Latest / Safe As Podcast / E43: How 'no blame' can potentially subvert learning and improvement

Transcript

E43: How 'no blame' can potentially subvert learning and improvement · Safe As Podcast

• 0:06
  What if no blame logics is short circuiting effective
• 0:09
  accountability? Let's find out.
• 0:13
  Good day everyone. I'm Ben Hutchinson and This is
• 0:15
  Safe as a podcast dedicated to the thrifty analysis of safety,
• 0:20
  risk and performance research. Visit safetyinsights.org for
• 0:24
  more research. Today's study is from Share it
• 0:27
  ET L 2023, titled The Unintended Consequences of No Blame
• 0:32
  Ideology for Incident investigation in the US
• 0:35
  Construction Industry in Safety Science.
• 0:39
  In this study, the transcripts of 34 experimentally simulated
• 0:44
  incident investigation interviews from safety experts
• 0:47
  were evaluated, simply exploring how safety experts made sense of
• 0:52
  simulated accidents and particularly interested in how
• 0:56
  the safety experts described and allocated blame.
• 1:00
  Providing background from the research, they note that
• 1:03
  Conklin's neat mantra that blame fixes nothing is so prominent in
• 1:07
  safety discussions across North America that is arguably now
• 1:12
  simply part of the vernacular. Prior work from the UK found
• 1:16
  that construction workers fully expect blame and punishment as
• 1:19
  part of site safety. For some, they see it as a
• 1:23
  necessary counter to their need, and in some cases choice to bend
• 1:27
  and flex site rules in order to get the job done.
• 1:31
  Other workers argue that a unidimensional focus on creating
• 1:34
  a no blame environment carries its own safety risks.
• 1:38
  Moreover, others have argued that systemic safety's
• 1:42
  absolution of individual performance and accountability
• 1:46
  meant workers were effectively written out of the safety story,
• 1:50
  which in turn made it impossible to hold them accountable for
• 1:53
  poor behaviour. Also, the language used in
• 1:56
  investigations can be relevant as it can shape and influence
• 1:59
  all parts of the process. Even the order of that
• 2:02
  information announced presented can influence judgments of
• 2:05
  culpability. Some firms have shifted to
• 2:08
  incident evaluations rather than investigations to avoid punitive
• 2:12
  language associated with criminal investigations.
• 2:15
  In all, the blame has a dubious history in construction, such to
• 2:19
  the extent that blame has even been weaponized against workers.
• 2:23
  So what did they find? A kind of bias was found not a
• 2:27
  psychological bias, but rather, quoting the paper, a dogmatic
• 2:30
  adherence to no blame within the investigative process.
• 2:35
  This dogmatic adherence bias was called new blame and was noted
• 2:39
  to have detrimentally influenced and shaped investigation
• 2:42
  interviews. This means that a rigid
• 2:45
  commitment to no blame actually became a problem itself,
• 2:48
  changing how investigations were conducted.
• 2:51
  Quoting the paper you blame renders interviewers reluctant
• 2:55
  to unpack the people in the process and instead directs
• 2:59
  focus on things easily blamed without consequence, such as
• 3:04
  inanimate objects, organisational procedures, or
• 3:06
  paperwork. In other words, investigators
• 3:10
  avoid looking closely at what people actually did, preferring
• 3:14
  to blame things that couldn't argue back, like equipment or
• 3:16
  paperwork. They argue that without
• 3:19
  understanding how workers interact within their work or
• 3:23
  what went wrong this time, learning is inevitably limited.
• 3:27
  Interestingly, they say, at the most extreme level, new blame
• 3:30
  may take safety back to acts of God.
• 3:34
  New blame was often seen as a comment the investigator made to
• 3:37
  the worker before the questioning commenced.
• 3:39
  For instance, investigators would start with comments like
• 3:43
  we're not looking to fix people, we're looking to fix our
• 3:46
  processes and systems so that you guys on the frontline will
• 3:48
  be set up for success. In some cases, these preamble
• 3:53
  statements were presented in a rehearsed manner and with the
• 3:56
  language of hope clearly evident.
• 3:58
  That's a direct quote. Interviewers drew clear
• 4:01
  distinctions between people and processes, emphasising the need
• 4:05
  to focus on improving processes that support people when people
• 4:09
  inevitably make mistakes. The researchers note that
• 4:12
  despite the investigator providing reassurance to the
• 4:14
  workers, blame was still present.
• 4:17
  In fact, investigators, in trying to avoid blame, brought
• 4:21
  blame back into the conversation within the very first sentence.
• 4:25
  As the authors argue, using this type of blame avoidance language
• 4:28
  in the preamble statements before the questioning begins
• 4:32
  has the practical effect of actually placing blame at the
• 4:34
  very top of the agenda and quoting paper again.
• 4:37
  It becomes the first thing mentioned, elevating its
• 4:40
  importance within the situation for both interviewer and
• 4:43
  interviewee, and that sets the precedent for the rest of the
• 4:46
  interview. So put simply, by constantly
• 4:49
  saying no blame, the idea of blame ironically became the main
• 4:53
  focus right from the start and blame was positioned in the
• 4:57
  negative frame and associated with punishment.
• 4:59
  Despite being different activities.
• 5:02
  They argue that via new blame, blame is centrally positioned
• 5:05
  throughout the interaction from the very start.
• 5:08
  However, such early promises of no blame before the
• 5:12
  investigation has even started shouldn't be made, according to
• 5:15
  the authors, since it's unknown at this point what appropriate
• 5:18
  actions will be necessary. Further, early assurances of no
• 5:22
  blame become both presumptive and pointless, and in cases
• 5:26
  where disciplinary action is actually warranted according to
• 5:29
  paper, and these early assurances become just another
• 5:32
  management lie. Therefore, introducing promises
• 5:36
  that may have to be broken can damage trust.
• 5:39
  This highlights that making blanket no blame promises too
• 5:41
  early can be seen as dishonest. Later on, New blame was found to
• 5:46
  direct attention to things that could be blamed without
• 5:48
  immediate consequences for the interviewee, including poor
• 5:52
  materials or inanimate objects or lack of training, while also
• 5:56
  avoiding questions around the actions or behaviours of the
• 5:58
  workers who were directly involved in the event.
• 6:02
  So here, investigators frequently shortchanged the
• 6:05
  collection of information around behaviours and individuals,
• 6:09
  thereby not properly exploring the context of workers and how
• 6:12
  they fit into the incident. So instead of understanding
• 6:15
  where people acted as they did, investigators who are so
• 6:18
  cautious around this idea of blame focused on easy targets
• 6:22
  like faulty equipment or bad procedures.
• 6:25
  And while focusing on inanimate objects seemed to be sort of a
• 6:28
  logical direction to explore, it also provided A relatively
• 6:32
  comfortable space for inquiry. So blaming the staircase for
• 6:37
  instance, over an individual that fell.
• 6:39
  Blaming the staircase seems to be utterly unresponsive to
• 6:42
  criticisms or blame as opposed to individuals.
• 6:45
  So it makes blaming the staircase a really comfortable
• 6:47
  target. Also, potential cues relevant to
• 6:51
  the investigation surrounding individual behaviour were rather
• 6:55
  glossed over in preference to factors of the objects or system
• 6:58
  factors. So you blame directed attention
• 7:01
  towards the role of procedures and as the author state, the
• 7:06
  focus in objects and systems provides the blameable faceless
• 7:09
  organisation. Not all investigators avoided
• 7:13
  the role of people in the process, so some of them did
• 7:16
  actually focus around agency. But even in these cases, new
• 7:20
  blame still had some level of influence.
• 7:23
  So even when investigators tried to look at human actions, the
• 7:26
  new blame logic still pushed conclusions towards blaming
• 7:29
  impersonal organisational issues rather than understanding
• 7:33
  individual contributions. And this is said to be up to
• 7:36
  limit learning since organisations are still made up
• 7:39
  of individuals. Therefore you blame
• 7:42
  oversimplifies the situation and prompts A predetermined
• 7:46
  conclusion of organisational issues.
• 7:48
  Importantly, if responsibility cannot ever be attributed to
• 7:52
  individuals, then we will struggle to fully understand the
• 7:55
  actions of people and why they did what they did.
• 7:58
  Although I have to say I've never really seen this
• 8:01
  hypothetical no blame organisation in practise and
• 8:05
  even in the more progressive, just and restorative
• 8:08
  environments I've worked in, individual accountability and
• 8:11
  responsibility always existed. In fact we really didn't need
• 8:14
  any more incentives to blame people.
• 8:17
  But something I should highlight is we do need to separate
• 8:20
  responsibility and accountability and forward
• 8:24
  looking and backward looking accountability.
• 8:27
  Was that individual even aware of what they were supposed to be
• 8:30
  doing or what the expectations were and where they
• 8:33
  appropriately supported resourced to actually do what
• 8:37
  was expected? But anyway, what they argue is
• 8:40
  by shifting focus away from individuals, like always, it
• 8:43
  becomes harder to truly learn from incidents and understand
• 8:46
  why people made the choices they did.
• 8:49
  And finally, to reiterate, new blame at the extreme level may
• 8:52
  simply reintroduce the old concept of acts of God.
• 8:56
  Largely intangible are uncontrollable forces that are
• 8:58
  readily flammable. So what can we make of the
• 9:02
  findings? There's a bit to say here.
• 9:04
  Remember I mentioned before around separating things like
• 9:07
  blame, accountability and responsibility.
• 9:10
  Particularly accountability should be more forward looking
• 9:13
  rather than retrospective. In any case, some things that
• 9:16
  came out of this paper I think is acknowledging that there
• 9:19
  could be some sort of new blame logic at play.
• 9:22
  So recognising that saying no blame and having this no blame
• 9:25
  approach can ironically create a new blame discourse that still
• 9:30
  possibly hinders genuine learning and safety
• 9:32
  improvements. We should be careful around
• 9:35
  being overly premature with blanket no blame promises as
• 9:39
  this can reinforce the concept of blame and damage trust.
• 9:42
  If we can't stick to it later on.
• 9:45
  Maybe as I suggest, shift towards terms like incident
• 9:48
  evaluations to foster an environment of learning rather
• 9:51
  than punishment, probing deeper than surface causal issues, so
• 9:56
  looking beyond easily blamed factors like inanimate objects
• 9:59
  and procedures or paperwork, and focus on understanding how
• 10:02
  workers interact with their tasks.
• 10:04
  So normal work promoting perhaps the restorative just culture
• 10:08
  ideas which do actually have accountability and
• 10:12
  responsibility baked into them already, and focus on deeper
• 10:17
  systemic learning. Easier said than done, but
• 10:20
  prioritise fixing your overall environment, processes and and
• 10:24
  interactions. Can, might, maybe help frontline
• 10:26
  workers, but this still requires delving into how human action
• 10:31
  influences and is influenced by underlying factors.
• 10:34
  Also, I'm reminded here around an early quote from James
• 10:37
  Reason. So Reason argued that perhaps
• 10:40
  the pendulum may have swung too far in our present attempts to
• 10:44
  track down possible errors and accident contributions that are
• 10:47
  widely separated in both time and place from the events
• 10:50
  themselves. So what he was arguing is that
• 10:53
  while systems level analysis are probably absolutely critical and
• 10:57
  still to this day probably not emphasised enough, sometimes we
• 11:01
  can try too hard and move well away from the actual incident in
• 11:06
  trying to find really distant latent contributing factors and
• 11:10
  under emphasising the actual active event itself.
• 11:14
  And just remember, a limitation of this study is that it wasn't
• 11:17
  actually looking at real investigations, but rather it
• 11:21
  was using simulations. Still useful, still insightful,
• 11:25
  but it may not necessarily accurately carry over to the way
• 11:29
  that we do investigations sort of in the real world.
• 11:33
  That's it on Safe. As I'm Ben Hutchinson, please
• 11:36
  help share, rate and review and checkout Safety insights.org for
• 11:41
  more research. Finally, feel free to support
• 11:44
  Safe AS by sharing a coffee link in the show notes.