Blog: HR analytics: Dealing with 'anchoring bias'

HR Analytics

HR analytics: Dealing with 'anchoring bias'

There are many biases that may exist in organizations. Some of them are evident while others may not be very easily observable. How can organizations do a better job and proactively address anchoring bias?
HR analytics: Dealing with 'anchoring bias'

I have been reading some rather emotional posts recently on the ethics shown by recruiters/companies in asking candidates for their current salary details. They ask questions like 'Why should a candidate share his current salary with a prospective employer?' Their argument is that the company has a budget for the role and they should pay accordingly as long as the candidate meets the role requirements.

The recruiter may want to know a candidate’s current compensation, for (one or all) the following reason(s):

  • He wants to find candidates who have a lower compensation than his budget so that he can attract the suitable candidates with a reasonable hike.

  • He wants to save time because most candidates will not be interested in a job that has a compensation lower than his current salary. Most candidates will be offended if they are taken through a selection process and offered a salary lower/equal to his current pay. 

  • The recruiter variable pay depends on the savings that he is able to deliver on his budget.


Making an offer to candidates based on their current salary is what is referred to as a classical anchoring bias, where the last earned salary becomes the anchor for the recruiter. There is a good probability that similar anchors could lead to a plethora of unintended biases based on gender, economic background of the candidate, schools attended... A large majority of firms could be guilty of advancing some (or all) of these biases. Research in the US on gender bias pegs average earnings of a woman at 78 cents to a dollar for men for the same job. Add ‘colour’ as a variable and this gap could go up higher. So, offering her a compensation package based on her current earnings will ensure a pay gap consistent with past research.

“Companies like #Google have invested deeply in the people analytics space. They were surprised to see some of these biases lurking inside their data. This resulted in re-thinking the recruitment process thereby ensuring that the anchoring bias was addressed”

There are many such biases that may exist in organizations. Some of them are evident while others may not be very easily observable. How can organizations do a better job and proactively address anchoring bias?

I believe people analytics is a plausible solution to counter this problem. Let’s look at 3 examples where it could lead to better and productive solutions:

Interviews as a selection tool: It is fair to assume that managers love to trust their gut feeling and do not necessarily question their own hiring decisions. But gut decisions are not necessarily correct. Research says that a normal interview process has a very low validity (approx. 0.2). This could go up significantly higher if the interview is structured for the role. Structuring an interview needs to be based on insights from the existing database of past interviews-what worked and what didn’t.

“Interview questions with sharper focus on the job requirements can help a) the candidate share relevant experiences and b) the interviewer in evaluating the response”

Internal equity: This is one area which can hurt an organization at a much deeper level, especially companies which are growing at a rapid pace and are inducting lateral recruits to fill positions. It is a well-known fact that it is much easier to get a significant raise at the recruitment stage than compared to yearly salary increments. So, you may very quickly have employees at similar positions and performance levels drawing different salaries. Usually, the employees who have stayed and grown within the system get impacted negatively.

“Good compensation professionals are able to see through this bias and build logical arguments to influence decision-makers, which can have a significant impact on engagement levels for employees”


Forced distribution for performance evaluation: I have found the bell curve to be an efficient technique to distribute performance pay/bonus. It uses relative performance as an anchor and is relatively easy to administer. But I have serious doubts about its ability to manage performance. A good performance management system would ideally help the employee understand how he performed against his key result areas and would provide objective inputs on how to build capabilities for his current and higher roles. A comparison with another individual may not necessarily motivate him to do better. One of the alternatives is to create a system which looks at KSAs (Knowledge Skill and Attitude) for a role. Once you have KSAs for each role, it is easy to set performance standards on a Likert scale. This will enable managers to have productive feedback sessions which don’t necessarily focus on a post-mortem of past performance. They can instead focus on helping team members build their capabilities by bucketing actions under Knowledge, Skill, and Attitude. I have used this methodology in my leadership development journeys across organizations and hence can confidently state that this approach works beautifully and gets great business outcomes.

Have you seen/experienced ‘anchoring bias’? Do you have any interesting examples of biases that may have crept in your organization? How did you solve them?

Read full story

Topics: HR Analytics

Did you find this story helpful?



How do you envision AI transforming your work?

Be Heard: Share Your Feedback and Recommend Our Content!

Selected Score :