Do I think I know what happened?

There are a number of reasons why we can think we know what happened during the interviews – despite the fact that conclusions should not be drawn until all the information has been collected and analyzed.

For example, if the last three incidents investigated were falls from height due to poor ladder maintenance, then representativeness bias may lead us to presume poor maintenance is also the cause of the fourth incident involving a fall and a ladder when it happens. Representativeness bias can result in matching current events back to our previous experiences, which can lead to perceived similarities in situations that aren’t there in reality.

Alternatively, availability bias can result in our over-valuing of information simply because we can easily remember it. Things that are more impactful, emotional or simply things that happened more recently can be included in our investigation, simply because we can recall them quickly and easily. However, this remembered information may have no bearing on the current situation and should be treated with caution.

Information can also be collected to fit previous patterns of events, rather than be collected to explore what has actually happened in this specific instance. For example, if we fixate on ladder maintenance as the causal factor (rather than perhaps poor access planning that resulted in workers having to carry tools and materials up the ladder without being able to maintain three safe points of contact) confirmation bias can lead us to ask questions that generate information to support the problem of a lack of maintenance, rather than questions that seek new information. This can limit learning as the ladder could easily have been badly maintained, but that may have had no influence on this particular incident – and confirmation bias means deeper learning about planning and management is missed.

Overconfidence in conclusions can also make us vulnerable to conservatism in belief revision, as we then struggle to change our thinking, even in the face of new contradictory information.

Jumping to conclusions and thinking you know what happened before all the information has been collected should always be avoided. Conclusions can and should only be drawn once all the information has been collected and analyzed.

Suggested actions:

For Interviewers

• Be mindful not to make any judgements or reach any conclusions until the information collection process and its analysis has completed.
• Actively listen to what the interviewees are saying.
• Let the interviewees talk freely, don’t interrupt or finish their sentences for them – keep quiet and let them explain what they think actually happened.
• Prior knowledge of a process or system of work can help, but be careful you are not assuming things happened in the same way in this particular instance.
• Reflect on your lines of questioning, are you trying to find evidence to support a preconceived idea?
• Be sure to explore all avenues of inquiry, not just those that back up your ideas. Use a checklist of interview question topics to ensure nothing obvious is missed.
• Have a non-professional member in the team to bring a neutral starting perspective, or run your thoughts past a colleague to get a fresh perspective.

For Managers

• Be flexible in your resourcing and ensure enough headroom in scheduling to allow colleagues to collaborate on investigations as needed.
• Create investigative teams or pairs with different expertise and experience to bring different perspective to the process.

Find out more about availability bias, representativeness bias, confirmation bias and conservatism in belief revision here.