Not All Interview Signals Are Equal
Treating all interview feedback as equally valid is the most common structural error in hiring. A 30-minute peer coffee chat produces different quality signal than a 90-minute structured behavioral interview conducted by someone who has managed the same role.
Signal weighting gives each piece of evidence an effective contribution to the final decision that reflects its reliability — not just its recency or the seniority of the person who provided it.
The Confidence Dimension
Every score includes two components: the score value, and the confidence of the scorer. An interviewer who observed a candidate describe three distinct technical architectures in detail holds high-confidence evidence. An interviewer who asked one theoretical question and received a plausible-sounding answer holds low-confidence evidence.
Explicitly capturing confidence alongside scores — even with a simple high/medium/low tag — changes how you aggregate. A confident 2/4 should outweigh an uncertain 4/4 when determining whether a candidate has a dimension-level gap.
Signal Source Hierarchy
Different interview formats produce different quality signals for different attributes. Work sample tests outperform behavioral interviews for assessing technical skills. Behavioral interviews outperform hypothetical questions for predicting execution behavior. References provide signal on patterns over time that no single interview captures.
Knowing your signal hierarchy prevents over-indexing on easy-to-generate signals. 'They were impressive in the culture-fit conversation' is a low-fidelity signal for predicting technical execution.
Strong vs. Weak Signal
Strong signal is specific, behavioral, and independently observed. A candidate who built a specific system and explains the tradeoffs made is producing strong signal on technical depth. An interviewer who says 'they seemed technically sharp' is reporting an impression, not a signal.
Weak signal should inform only when strong signal is absent — and should be labeled as such. Weak signals that drive decisions are the primary structural cause of fragile hires.
Aggregating Conflicting Signals
When interviewers provide conflicting signals on the same dimension — one strong positive, one strong negative — the correct response is not to average them away. Opposite-direction high-confidence signals warrant investigation before a decision is made.
Unresolved dimension conflicts are a structural hold trigger, not a calibration rounding problem. Surface them in the debrief and require each interviewer to present behavioral evidence before a resolved score is recorded.