_bayer.com
• ? Bayer Corporate and Business Services LLC

2
Why Are We Here?

• Organizations, like geological features, are
  subject to earthquakes - sudden upheavals that
  can later be attributed to hidden, underground
  faultlines that are sensitive to stresses and
  can, without warning, let lose with disastrous
  consequences.

3
Faultlines

• Some of us have chronic problems that have defied
  previous attempts at resolution.
• Old problems rear up out of the blackness and
  bite us again and again.
• Meaning we never really fixed them the first
  time.

4
What Are We Going to Do Today?

• Examine techniques for uncovering faultlines.
• Practice building logic trees, a key tool for
  uncovering Latent Causes.
• Differentiate between Direct Causes, Symptoms,
  Contributing Causes, and Latent Causes.
• Learn to build effective fixes to these problems.

5
When Something Goes Wrong Typical reaction

• Shoot the messenger.
• Jump to quick conclusions about why something
  happened.
• Find someone to blame - THEN Hang em high!
• Review procedures with bad person.
• Re-train and discipline bad person.
• How about blame the supplier!
• Better yet Blame the CUSTOMER!
• Hope it doesnt happen again.

6
We Need A Better Approach

• Find out what really went wrong.
• How do the quakes happen?
• What can we do to prevent bad things like this
  from happening again?
• Where do you find evidence for the hidden
  faultlines?
• Utilize a systematic approach using tools that
  avoid simple blaming.
• Develop practical solutions that fix it forever.

7
Finding the Faultlines

• We have learned to look in these two places
• Customer Complaints
• ISO audit Corrective Action Requests (CARS)

8
COMPLAINTS

• Individual complaints are like viewing the
  company using tunnel vision.
• You can get trapped in the specific details of
  any one case.
• Instead look at a broad range of similar
  complaints looking for patterns.
• These patterns appear to be families of
  complaints.
• The underlying causes of these patterns are what
  we are looking for.
• Doing this is easier if you have a comprehensive
  customer complaint database.

9
ISO Corrective Action Requests

• In a similar manner look into the requests for
  corrective action that are written as a result of
  internal ISO audits.
• Are there relationships between complaints and
  CARS?
• As with complaints, you need a database of CARS.

10
Steps in searching for that hidden FaultLine

• Work backwards from the visible symptom of a
  hidden faultline.
• The visible symptom is evidence that you have a
  problem.
• The kind of problem we are interested in shows up
  multiple times and sometimes in varying places
  with often a variety of symptoms.
• You dont know this at first because you start
  with visible symptoms.
• Sometimes this backwards analysis uncovers just a
  local issue that hardly counts as a faultline.
• In that case you find the cause of the problem
  and fix it.

11
Variety of Problem-Solving Tools

• Fishbone Diagram (Cause and Effect).
• The Five Whys.
• Systematic Root Cause Diagramming Methods
• Commercial systems (see bibliography)
• Computer programs and chart-based analyses.
• Chronological timeline
• Logic Tree
• Whats Different Analysis, also known as
  IS/IS-NOT
• (Kepner Tragoe)
• Be prepared to apply multiple tools as the
  circumstances dictate

12
Two Models for Uncovering Faultlines

• Single Investigator
• Has a virtual team lurking in the background.
• Can use all the tools described later.
• BUT might be biased or jump too quickly to
  conclusions.
• AND might miss something hidden under the
  surface.
• May fail to come up with a good mix of corrective
  actions.
• Team with Facilitator
• Expensive to get everyone together.
• Used when a highly visible problem really needs
  big-time attention.
• Used when many departments touch the problem.

13
If You Need a Team ...

• Include people who know something about what
  happened.
• But hold down the size of the team! 5 - 7 seems
  optimal
• Want persons with different expertise and
  backgrounds. May want a vendor or a customer
  representative on the team if appropriate.
• Some team members might feel guilty!
• The guilty-feeling persons should not be
  spotlighted - we need information and not
  remorse.
• Hold kick-off meeting - carefully define the
  problem.
• Determine what sorts of information are likely to
  be needed.

14
Gather Information

• Interviews, copies of procedures, copies of logs,
  charts, test results, reports, photographs,
  maintenance records, audit reports, process flow
  charts and diagrams.
• Has this happened before? Retrieve reports from
  earlier investigations.
• This is why archiving investigations in databases
  is useful!
• Do not assume anything - fill in the details with
  facts.
• Expect the unexpected - look for the surprise!

15
A Lesson In Assuming
Bill owns a company that manufactures and
installs car-wash systems. Bill's company
installed a car-wash system in Frederick, MD.
These are complete systems, that include not only
the car wash itself, but also the money-changing
machines.
16
A Lesson In Assuming
Lots of money turned up missing - was it the
manager? Or had someone stolen the key from the
manager to make a copy? Bill just couldn't
believe that his people would do that, so they
set up a camera to catch the thief in action.
Well, they caught him on film!
17
It was not just one bird there were several
working together. Once they identified the
thieves, they found over 4000 in quarters on the
roof of the car wash and more under a nearby tree.
18
(No Transcript)
19
A Lesson In Assuming
No matter what the circumstantial evidence may
be, dont jump to conclusions until you have all
the facts. In this case, the new owner made the
assumption that sincea) Money was missing on a
regular basis.b) The machines were not being
broke into (no damage).c) The only other keys
would be the dealer or one of the dealers
employees. that it must be theft by the dealer
or a stolen key. WRONG!!
20
Key Tool for Identification of Hidden Faultlines

• The Logic Tree

21
How to Construct a Logic Tree

• For training purposes we will play around with a
  trivially simple case
• First define the PROBLEM by examining symptoms
• We lost 20 hours of production
• The customers plant had to be shut down
• An employee broke his leg
• My son was ticketed for speeding
• Search for the pain, where does it hurt?
• The problem has a so what dimension, check why
  do we care?
• The cost of a ticket is an OUCH!
• The possible increase in insurance premiums
  HURTS!
• Your son has run afoul of the LAW not a good
  thing.

22
Construct the Sequence of Events and Conditions

• Begin with the bad thing.
• Ask How did that bad thing happen? or What
  immediately preceded the problem event to
  directly cause it to happen?
• EXAMPLE
• My son was driving his car AND His speed was
  35 AND The speed limit was 25 AND A Police
  Officer observed him.
• Think in terms of events and all of the
  necessary conditions that conspired to cause
  something bad to happen. In this example, the
  two events and two conditions had to all be
  present to lead to the end result.

23
Speeding Logic Tree
Each box contains a single item. Avoid
statements like Driving his car at 35.
24
Building The Logic Tree

• The structure looks like a sideways tree.
• It spreads out with multiple limbs.
• Develop each limb of the tree by asking What
  caused this to happen?.
• Capture events and conditions necessary to
  describe what happened - working backwards.
• If you dont know the why, terminate that
  branch with a ? mark. You may need to research
  that limb further.
• Eventually each branch ends.
• Judgment is required here. Dont terminate a
  branch prematurely (you may miss a significant
  organizational fault, but on the other hand dont
  keep going back forever to the origin of time.

25
What the Logic Tree Looks Like
26
Speeding Logic Tree
?
Need to investigate why he was Not aware that
speed limit was 25
27
Speeding Logic Tree
The root cause here might be attributed to
Inattention. BUT why the Inattention?
28
Try a Simple Example

• Take these eight statements and identify the
  symptom of the problem - then draw the events and
  conditions in a Logic Tree Chart.

29
Eight Statements

• Did not see debris
• Driving to work
• Left by previous car?
• Got a flat tire
• Looking backward to pass car
• Debris shredded tire
• Slow car in front
• Debris in road

30
Case of the Shredded Tire
Case of the Shredded Tire
31
After The Logic Tree is Constructed ...

• Identify Direct Causes.
• Key events or conditions (e.g., ran over
  debris) that led directly to the undesirable
  event.
• Appear to the untrained person as the root
  cause but is not.
• Identify Contributing Factors.
• They have an influence on the problem, but if
  they were not present, the event could still have
  occurred.
• Example Talking on the cell phone

32
Now Look Deeper

• Identify Latent Causes.
• Affect not only this incident but influence
  spreads over a wide area and could generate many
  other similar incidents.
• The process of checking for access to the passing
  lane

33
Definition of Latent

• Present but not visible or Active
• Dormant
• Quiescent

34
Searching for Latent Causes

• One reason for identifying Direct Causes and
  Contributing Factors is to avoid calling them
  Latent Causes.
• Direct Causes and Contributing Factors affect
  this particular case.
• Fixing these factors is sometimes called
  Containment.
• Latent Causes will generate new problems of a
  similar nature at a later date.
• Addressing Latent Causes leads to sustaining
  corrective actions.

35
Finding a Latent Cause

• In the speeding example, Speed was 35 is a
  Direct Cause to Ticketed for speeding, but not
  the latent cause.
• Latent Causes underlie Direct Causes.
• Latent Causes are at the end of the
  cause-and-effect chain yet still within the
  control of the organization.
• Although the Direct Causes lead directly to the
  problem, the Latent Cause sets up circumstances
  to bring about the Direct Causes.

36
Test for statements identifying latent causes

• If a statement merely summarizes a bit of factual
  information about something that took place it is
  not a good latent cause statement.
• Example Pipe broke
• This is an accurate statement describing what
  happened. BUT the statement does not drill deep
  enough beyond describing what happened, hence it
  is not identfying a latent cause.

37
Latent Cause Tests

• Events are not typically latent causes. Latent
  causes are more likely conditions that allowed
  events to lead to the (usually) undesired effect
  .
• Think in terms of inadequate systems, processes,
  and procedures.

38
Other Tests for a Latent Cause Statement

• If you were to remove the latent cause, or fix
  it, or change it so that the influence it had
  before is gone...the problem should go away
  permanently.
• Sometimes it might take fixing or removing more
  than one thing, in that case you have more than
  one latent cause.
• One of the causes is necessary but not
  sufficient.
• This shows up as fixing the problem under certain
  circumstances but not all circumstances.

39
WARNING Symptoms are not latent causes.

• Symptoms partially describe the problem.
• Symptoms tell you something about whats wrong.
• BUT Fixing the symptom rarely stops the problem
  from happening again.
• We are having processing problems at the
  customer site, and their filters are showing
  evidence of a solid contaminant in our product.
  
• The solid contaminant is a only a symptom of the
  underlying cause.
• Ineffective corrective action They should
  switch to larger filters.

40
Corrective Actions

• After the Logic Tree chart is completed
• Check each box and ask if there is anything that
  can be done about it.
• Build a corrective action list from these ideas.
• Corrective Actions should
• Be practical and achievable.
• Reduce the likelihood of problem repetition.
• Be compatible with other departments or
  functions.
• Be accountable in terms of persons and time.
• Be sure you have done something about the Latent
  Causes and the various contributing factors.

41
EXAMPLE

• The Case of Something That went Wrong?

42
Examples of Actual Investigations
43
EXAMPLE

• For the want of a nail, the shoe was lost for
  the want of a shoe the horse was lost and for
  the want of a horse the rider was lost, being
  overtaken and slain by the enemy, all for the
  want of care about a horse-shoe nail. Benjamin
  Franklin, Poor Richards Almanac

44
EXAMPLE OF AN INEFFECTIVE LATENT-CAUSE ANALYSIS

• Complaint Five skids are misidentified. Labels
  exhibit code 160200 instead of 160280.
• Latent Cause The latent cause of this error is
  that the label was not generated with the correct
  code
• IS THIS CORRECT?

45
ANOTHER EXAMPLE OF AN INEFFECTIVE LATENT-CAUSE
ANALYSIS

• Customer profiles define shipping requirements.
  If the shipment arrived at the wrong temperature,
  it is because the temperature was not in the
  customer profile.
• HAVE WE FOUND THE LATENT CAUSE?

46
The Case of the Missing Bar-Code Label

• Complaint Description Section
• Missing bar code labels - customer requires
  bar-code labels on every box showing the part
  number.
• Investigation Section
• All messages are in place for customer to
  receive bar-code labels on their shipments.
  Order Entry tested a dummy order to make sure
  everything was in place for them to receive them
  on the next order and the test ran perfectly. We
  can only conclude this was a system-related
  problem that should not occur again.
• WAS THIS AN EFFECTIVE INVESTIGATION?

47
A Less than Effective Latent Cause and Corrective
Action

• Latent Cause Section
• Isolated incident that may have been a
  system-related problem. All procedures are in
  place for customer to get bar-code labels. Test
  confirmed this.
• IS THIS A LATENT CAUSE?
• Corrective Action Section
• Make sure before printing a bill of lading that
  bar-code indicator is set to N in other words
  do not bypass bar code labeling. This has been
  noted on customer profile.
• IS THIS AN EFFECTIVE CORRECTIVE ACTION?

48
The Case of the Scrambled Boxes

• Complaint Description
• Customer received sample with two different
  labels on one box. One label read 248-1050 and
  the other label read 348-012002. The material
  ordered was 248-1050.
• Investigation
• There were two orders from the customer
  scheduled to ship. One order was for 100 pounds
  of 348-012002 and the other for 100 pounds of
  248-1050. As the technician was processing the
  samples the shipping labels were attached to the
  wrong boxes. In essence it is inattention on the
  part of the technician which resulted in these
  orders being labeled inappropriately.

49
Next Step..

• Latent Cause Analysis
• The latent cause was human error, affixing
  address labels on two orders incorrectly.
• This is also a procedure short coming. We do
  not, in detail, define the steps that should be
  taken by the technician to eliminate the
  potential for mixing sample orders.

50
Following Step...

• Corrective Action
• We have met with the technician responsible for
  sample shipments, we reviewed the incident with
  the technician. This issue will be discussed
  with all technicians in the next team meetings.
• The procedure covering the preparation and
  shipment of sample orders will be reviewed and
  updated as necessary to address this type of
  issue.

51
On changing procedures ...
Warning labels and large instruction manuals are
signs of failures, attempts to patch up problems
that should have been avoided by proper design in
the first place. The Design of Everyday Things,
by D. A. Norman, 1988, Doubleday.

• Procedures are the scar tissue of past mistakes
• Sherry Poriss, Performance Review Institute,
  Warrendale, PA

52
NOW YOU TRY IT ...

• The Case of the Capsized Ferry
• Get into teams
• Review the facts of the case,
• What is THE PROBLEM?
• Build a Logic Tree with post-it-notes
• Identify key Direct Causes
• Identify Contributing Factors
• Identify Latent Cause(s)

53
Other Tools
54
Chronological Timeline Tool

• The Logic Tree does not follow a timeline.
• Can be confusing.
• Building a timeline helps sort-out the order in
  which events took place.
• Might offer some cause/effect clues.
• BUT is often multiple tracks with parallel lines
  which can itself be confusing.

55
For Simple Problems, A 5 Whys Analysis May Work

• Why is our production rate so low?
• Well, the preheater pressure is maxed out.
• Why is the preheater pressure maxed out?
• Because the polymer melt viscosity is too high.
• So why is the polymer melt viscosity too high?
• Because the polymer melt temperature is low.
• Why is the polymer melt temperature low?
• The second-stage reactor temperature is low.
• Why is the second reactor temperature low?
• Because the process chemistry requires it to be
  low.

56
Need Another Why

• Why does the process chemistry require it to be
  low?
• Hey ... if we go to this new catalyst we can run
  the second-stage reactor at a higher
  temperature.
• Just hit upon a corrective action!
• BUT Have we considered other factors?
• This is really just one branch of a Fault Tree.

57
Sometimes the Tree Doesnt Help Much

• If you have a process with numerous inputs a
  variety of process variables and an unpredictable
  output, Logic Trees are not the best way to find
  out why the output is varying.
• Example Product X is frequently off-spec in
  color. Color is influenced by temperature,
  pressure, production rate, and an impurity in the
  feedstock.
• Statistical techniques are required.
• If you have a long list of potential causes.
• Not sure which ones need closer examination.
• The Whats Different tool is then useful.

58
Whats Different Analysis

• Use this tool when it is not clear what changed
  to lead to the undesired effect.
• Compare problem cases to no problem cases.

59
Whats Different Analysis

• Break the problem down into
• WHAT IS THE PROBLEM vs WHAT IS NOT?
• WHEN IS THE PROBLEM vs WHEN IS IT NOT?
• WHERE IS THE PROBLEM vs WHERE ISNT IT?
• THE SEVERITY OF THE PROBLEM?
• Look for a pattern in the difference
  comparison.
• Based on Kepner and Tragoe IS/IS NOT analysis

60
EXAMPLE Excessive leaks in Unloading Hoses

• Customer complained that carrier hose fittings
  often failed hydrostatic tests at their facility.
• Truck sent back to originating plant with
  customer complaint.
• Yet hose fittings tested OK back at the terminal.
• Logic Tree working backward through series of
  events did not pick up on any obvious causes.
• Special hoses and connections for that customer
  site?
• Poor unloading procedure?
• Incompetent unloading personnel?
• Inadequate unloading facilities?
• Events and conditions seemed somewhat
  contradictory.

61
IS/IS NOT Analysis Kepner Tragoe

• Look for instances where the problem does NOT
  occur.
• Compare to where and when it DOES occur.
• In fact the customer had another site where the
  leaks were not occurring (IS NOT).
• Asking questions of the driver involved revealed
  a critical DIFFERENCE between the two sites.
• At the no-leak site the standard practice was to
  have a cleaned hose from the previous trip
  off-loaded and waiting in an interior location.
• An arriving truck swapped hoses they used the
  previous trucks hose to unload the arriving truck
  - the new hose was stored inside for the next
  load - NO LEAKS!

62
IS/IS NOT Questions lead to new insights

• When leaking hoses from problem customer site
  were tested later they had warmed up and did not
  leak.
• Good hoses left outside in the cold of winter
  showed leaks.
• DIRECT CAUSE
• Hoses that cooled down had couplings that shrank
  away from the hose material allowing for slight
  leaks.
• LATENT CAUSE
• Inadequate unloading procedure that failed to
  take into account the affect of changing
  temperatures on hose coupling.
• CORRECTIVE ACTION
• Apply same hose swap with other customer site.

63
Corrective Action

• Develop corrective actions with the team or
  virtual team
• Monitor corrective action implementation
• Verification.

64
Verification

• Review the corrective action after sufficient
  time has elapsed since it was implemented.
• Check that the corrective action was indeed
  implemented and still in place.
• Determine if there have been any recurrences
  after the corrective action was implemented.
• Dont have tunnel vision -- if the corrective
  action was implemented because of a problem with
  Customer X, it is not effective if it occurs
  later with Customer Y.
• If there is a recurrence, the corrective action
  must be deemed Ineffective and the investigation
  re-opened.

65
Some key tools for developing effective
corrective actions

• Process Mapping
• RACI
• Applied Behavioral Analysis.
• Cognitive Psychology of Human Error.
• ISO Standards.

66
PROCESS MAPPING

• Often a latent cause is tracked to some
  malfunctioning part of a process.
• Unfortunately processes work horizontally while
  management is designed vertically.
• Who is in charge of a horizontal process?
• Who can take managerial charge of a process that
  needs to be fixed?
• The answer is often unclear.
• Mapping the process with all stakeholders can
  force the issue.

67
Silos vs Processes

• Vertical organizational silos handling a process,
  but who is the process owner?

Silo 1
Silo 2
Silo 3
Process
68
PROCESS MAPPING

• The action of mapping a process with the various
  silo owners present can lead to turf wars but
  it can also lead to agreement on who owns
  slices of the process.
• Then you gain enough ownership to bring about
  managerial impact on the changes needed for the
  corrective action.
• It helps to be building this for ISO - adds
  respectability to the mapping process!

69
PROCESS MAPPING

• For the purpose of building a corrective action
  you need to define the process needing mapped.
• Then identify the persons most likely to be
  process owners.
• Bring them together and with post-it-notes
  construct the elements of the process paying
  particular attention to the information flow
  through the process.
• Search for ways to change the process so that an
  undesirable incident is less likely to occur.

70
NEED RACI

• Although you may build acceptance regarding the
  process map there is still room for
  misunderstanding among the owners regarding
  exactly who is responsible for doing what.
• The RACI tool helps clarify the details so that
  specifics of process interaction are precisely
  defined.
• RACI stands for RESPONSIBLE, ACCOUNTABLE,
  CONSULTED, INFORMED

71
RACI TOOL

• RESPONSIBLE Who is the person responsible for
  carrying out the process step under examination?
• ACCOUNTABLE Who is the person accountable for
  seeing to it that the person responsible carries
  out the task?
• CONSULTED Is there a position or positions
  within the organization who needs to be consulted
  about the task?
• INFORMED Is there anyone who needs to be
  informed about the execution of the task?

72
RACI Charting

• Identify all activities and decisions necessary
  to run the day-to-day process effectively
• Identify
• Who is Responsible (R)
• Who is Accountable (A)
• Who must be Consulted (C)
• Who must be Informed (I)
• Document on charts for reference

73
RACI Definitions (cont.)

• Role Players are the positions in the
  organization that have a task to perform.

Role Players

• Activity
• An action or decision that is one of several
  sequential steps in the completion of a business
  process.

74
Analysis of RACI Chart

• Look For
• No or too many Rs
• No or too many As
• Too few As / Rs
• Every box filled in
• Lots of Cs
• Lots of Is

75
Get Feedback and Buy-In

• Show the RACI chart to the representative groups
  of people covering the roles on the chart
• Capture their comments and revise the RACI chart,
  if appropriate

76
HUMAN ERROR

• There are circumstances where an undesirable
  event was partially caused by a person within the
  system who failed to do something that needed
  done.
• Dont just blame them! Judge how the failure
  occurred.
• Use the following test
• Didnt know how to do it.
• Didnt know to do it.
• Wouldnt do it.
• Didnt know they didnt do it.

77
Corrective Actions for Behavior
Determine whats needed, i.e. resources,
training. Remove barriers.
Cant Do
If root cause is determined to be human error
or simply a case of people not doing what they
are supposed to do. Conduct a Root Cause
Analysis for Behavior.
Develop and communicate clear and agreed-upon
RACI.
Doesnt Know To Do
Ensure the right things are being reinforced,
especially when there are competing behaviors.
Wont Do
Slips
Need Mistake Proofing
78
Typical Corrective Actions
Cant Do

• The cant often relates to undeveloped skill or
  knowledge. Training (one of the knee-jerk
  automatic corrective actions) is the relevant
  corrective action.
• Possibly cant has more to do with unavailable
  tools or materials.
• Cant might even be involved with politics or a
  decision that someone is not allowed to do it.

79
What do you do with Didnt know how to do it?

• Cant do it relates often to undeveloped skill
  or knowledge.
• Sounds like training is needed.
• Possibly tool availability or use might be an
  issue - computer example.
• Ask ISO
• 6.2.2 Competence, awareness and training
• Someone needs to determine the necessary
  competence.
• Someone needs to offer the necessary training.

80
Doesnt Know To Do Situations

• When the root cause is due to individuals who can
  do the behavior, but do not know that they must
  complete the task, then use the RACI tool.
• RACI defines the types of participation and
  involvement of all impacted team members.

81
What do you do with Didnt know to do it?

• The person responsible did not know they were
  responsible?
  
  
  The person
  accountable needs to straighten this out.
• And check ISO
• Update job description etc.
• 6.2 Human resources
• 9004 points to the involvement of people.by
  defining their responsibilities and authorities

82
What do you do with Wouldnt do it?

• Put a gun to their head?
• Tell the boss?
• (Bosses are not always effective)
• Application of Behavioral Principles can help.
• Behavior affected by
• Antecedents
• Consequences
• Corrective actions can be centered around
  creating the right mix of antecedents and
  consequences.

83
Applied Behavioral Analysis

• Solution based on a long history of psychological
  research.
• The term Behavior is precisely defined.
• Behavior is measurement based.
• Studies show behavior can be triggered by
  incidents preceding it called ANTECEDENTS.
• Behavior can be strengthened by the application
  of resulting incidents following the behavior
  called CONSEQUENCES.

84
Wont Do Situations

• We address the individual who Wont Do a task
  through a Consequence Plan.
• Wont Do usually means
• A lack of natural positive consequences exist for
  task completion.
• Naturally occurring negative consequences exist
  for the desired behavior.
• Answer
• Plan and provide positive consequences for the
  right behavior (to override the naturally
  occurring negatives.)
• Analyze whats being rewarded (from the
  performers point of view.)

85
Building a Consequence Plan

• If they dont do it there needs to be a
  consequence (or twoetc) clearly following the
  inaction that is recognized by the non-doer as
  undesirable punishment.
• When they DO do it there needs to be resultant
  consequences that have a positive, reinforcing
  effect.
• Antecedents include clear messages that it needs
  to be done and this is why.etc.
• Antecedents can instill a belief that doing it
  is a good thing.

86
What do you do with Didnt know they didnt do
it?

• Asleep at the switch?
• YES, The psychology of human error
• Slips are a class of human error involving
  unawareness of the error.
• Occurs during skilful acts.
• An example could be typing errors
• Corrective actions do not include
• Punishment
• Re-training
• Reviewing procedures

87
Types of Slips

• Capture Slip
• A pattern similar to another pattern (the desired
  behavior) is triggered because it has a higher
  frequency of execution.
• Example Calling the wrong number because you
  are used to calling it.
• Associative Activation Slip
• Your actions are applied on the wrong thing, the
  intention is correct but is misapplied.
• Example The case of the wrong dog.

88
Types of Slips

• Loss of Activation Slip
• The intent begins the action but short term
  memory is suddenly loaded with something else and
  you find yourself part-way through an action but
  cannot remember why hence cannot complete the
  act.
• Example Why did I go upstairs?

89
Correcting Slips

• Change the system under which the slip occurred.
• Redesign the way people interact within the
  system.
• Put into place immediate feedback loops.
• Design in checker routines.
• Re-program the computer.
• Include Mistake Proofing (Poka yoke)

90
More Human Error Types

• Rule based errors
• A rule (known to be a good and true rule) is
  applied to the wrong thing in the wrong setting
  hence causing a problem.
• Example
• RULE - If your car is going too fast, put on the
  brakes.
• But suppose you discover you are driving on ice?

91
More Human Error Types

• Knowledge-based errors
• The situation is too far out-of-control for rules
  to work
• Rules were not constructed for this situation
• Instead you must figure out what to do from
  scratch based upon your knowledge of the system.
• Unfortunately your analysis can lead you
  off-track and an error is made.
• HELP? The more information you can provide the
  better

92
HUMAN ERROR What Do Researchers Tell Us?

• Instead of blaming the human who happens to be
  involved, it would be better to try to identify
  the system characteristics that led to the
  incident and then to modify the design
• One major step would be to remove the term
  human error from our vocabulary.
• One conviction that seems to be shared by all
  members of the field studying human erroris a
  rejection of the conventional approach to error
  prevention, that of TRAINING and PUNISHMENT.

93
FINAL MESSAGE

• Search for patterns underlying a family of
  complaints.
• Improving corrective actions means looking
  further and deeper than fixing the direct cause
  of a problem.
• Let ISO help cue that search for deeper places to
  look.
• Change the system instead of blaming the
  performers in it.
• Identify and map out processes looking for ways
  to improve the processes so the problem is less
  likely to recur.
• Apply principles of behavioral reinforcement to
  susstain a true change in behavior.

94
THE END.References

• Guidelines for Investigating Chemical Process
  Incidents American Institute of Chemical
  Engineers (1992).
• The deductive approach starts at one point in
  time (the event) and looks backward in time to
  examine preceding events.
• See also Current Reality Tree tool described in
  Goldratts Theory of Constraints by H. William
  Dettmer (ASQ Quality Press, 1997)
• In building a Current Reality Tree, we work our
  way from Undesirable Events back through the
  chain of cause and effect to root causes.