How to Make Interviews More Effective and Avoid Interviewer Bias.

Why is the average interview as reliable as a coin flip? It's psychology!

If you were to ask the average recruiter or hiring manager to rate their interviewing skills, almost all of them would praise themselves as excellent. However, it's well known that most companies struggle to achieve the intended results of the interview process. The purpose of interviewing is to discover quality hires who are both productive and satisfied. There is an apparent disconnect between perceived interviewing skills and actual success in finding quality candidates. As a researcher, I reviewed published data in my field of industrial and organizational psychology, which dates back almost 40 years. In this post, I'll review the best research and offer tips on how to ensure interviews yield the intended outcomes.

Outside of the resume, the interview is the preferred procedure companies use to screen candidates. A large body of research suggests that interview impressions sway the employment decision more than the resume does. This evidence suggests that subjective impressions influence hiring decisions, often more than any other factor. Whereas a resume moves a candidate forward to the interview with what is assumed to be objective information, such as:

• Work history

• Education

• Experience

• Core competencies (Knowledge, skills, and abilities) related to the job description.

The next step, the interview, is much more subjective. Findings on this topic supports that most hiring decisions are not based on objective factors related to what is required for success. Instead, offers are too often extended based on:

• Appearance

• Attractiveness

• Responses to unstructured questions

• Likeability

• Non-verbal behaviors

The common psychological errors of interviewing

Impressions dominate decision-making. This phenomenon is known as the halo effect. We do it all the time when dating someone new. Most decision makers say, “I just have a gut feeling about this person.” However, this statement lacks any quantifiable validity. Additional research supports that many hiring managers fall into what is known as the “four-minute phenomenon.” Decision makers often decide about a candidate within the first four minutes of the interview. Even though the first four minutes of the interview are usually the information-gathering stage (Verifying information such as previous positions and other details on a resume).

A 1996 study found that attractive people are perceived as more sociable, dominant, warm, psychologically stable, intelligent, and socially skilled than unattractive people (Marlowe, Scheider, & Carnot, 1996). Yet the goal of hiring is to identify a high-performing individual over a low-performing one. Whether hiring managers and recruiters care to admit it, the two factors most often used to make hiring decisions are attractiveness and likability. These subjective practices have been researched and known in industrial and organizational psychology as “leniency errors”. Leniency errors produce inflated ratings of candidates without any objective justification.

Even worse, Marlowe, Scheider, and Carnot (1996) found that the less experience a candidate has, the more likely interviewers are to hire based on attractiveness. This suggests that the more interviewers are unsure about qualifications, the more they rely on attractiveness and likability. This is concerning considering that when a candidate lacks job history or experience, there are proven methods for assessing their potential or likelihood of success in the role. Outside structured interview questions, intelligence (GMA) and predictive personality assessments can easily measure the potential for success in most cases and for most positions, but are rarely weighted heavily in the selection process.

Several other well-documented psychological errors often hinder a decision maker's ability to make quality hiring decisions during interviews. Going back to the dating analogy, it’s clear that a brief sampling of behavior in an interview is not representative enough of someone’s stable traits to make a hiring recommendation (or to commit to a relationship), but it happens, nonetheless. A common psychological trap that decision makers fall victim to is mental heuristics. A heuristic can be defined as a rule of thumb used to make decisions that are easy and save time. Heuristics are errors in decision-making because they are considered a form of mental laziness. People like to make mental shortcuts.

Here are a few common examples of mental heuristics in the hiring process. Many recruiters believe that “95% of quality candidates are already working" suggesting that unemployed people have no talent to offer. Other mental heuristics include dogmatic practices, such as "If someone’s LinkedIn profile does not match up 100% with a resume, they are disqualified.” While it can be understood that many companies today must sift through an average of 300 resumes for one position, making it necessary to be efficient in eliminating candidates, heuristics in thinking exclude candidates without any objective justification.

The most dangerous heuristic is known as the fundamental attribution error. People, regardless of their social situation, tend to give themselves too much credit for the good things that happen. So, when a candidate performs well, it's all too common for hiring managers to use self-serving attributions. Yet when candidates fail and fall victim to turnover, they tend to underestimate the situational factors that may have contributed to it. Managers tend to attribute all the positive outcomes that occur within the organization to their own choices and personal traits. Therefore, they fundamentally attribute the bad things to everyone they manage, and all the good things that happen to themselves. This creates a horrible cycle of non-accountability.

What the data shows

Most companies have a success rate of predicting retention with new hires that is less than a simple coin flip. Let's review some correlational evidence. First, we all know that correlation does not mean causation. However, the higher the correlation, the more likely causation exists. It is generally accepted that if you flip a coin 100 times, 50% will be heads, and 50% will be tails. This would mean the correlation between coin flips and landing on heads is 0.50. A correlation of zero would suggest no causation, and a correlation of 1.00 would indicate a perfect correlation. Suggesting that there may be no other causation for the measured variables (no need to continue investigating other variables of causation).

In a study conducted with nurse applicants (Kinicki 1990), three Nursing Directors were assessed regarding their hiring process of 186 Nursing applicants. Like other studies, the authors found that when compared to the resume, impressions had the most significant impact on the hiring recommendations. The variance that accounted for resumes as a factor in future job performance was only .09. To make matters worse, “comparisons in the change of the squared multiple correlation show that every manager primarily made hiring decisions based on impressions interviewees evoked in face-to-face interviews.” While manager “B” did better than his/ her peers, the resume (objective criterion) was not found to be a factor in the ultimate hiring decision.

The authors also followed the new hires and assessed the variables of interview impressions and relevant job qualifications to new hire performance, retention, organizational commitment, and job satisfaction. As you can see, because interview impressions dominated the hiring process, the validity of the new hires variables listed below measured either no correlation or a negative correlation.

While this is only one of many studies, research supports hiring based on resume reviews and unstructured interviews alone yields a correlation of .14-.33 (Campion & Palmer, 1997). This suggests that the average interviewer would have a better chance of flipping a coin, or simply hiring random people off the street! When considering the costs of training, lost production, and all other associated revenue losses resulting from a poor hire, it makes sense to avoid unnecessary turnover. On average, most companies have a correlation coefficient between 0.18 and 0.20. This includes companies that outsource the hiring process to external recruiters, who typically take 20% of a new hire's salary.

How to fix it.

Interviewing needs to be a process of mathematical measurement. This is done by creating a very objective, structured process, and translating interviewee answers into scores. Let's break this down. Using the proper structured interview techniques has increased the correlation between interviewing and quality hires by .30-.40 (Campion and Palmer, 1997). It's essential to measure (score) interview questions and avoid getting sidetracked with unstructured conversations. Here is a condensed list I have compiled to help make the interview process more effective:

• Conduct (or reassess) a proper job analysis of the position. Do all decision-makers agree on the competencies outlined in the job description? If not, it's impossible to measure them accurately in an interview. Issues in the selection process begin with the inability to fully define the knowledge, skills, abilities, and other characteristics required for role success. Additionally, subject matter experts within the organization often disagree about these characteristics and what the behaviors that lead to success look like. This creates “bad hires.” Tough conversations frequently need to be had, and the subject matter experts need to be in complete agreement before a job description being published and advertised.

• Consider a behaviorally-based job description. This should match the desired workplace behaviors that can then be measured in a quality psychometric. Before interviewing, a job profiler should be completed that converts desired behaviors into cut scores (statistics) that serve as a hurdle system. In other words, if candidates can't clear the hurdle, they have been objectively eliminated. For example, they must have a T score above 68 (second standard deviation off the mean) in the domain of agreeableness on the big five assessment. This can be accomplished through a one-time process called “competency modeling”. Learn more here.

During the interview, do the following

• Structure your interview questions to align with your job description and ask the same questions in every interview, with every candidate. Create structured questions that are behaviorally-based. An example of this would be “How would you help a dissatisfied customer?” (Choose an example relevant to the business). Have desired responses listed out on your evaluation form, such as “Our desired candidate would respond like this”:

  • An immediate response that acknowledges the customer's concern

  • Resolution is a top priority

  • Doesn't hear the problem as a personal attack

  • Finds a convenient solution that the customer finds non-disruptive to their lives.

• Have multiple interviewers. Compare your scores and discuss. Keep a record of every candidate interview. Digital records work best. Rate or score every answer on a Likert scale of 1-5. Let's use an example of a salesperson and asking structured questions related to skills.

• Be aware of any “contamination”. This means be aware of avoiding questions that can lead to faking, socially desirable responses, or irrelevant information from a candidate. Control ancillary information. Don’t allow the conversation to take another direction.

• Verify all possible competencies such as soft skills (negotiation, persuasion, etc.), through psychometric testing after the interview and only with the candidates you choose to move forward with. This allows the firm to save money instead of administering assessments to every qualified applicant who submits a resume.

• Make a gerneralized final determination on a likert scale at the end of the interview. You would then score a simple 1-5, or something similar to this:

  · 5 - Excellent - Always matched the points to look for.

  · 3 - Average - Overlapped with at least half of the points to look for.

  · 1 - Poor - Matched less than half of the points to look for.

  · 0 - Unacceptable - Did not match any of the points to look for.

• Over time, using the above techniques, keep track of your human capital expenses with an employee turnover calculator. According to Hartwell and Campion (2016), knowing “that the risks of hiring a bad employee outweigh the risk of not hiring a good employee,” human resources and management continue to incorporate willful blindness by ignoring the financial impact of a bad hire.

The above reccomendations can dramatically increase the liklehood of hiring productive and satified employees. A proper, structured interview process can increase the correlation of interviews and quality hires by about .30 (above chance / a coin flip). Then adding in the proper psychometric, it is possible to have a correlation associated with your interview process rise to 0.60 - 0.70 over traditional, outdated, subjective hiring procedures. To ensure success, all the puzzle pieces must be in place to hire quality candidates. Few companies make the required efforts, but the solutions only need to be implemented once. To achieve the results you want from interviewing, you must do the things that your competitors refuse to, and be aware of the psychological traps!

References:

Marlowe, C., Scheider, S., & Carnot, E. (1996). Gender and Attractiveness Biases in Hiring Decisions: Are More Experienced Managers Less Biased? Journal of Applied Psychology. Vol 81,1.

Levashina, J. & Campion, M. (2007). Measuring Faking in the Employment Interview: Development and Validation of an Interview Faking Behavior Scale. Journal of Applied Psychology. Vol 92, 6.

Campion, M. & Palmer, D. (1997). A Review of Structure in the Selection Interview. Personnel Psychology. Vol 50.

McDaniel, M., Whetzel, D., Schmidt, F. & Maurer, S. (1994) The Validity of Employment Interviews: A Comprehensive Review and Meta-Analysis. Journal of Applied Psychology Vol 79, 4.

Hartwell, C. & Campion, M. (2016) Getting on the Same Page: The Effect of Normative Feedback Interventions on Structured Interview Ratings. Journal of Applied Psychology. Vol 101, 6.