Severity Score vs. Evidence Score: Why Merging Them Is a Mistake
A lot of intake scoring systems - not just in this category - collapse everything into one number: a priority score, a risk score, a triage score. It's an understandable design choice. One number is easy to sort by, easy to explain in a demo, easy to build a dashboard around. It's also the wrong choice for workplace misconduct intake, because severity and evidence strength are answering two genuinely different questions, and merging them hides information the person handling the case actually needs.
The two questions, kept separate
Severity asks: how serious is what's being described, if true? A single yelling incident and a pattern of sustained retaliation over months are not the same severity, independent of how well-documented either one is.
Evidence asks: how substantiated is this account right now? Named witnesses, dates, documents, prior complaints on record - all evidence signal, independent of how serious the underlying allegation is.
These two axes move independently, and the four combinations they produce require genuinely different handling.
Why the combinations matter
A high-severity, high-evidence case - serious allegation, strong documentation - is usually the clearest to route: it needs a skilled investigator, quickly, and the evidence base means the investigation has somewhere real to start.
A high-severity, low-evidence case is the one that gets mishandled most often by systems that merge scores into one number, because a merged score might rank it lower than it deserves simply because the evidence isn't there yet - when the right response is usually the opposite: prioritize it precisely because the potential severity is high, and use early follow-up questions to build out the evidence before the trail goes cold.
A low-severity, high-evidence case is often exactly what it looks like - a minor, well-documented issue that can move through a lighter-weight process without an investigator spending disproportionate time on it.
A low-severity, low-evidence case is genuinely the hardest to route confidently, and honestly, some of these are just vague or preliminary reports that need a clarifying follow-up before either score means much - which is a legitimate outcome, not a system failure.
What a merged score would actually hide
Picture a single "priority score" that happens to land two very different cases at the same number: one is a serious allegation still light on evidence, the other is a moderate issue that's extremely well documented. A merged score treats them as equivalent priorities. An investigator working from that single number has no way to tell, without opening both case files, which one needs an urgent follow-up question and which one is ready to move straight to interviews. Keeping the scores separate means that distinction is visible before anyone has to dig.
The tradeoff of keeping them separate
It's genuinely a little harder to build a simple sorted queue view around two numbers instead of one - "sort by priority" is a cleaner UI pattern than "sort by severity, then check evidence gaps." That's a real cost, and it's worth naming rather than pretending the two-score design is free. The reason it's still the right call: the alternative optimizes for a clean dashboard at the expense of the actual judgment call an investigator has to make. Software that makes triage look simpler than it is doesn't make the underlying decision easier - it just moves the complexity somewhere less visible, usually onto the investigator who has to reconstruct the missing distinction by opening every case file anyway.