I have sat on more interview panels than I can count. For over 25 years, I have watched hiring managers open with the same question: "Tell us about yourself." The candidate talks for two or three minutes. The panel nods politely. Everyone feels good about the conversation. And then the panel picks the candidate who told the best story, not the one who can do the job.
This is not a harmless icebreaker. It is one of the worst questions you can ask in a job interview, and the research on this point is overwhelming. More than a century of data on hiring methods tells us that open ended, unstructured questions like this one fail to predict job performance. Worse, they actively invite bias into the room and give panels a false sense of confidence in their decisions.
If your interview panel still opens with "Tell us about yourself," you are not just wasting time. You are making worse hiring decisions than you would by skipping the interview entirely.
The Science Is Clear: Unstructured Questions Do Not Work
The most important study in hiring history was published in 1998. Schmidt and Hunter analyzed 85 years of research across hundreds of studies covering tens of thousands of employees. They compared 19 different methods organizations use to select employees. Their findings changed everything we thought we knew about interviews.
Structured interviews, where every candidate gets the same job related questions scored against clear criteria, predicted job performance at r = 0.51. That is a strong relationship. Unstructured interviews, the kind where panels ask whatever comes to mind including "Tell us about yourself," predicted performance at just r = 0.38. That is a 34% drop in accuracy.
To put that in plain language: for every 10 good hiring decisions you make with structured interviews, you would make only about 7 with unstructured ones. Over dozens of hires per year, that gap adds up to millions in lost productivity, failed probations, and repeated recruitment costs.
Why This Question Invites Bias Into the Room
When you ask "Tell us about yourself," you hand the conversation to the candidate. That sounds respectful, but it creates a measurement problem. Different candidates interpret the question differently. Some talk about their family. Some talk about their hobbies. Some talk about their career. The panel ends up comparing completely different types of information across candidates, which makes fair evaluation impossible.
The bias research is damning. A meta analysis on interview bias found that unstructured interviews produce bias effect sizes of d = 0.59 compared to d = 0.23 for structured interviews. That means unstructured formats produce more than double the bias. The biases documented in this research include discrimination based on physical attractiveness, pregnancy status, body weight, sex, race, and nonverbal cues like handshake firmness and eye contact.
Think about what "Tell us about yourself" actually does. A woman might mention her children. An older candidate might reference decades of experience, revealing their age. A candidate from another country might discuss their background, triggering accent bias or assumptions about cultural fit. A candidate with a disability might feel pressured to explain gaps in their career. None of this information predicts job performance. All of it activates stereotypes.
Research on gender bias in interviews shows that women receive more questions about marital status and family plans than men. They get interrupted more often. In unstructured formats where the conversation flows freely, these patterns get worse because there are no guardrails to keep the discussion focused on job related criteria.
The most insidious bias at work here is similarity bias. We like people who remind us of ourselves. When a candidate opens up about their personal background, the panel member who shares a similar background rates them higher. The candidate who went to the same university, supports the same football club, or grew up in the same town gets an invisible bonus that has nothing to do with whether they can do the job. Structured questions eliminate this problem because every candidate answers the same job related questions.
The Candidate Who "Interviews Well" Is Not the Best Hire
Here is where the research gets uncomfortable. Most hiring managers believe they can read people. They trust their gut. They think the interview gives them valuable information about a candidate's character, motivation, and fit.
They are wrong.
A landmark study by Dana, Dawes and Peterson tested what happens when interviewers use unstructured interviews to form impressions. In one experiment, they gave interviewers completely random information about candidates. The responses had no connection to the candidates' actual abilities or future performance. Despite this, interviewers formed confident impressions and made predictions they believed in.
The researchers identified two mechanisms that explain why. The first is "sensemaking." When humans receive random, disconnected information, our brains automatically weave it into a coherent story. The interviewer hears a candidate ramble about their childhood, their degree, and their last holiday, and the brain builds a narrative: "This person is well rounded and adaptable." The narrative feels real. It is fiction.
The second mechanism is "dilution." When you mix irrelevant information with relevant information, the irrelevant stuff does not just sit there harmlessly. It actively drowns out the useful signals. A candidate's genuine technical expertise gets lost in the noise of their personal anecdotes. The panel walks out of the room remembering the funny story about the candidate's gap year, not the specific example of how they solved a complex technical problem.
Perhaps the most striking finding: participants in the study preferred having an unstructured interview over having no interview at all, even when the interview information was provably random and useless. We are addicted to the feeling that we are learning something, even when we are not.
More Information, Worse Decisions: The Overconfidence Trap
"But surely more information is better than less?" That is the intuition most managers have. Kausel and colleagues tested this directly and found the opposite.
In their experiments, they gave managers two types of information about candidates: objective test scores and unstructured interview impressions. Managers who received both types of information were more confident in their hiring decisions. But they were less accurate. The unstructured interview data did not add useful signal. It added noise that managers mistook for insight.
The researchers then tested whether this overconfidence had real consequences. They set up a betting exercise where participants could wager real money on their predictions. Managers who had received unstructured interview information bet more aggressively, but won less. They were literally paying for the privilege of making worse decisions.
This finding has a direct parallel to what happens in every boardroom where a panel discusses candidates after interviews. The person who says "I just had a really good feeling about Candidate B" carries more weight than the person pointing to the assessment scores. The feeling wins because it comes with confidence. But the confidence is a mirage created by irrelevant information, much of it gathered through questions like "Tell us about yourself."
The Legal Risk Nobody Talks About
Beyond the accuracy problem, "Tell us about yourself" creates legal exposure. Employment discrimination law in most countries prohibits making hiring decisions based on protected characteristics, including age, sex, race, religion, disability, marital status, and pregnancy.
When you ask an open ended question, candidates often volunteer protected information without being asked. A candidate says, "Well, I just moved here with my wife and three kids." Now the panel knows the candidate's marital status, that they have children, and potentially their sexual orientation. None of this is job related. All of it can influence decisions, especially in an unstructured format where there are no scoring rubrics to anchor judgments to job criteria.
The U.S. Equal Employment Opportunity Commission guidelines are explicit: selection procedures should be job related and consistent with business necessity. An unstructured question that elicits personal information unrelated to the job fails this test. The U.S. Office of Personnel Management specifically recommends structured interviews as the legally defensible approach. If a candidate files a discrimination complaint, a panel that used structured, job related questions with documented scoring criteria has a strong defence. A panel that opened with "Tell us about yourself" does not.
"But We Use It as an Icebreaker"
This is the most common defence I hear. "We don't score it. We just use it to relax the candidate." That argument fails for three reasons.
First, it does not relax everyone equally. Extroverted candidates who enjoy talking about themselves get a confidence boost. Introverted candidates, who may be outstanding performers, start the interview feeling uncomfortable because they have been asked to perform a task they find unnatural. You have introduced a systematic advantage for extroversion, which has a near zero correlation with job performance in most roles.
Second, even if you say you do not score it, the first impression is already formed. Decades of research on the halo effect shows that first impressions colour every subsequent judgment. If the candidate charms the panel with their opening monologue, the panel will unconsciously rate their later answers more favourably. The "icebreaker" is shaping your decision whether you intend it to or not.
Third, there are better ways to put candidates at ease. Start by explaining the interview format. Tell them how many questions you will ask. Offer them water. Give them a moment to settle. These simple courtesies relax candidates without contaminating your assessment data.
What to Ask Instead: Evidence Based Alternatives
If you remove "Tell us about yourself" from your interview, what should you replace it with? The research points clearly to three types of questions that actually predict job performance.
Behavioural questions ask candidates to describe specific past experiences relevant to the job. "Tell me about a time you had to manage a project with a tight deadline. What did you do, and what was the result?" These work because past behaviour is the best predictor of future behaviour, and they force candidates to provide concrete, verifiable examples rather than vague self descriptions.
Situational questions present candidates with a realistic job scenario and ask how they would handle it. "Imagine you discover that a team member has been submitting inaccurate reports. What steps would you take?" These test judgment and problem solving in context, which is far more predictive than asking someone to summarize their life story.
Job knowledge questions test whether candidates have the technical expertise the role requires. "Walk me through how you would conduct a risk assessment for this type of project." These directly measure the competencies you are hiring for.
All three question types should be standardized: every candidate gets the same questions, and every answer is scored against a predetermined rubric. This is the foundation of effective HR management. The science is unambiguous. Structured, job related questions with consistent scoring outperform open ended questions by a wide margin.
Organizations that invest in HR analytics can track the predictive validity of their interview questions over time. This means you can measure which questions actually predict who performs well and who does not. When you track the data, open ended questions like "Tell us about yourself" consistently show no value..
Your Interview Process Is a Measurement Tool. Treat It Like One.
An interview is not a conversation. It is a measurement tool. Its job is to collect data that predicts how well a candidate will perform in the role. Every question you ask either adds useful measurement data or adds noise. "Tell us about yourself," adds nothing meaningful.
The combined evidence from over a century of research points in one direction. Structured interviews predict job performance. Unstructured interviews do not. Unstructured interviews introduce bias. They create legal risk. They make interviewers overconfident. And they produce decisions that are worse than what you would get from test scores alone.
If your organization is serious about hiring the best people, start by eliminating the questions that make your process worse. "Tell us about yourself" should be the first one to go.
Stop asking candidates to tell you about themselves. Start asking them to show you they can do the job. Your HR policies and procedures should reflect what the evidence demands: structured, job related interviews with standardized scoring. Anything less is not just bad practice. It is bad science.
