The 'Anchoring Bias' - in practice

A couple of videos here on the 'anchoring bias' - the phenomena where estimates and assessments are influenced by an initial estimate or even a spurious number or value.

I have found anchoring can be a problem when assessing risk or uncertainty on values in a business for project plan. For instance if evaluating risk to a project time-plan, if the existing planned duration is visible then assessments on upper and lower estimates will lie either side of the established planned value (very frequently but not always). If there is no established or baseline duration visible to the assessors then they will far more frequently provide lower and upper estimates that challenge the planned value. This is partly down to the anchoring effect and partly down to the way the problem is presented.

The role of an independent facilitator is to manage these type of phenomena and gather the best quality and most relevant inputs in the time available ("clarity efficiency"). If a plan with pre-existing values is shown to the assessors this could be an implicit message that 'the plan is basically sound but could vary somewhat'. This might lead people to frame their responses in terms of a percentage difference from the baseline or 'anchor point', however these variances will typically be too small due to over confidence and anchoring. The facilitator can manage this in part by stating that the values are sacrificial and providing sufficient time in meetings and workshops for challenge. Not providing an established baseline would also imply more uncertainty and remove the anchor. Real life situations are complicated by the facts that plans frequently contain some robust numbers and some wild guesses - often presented with equal gravitas.

If a baseline or existing value is not provided anchoring still comes into play. When asked to provide upper and lower estimates, whichever is provided first assessors will typically not move far for the next value. For this reason I will generally ask for an upper value first if estimating a risk, and a lower value first if estimating an opportunity. This acts somewhat protectively against overconfidence, however awareness of anchoring and overconfidence is a much better solution (see "Risk Judgement Training").

The first video explains the Anchoring Phenomena, and the second gives a real demonstration of anchoring (I dislike the 'Why are we all so stupid?" sentiment). Although a clear anchoring effect is demonstrated it is under 'laboratory conditions' of estimating the quantity of sweets in a jar. People and organisations would generally take a very different approach in a real world situation where meaningful consequences are recognised. However the anchoring effect is real and there is a risk that something might not be recognised as important as a result.


©2019 by intorisk. Proudly created with