Is Your Investigator More Biased Than You Think? Part II: How to Keep Truthiness Out of Your Workplace Investigations

All companies should have mechanisms through which employees can voice concerns and feel confident that (a) those concerns will be taken seriously, (b) those concerns will be fairly and promptly investigated, and (c) misconduct, if it is found, will be corrected. However, if unconscious bias infects decision-making when it comes to workplace investigations, employees will feel the system is rigged against them—no matter how well-written a company policy is, no matter how many avenues are given to employees to lodge complaints, and no matter how much a company talks about a culture of accountability. Fortunately, there are a number of steps workplace investigators can take to minimize the chances that this will occur.

Generally speaking, unconscious bias can be most effectively rooted out when the person in the position to make decisions recognizes and can identify possible biases and takes steps to correct what might be negatively affecting the decision-making process. For workplace investigations, this process might involve taking a number of corrective measures at each stage of the investigation.

I. The Complaint Phase

At the complaint phase (whether an employee reports a concern verbally, in writing, or through a complaint-filing system), investigators might take the following steps to combat unconscious bias:

Don’t Prejudge

Take a deep breath and make sure not to prejudge the validity of the complaint. Make sure you make a fair decision about whether the concerns raised need to be formally investigated. Even if your first instinct ends up being correct (that the allegations were or were not substantiated), keep your middle school math teacher’s words in mind: you might not get credit for the right answer if you don’t show your work (i.e., be prepared to justify the conclusions of your investigation with the information you’ve gathered).

Be Conscious of Communicating a Bias

Beware of making comments—in writing or to others about the validity of the allegations even before you start your work. In a number of investigations I’ve reviewed, the person assigned to look into concerns of inappropriate behavior began his or her work by writing notes on the written complaint, including pre-judgments such as “this is a lie,” “that is only her perspective,” and “this didn’t happen,” among others.

II. The Pre-Investigation Phase

Prior to starting the investigation, the investigation may take the following steps:

Test Yourself

A good measure on the issue of sexual harassment is to think back on your immediate reaction when news broke about recent high-profile sexual harassment scandals and sexual assault allegations brought against celebrities. If your first reaction was that the women accusing these men of improprieties were lying, waited too long for the allegations to be believable, or were making the allegations for self-serving reasons, such as financial profit and fame, then you might have a predisposition to disbelieve people who complain about this type of behavior or to believe people accused of engaging in this behavior. If, by contrast, your first reaction was to say that you knew the allegations were true—perhaps because you did not like these men for reasons unrelated to the allegations of misconduct or because you believe these types of allegations are never made up—then you might have an unconscious preference to generally believe accusers.  The good news is that the mere existence of these preferences does not mean you cannot be an effective and fair investigator—it simply means you need to recognize that you have this bias and take steps to prevent that bias from influencing the way you conduct your work or the way you make decisions about an investigation (see below for strategies for being an unbiased decision-maker).

Prepare, Prepare, Prepare

Investigators should take a page from programs that have effectively diminished the effects of unconscious bias from their own recruiting, interviewing, and hiring processes by being meticulous about the work conducted up front—at the selection phase. In recruitment and hiring, companies committed to diversity have implemented the following mechanisms to ensure that employee selection is unbiased:

• reviewing job postings to make sure they do not discourage certain groups from applying,
• using automated methods to select candidates for interviews in which a program selects candidates based on predetermined criteria such as education, experience in a particular industry, or other objective job-related factors, and
• ensuring that hiring panels are diverse and that they ask candidates identical questions.

Analogous strategies can be used for workplace investigations. This might include:

• preparing an investigation strategy that objectively and precisely defines the scope of the investigation,
• having a methodology for how to conduct interviews (including consistent strategies for opening the interview and making the witness comfortable), and
• preparing a detailed interview outline that uses open-ended questions related to the concerns raised.

III. The Information-Gathering Phase

During the information-gathering phase, investigators can take the following steps:

Review and Analyze ALL Evidence

In order to prevent confirmation bias, interview witnesses and review documents that corroborate the allegations and evidence that contradicts the allegations (particularly if you recognize that you might be forming impressions or arriving at conclusions before you finish gathering evidence). 

Create a Setting Conducive to Sharing Information

Take into account common-sense realities as you interview witnesses and evaluate their statements. People are often nervous, scared, or defensive during interviews. Make sure you ask questions that are precise and nonjudgmental, make it clear that you have not reached conclusions about the concerns, make sure you use techniques to make all witnesses feel comfortable and that they trust that you are simply trying to uncover facts (that you have no hidden agenda).

IV. The Analysis and Conclusion Phases

During the analysis and conclusion phase, investigations can take the following steps:

Collect and Consider all the Facts

An adage I often use for workplace investigations is that there are no “good” or “bad” facts in investigations, there are only facts. HR employees (and in-house counsel) sometimes see their role as someone who must defend or protect the company and its assets. However, these are not the roles played by an unbiased investigator. In fact, skewing the fact-finding in one direction or analyzing the evidence in a biased way is the opposite of what an independent investigator must do. Moreover, skewing an investigation in the company’s favor ultimately is worse for the company since the facts will eventually be revealed and the chances of the reporting party bringing a claim because of his or her dissatisfaction with the investigation process and conclusion increases exponentially if the investigator is a biased advocate rather than an independent fact-finder.

The learning lesson that decreases the chance of bias creeping into the analysis is to make sure that the verbal or written report the investigator prepares  at the conclusion of the investigation includes all the evidence—both the facts that support and those that might contradict the allegations. By doing so, an investigator can much more fairly analyze and judge the validity of the claims brought.

Rely on Tools

Investigators have many tools at their disposal that might help them see the evidence more objectively. Here are some examples:

• Create a separate timeline that simply shows dates and events (with no commentary about the meaning of those events). This timeline may serve as a way to view the evidence in a much more detached and fact-based way.
• Seek the help of experts. Your unconscious bias might be leading you to a conclusion that is based on your interpretation of the data, but perhaps someone with more experience in the substantive area can guide you to make a more judgment-neutral assessment. For example, I’ve worked on numerous investigations involving allegations of wrongdoing by leading scientists, including allegations of scientific fraud. During these investigations, I’ve often enlisted appropriate experts to help me interpret the scientific information relevant to the investigation so that I could ultimately arrive at a balanced and informed (and thus unbiased) conclusion.

Play Devil’s Advocate

If you are worried that you are viewing and interpreting the information presented in a biased way, test your conclusion. Ask yourself how an “advocate” for the party you are finding against would argue. Perhaps the party’s spouse, parent, or attorney would be able to present the evidence in a way that differs from your interpretation of the information you gathered.

Similarly, if you have a colleague with whom you can share this information—and if it would not breach any duty of confidentiality to do so—test your conclusion with that person to make sure your own biases aren’t affecting your decision-making.

V. Deciding Appropriate Remedial Measures 

When deciding which remedial measure to implement, investigations might want to take steps to ensure that bias is not affecting the investigation. Assuming there is some level of misconduct found, do not let assumptions or preferences dictate the remedial measure you will implement. Depending on the severity of the misconduct, some factors might be appropriate to use in the deciding which remedial measures to implement. For example, it may be appropriate to consider the fact that the misconduct was committed by a long-term employee with no prior record of misconduct. Other factors, however, are not appropriate and might be considered evidence of unconscious bias. A common example involves imposing a lenient remedial measure, or no remedial measure at all, because the misbehaving employee is a highly-valued employee or is an executive of the company.

As an initial matter, investigators should keep in mind that an appropriate remedial measure is on that is (1) commensurate with the level of wrongdoing and (2) tailored to prevent recurrence. One tool I use in deciding on the appropriate remedial measure is to create a list of possible remedial actions on a spectrum of 1 to 10.  A 1 might be a meeting to provide verbal coaching; a 10 is immediate termination. There are hundreds if not thousands of possibilities in between. When you reach the end of your investigation, think about the level of seriousness of the misconduct. If you use the same kind of scale to judge the misconduct to be a 3, do not implement a fix that is a 7 on the remedial measure spectrum. Conversely, do not implement a 3 fix (or no fix at all) when the level of misconduct was an 8.

Additional Tips

Track Conclusions and Remedial Measures

One of the most effective ways to measure possible unconscious bias is to keep track of your investigation decisions. Ideally, your company has a database or software that is sophisticated enough to capture a number of conclusions, including:

• The allegations were fully-substantiated.
• The allegations were partially-substantiated.
• The investigator failed to find a violation of the letter of the policy but did find a violation of the spirit of the policy.
• The allegations were unsubstantiated.

The database of conclusions should also include any learning lessons from the particular investigation. Similarly, keep track of remedial measures you have implemented at the conclusion of each investigation.

Regularly Analyze the Data

One of the benefits of tracking conclusions of investigations is it gives companies the ability to analyze their data to uncover evidence of unconscious bias in investigations. One client who supervises dozens of internal workplace investigators at her company told me she suspected that one particular investigator had a propensity to find against employees who made complaints. To test her hunch, she researched how often, over the last few years, this investigator “found” that the allegations were unsubstantiated. The data showed that the investigator ruled against the complaining employees in the vast majority of instances.

Although there is no percentage that is considered fair or bias-free, if your research shows that your conclusions are skewed one way or another, this is a strong indication that you have an unconscious bias to find in favor of or against certain parties. You may want to conduct a similar analysis of your findings as they relate to more nuanced characteristics:

• Do you find in favor of accused high-level managers more often than their lower-level counterparts?
• Do you always find for or against parties that allege certain types of claims (bias based on race or gender, for example)?
• Does your review indicate that you tend to find allegations made by “high-value” employees to be without merit?

Similar to researching and looking for possible patterns in your conclusions, look at the record of decisions for remedial measures:

• Do you tend to implement harsher measures against members of a particular group?
• Are you more lenient in your implementation of discipline to employees in a certain department or region?
• Do you tend to favor employees with whom you have something in common?

Share Information Across Departments

Finally, coordinate with your colleagues in other departments across your organization, including human resources, diversity, legal, and others. This will allow you to track decisions department-wide or company-wide by using the same criteria cited above: Are findings balanced based on employee characteristics, department, rank, etc.? Are remedial measures evenly handed out irrespective of these characteristics?

Conclusion

In 2016, comedian/TV personality Stephen Colbert coined a new meaning of the term “truthiness,” which made it into the Oxford English Dictionary to mean “the quality of seeming or being felt to be true, even if not necessarily true.” Independent fact-finders should strive to keep “truthiness” out of their vocabularies and instead be painstakingly vigilant about keeping unconscious bias out of their workplace investigations.

Part one of this two-part blog series covered what unconscious bias is and how it can disrupt workplace investigations.