The Bias You Do Not See
Phone screens feel fair. You ask questions, listen to answers, and judge who moves forward. But the format itself introduces systematic bias that no amount of good intentions fixes. Schmidt & Hunter (1998) showed that structured interviews are 2x more predictive of job performance than unstructured ones, and the phone screen is, by default, unstructured. Same questions get asked in different orders, different phrasings, different moods. The recruiter's Tuesday morning energy is different from their Friday afternoon energy. The candidate who gets the 9 AM slot gets a different experience from the one who gets 4:30 PM after six other calls.
Here are five screening habits that introduce bias, and what to do about each one.
1. Scanning Names Before Reading Qualifications
Identical CVs with different names get different callback rates. A meta-analysis by Quillian et al. (2017), published in the Proceedings of the National Academy of Sciences, found that white applicants received 36% more callbacks than equally qualified Black applicants. The bias is not intentional. It is automatic, fast, and invisible to the person doing the scanning.
The fix: blind the screening step. When candidates record async video answers to structured questions, the first thing you evaluate is their answer, not their name on a CV. Scoring against a rubric anchors the evaluation to the content. You are scoring what was said, not who said it.
2. Judging Voice and Accent Instead of Content
Phone screens put voice front and center. Accent, speech speed, filler words, confidence of tone. These feel like signal but they are not what predicts job performance. Schmidt & Hunter (1998) found that the structured content of an interview drives prediction, not the vocal delivery.
The fix: add a transcript. When every answer is transcribed, reviewers can read the content independently of how it sounded. The transcript strips away accent bias and lets you evaluate what was actually said. You can still watch the video for communication skills when that matters for the role. But for the screening step, the words carry more signal than the tone.
3. Asking Different Questions to Different Candidates
It happens naturally. The conversation flows, a candidate mentions something interesting, and you follow the thread. Twenty minutes later, you asked candidate A about their project management experience and candidate B about their weekend hobbies. Both conversations felt productive. Neither was comparable.
This is the core finding of decades of I/O psychology research. Schmidt & Hunter (1998) documented that structured interviews with consistent questions are 2x more predictive than unstructured conversations. The reason is simple: if every candidate answers different questions, you are not comparing candidates. You are comparing conversations.
The fix: use a template. Write five questions once. Every candidate gets the same prompts in the same order with the same time limits. Async video enforces this automatically. There is no drift because there is no conversation to drift.
4. Letting the First Impression Set the Score
The first 30 seconds of a phone screen set the tone for the rest of it. A confident opener anchors a positive review. A nervous start anchors a negative one. Tversky & Kahneman (1974) showed that initial impressions anchor subsequent judgments even when later evidence contradicts them. The remaining 20 minutes confirm the first impression rather than evaluating the evidence.
The fix: score per question, not per candidate. A rubric with separate scores for each answer breaks the anchoring effect. If a candidate stumbles on question one but delivers a strong answer on question three, the rubric catches it. A gut-feel impression misses it. Each question is evaluated on its own terms, not relative to a first impression shaped by nerves.
5. Reviewing Candidates One at a Time Instead of Side by Side
When you evaluate candidates sequentially, each one is judged against the memory of the last. That introduces recency bias (recent candidates score higher) and contrast effects (an average candidate after a weak one looks strong). Research by Iris Bohnet (Harvard, 2016) found that joint evaluation, comparing candidates side by side against the same criteria, reduces bias compared to evaluating them one at a time.
The fix: batch your reviews. Review all candidates for a role in one sitting, with their answers visible side by side. Compare candidate A's answer to question three against candidate B's answer to question three. Not candidate A on Tuesday morning against candidate B from Thursday afternoon's fading memory.
What This Looks Like in Practice
Structured async video interviewing addresses all five biases by design, not by adding a compliance layer on top. Every candidate gets the same questions (fixes #3). Answers are transcribed (fixes #2). Reviewers see responses side by side (fixes #5). Scoring is per question against a rubric (fixes #4). And the first thing you evaluate is the answer, not the name (fixes #1).
None of this requires more time. interviewstream (2025) reports that video review is 6x faster than phone screening. The EU AI Act, enforceable from August 2026, classifies AI in hiring as high-risk and requires documented, auditable, human-overseen decision-making. A structured, rubric-based process is not just less biased. It is what the regulation expects.
StormInterview is built around this structure. The platform transcribes every answer, supports structured scoring with rubrics, and shows candidates side by side for comparison. AI suggests scores with written reasoning that the reviewer can read, agree with, or override. Every decision is logged with the reviewer's name, so any future audit traces a person, not an algorithm.
Try It on One Role
Pick a role you are currently hiring for. Set up five structured questions in a template. Send the link to your next batch of candidates. Review them side by side with a rubric. One role is enough to see the difference between a structured process and a phone screen that felt fair but was not.
Start a free trial of StormInterview and screen your next batch without a single phone call. You will review the batch in under an hour instead of scheduling twenty separate calls across two weeks.