Cognitive debiasing 2: impediments to and strategies for change

The follow up paper on cognitive debiasing. This paper covers a scheme of cognitive change relating to debiasing, the constraints to change, and then a range of techniques directed at cognitive and affective debiasing.

[** Note that there’s more systematic and recent reviews on debiasing techniques. I’ll cover some of these in the near future.]

Biases are not a bad thing. The word bias has a negative connotation, but most biases serve or did serve an essential role in human sensemaking and survival. Almost all of our senses – sight, hearing etc. rely on bias to filter out noise and for sensory discrimination. Also regarding heuristics, other work has shown that they can frequently outperform slower, more deliberate algorithmic approaches or statistical/probabilistic approaches.

An early example of heuristics and debiasing was from 1772 where Ben Jonhson, an English playwright, referred to “moral algebra’ to improve his judgements and avoid rash decisions” (p65).

Debiasing is said to rarely come about via a discrete single event but rather through succession of stages. The stages are: precontemplation, contemplation, preparation, action, maintenance.

Precontemplation: The stage where many are thought to be. Unaware of the powerful influence unconscious factors have over their reasoning. Because they may not realise that cognitive and affective factors influence their decision making, they see no reason to take action to change their thinking strategies.

Introducing ideas at this stage is to create a need for change is a critical step for debiasing. This can include conventional forms of exchange like huddles, training, conferences, workshops etc. A single experience can also push change, like a critical event that’s particularly emotionally laden.

Of course, increasing awareness of the need for debiasing doesn’t guarantee debiasing occurs. One constraint is our own confidence in our judgements (and the lack of feedback that our judgements are not well calibrated). We may not believe we’re vulnerable to biases.

Also, “it is a human tendency to bolster existent beliefs rather than searching for new approaches, and it is easier to stay with the status quo rather than make efforts to learn new approaches and change current practice” (p66).

As noted in the first paper – “increased intelligence does not protect against biases”. [** Daniel Kahneman also quipped in an interview that after spending a lifetime studying biases, he’s still no better off avoiding them.]

A challenge of debiasing is that many act unconsciously and that “the same kinds of biases that distort our thinking in general also distort our thinking about the biases themselves” (p66).

Strategies for debiasing

Three groups of interventions are suggested: educational, workplace, and forcing functions.

The authors list ~27 different debiasing techniques, so I can only provide a few examples from each above group – adapted from tables 1 & 3 (pp67, 68).

Techniques under educational strategies rely on enhancing the knowledge and awareness to as to debias in the future. Workplace strategies rely more on in-situ cognitive processes and changes to the environment.

Forcing functions include absolute constraints, which prevent erroneous responses like removing potassium solutions from the ward so they cannot be mixed up, to explicit “if this then this” rules which encourage a desired response.

The authors then ask “are there specific cognitive pills for cognitive ills?”. Some biases have their origins in adaptive evolutionary terms, like availability and representativeness heuristics. Others are acquired through social and individual experiences.

They argue that perhaps the more evolutionary heuristics are probably more resistant to change and may need several debiasing strategies and interventions. Conversely, cultural and sociocentric biases are easier to learn and therefore, may be more amenable to change.

In short, it’s unlikely any one-size-fits-all technique will uniformly suit all biases and all contexts.

In wrapping up, they argue that in moving forward:

1.      A major challenge will be identifying the parameters for change. E.g. how to design the workplace to minimise bias and error traps. Which interventions should be selected?

2.      Following on from 1, altering the environment is likely to be critical. Many debiasing strategies focus on the individual, but better environmental design are forcing strategies that may reduce the consequences and variability from rapid decisions. E.g. better shiftwork arrangements to reduce sleep disruptions, reducing cognitive overload, extensive human factors integration via design of equipment or the structuring and sequencing of tasks etc.

3.      Evidence around debiasing is, in general, pretty shaky

4.      Training has traditionally focused on declarative knowledge (knowing what or information-based) compared to procedural knowledge (knowing how or application-based). More efforts towards procedural-knowledge may help with minimising sources of bias and noise. They note that avoiding bias is correlated with critical thinking.

5.      Many factors are difficult to control in workplaces. For healthcare, this includes patients, families, unique medical conditions, resource constraints, surges etc. This pushes people into less-than-ideal positions for decision making.

6.      There’s considerable scope in training, mentoring and other initiatives around intuition and heuristics. They note that “Type 1 processing is essential to cognitive functioning and generally serves us well; in fact we could not live without it ... Given that the vast majority of our daily decisions involve Type 1 processes,83 there is considerable ground to be made in educating intuition” (p70).

Link in comments.

Authors: Croskerry, P., Singhal, G., & Mamede, S. (2013). BMJ quality & safety, 22(Suppl 2), ii65-ii72.