Gap study vs. volume count: what each one actually measures

A planner asks for "traffic counts" at an unsignalized intersection, and someone on the team hands back a volume count: vehicles per approach, 15-minute bins, turning movements broken out. Then the warrant analysis stalls because what the study needed was a gap study, and those are not the same deliverable. They don't even use the same field equipment.
What a volume count measures
A volume count answers "how many." You park an observer, a tube counter, or a camera at a location, and you tally vehicles by approach and movement over a time window. The output is a number per interval: 42 northbound through movements between 7:00 and 7:15, say. Volume counts feed capacity analysis, signal warrant studies, level-of-service calculations, and design decisions like lane configuration. They tell you how much traffic exists, but nothing about how that traffic is spaced in time.
What a gap study measures
A gap study answers a different question: "how much usable time is there between vehicles on the major street, and can a driver on the minor approach get through it." Instead of counting vehicles, the observer times the intervals, the gaps, between successive vehicles on the conflicting stream. Each gap gets logged along with whether a waiting driver accepted it or let it pass. That accepted/rejected pattern is the raw data for everything downstream.
This is the core of unsignalized intersection analysis under HCM Chapter 20 for two-way stop control. A minor-street or left-turn movement's capacity depends on how often gaps long enough to cross or merge show up in the major-street stream, and that depends on gap size as much as raw traffic volume.
Critical gap, explained plainly
The critical gap is the gap duration at which a driver is equally likely to accept or reject it, roughly the median of the accepted-gap distribution once you account for drivers who reject gaps they could have taken. HCM publishes default critical gap and follow-up time values by movement type, number of major-street lanes, and whether the approach is single- or multi-lane, and most agencies start from those defaults rather than field-calibrating every site. Follow-up time is the second half of the equation: how closely spaced a queue of minor-street vehicles can use a single long gap, one after another.
Plug critical gap and follow-up time into the HCM gap acceptance formula alongside the conflicting major-street volume, and you get potential capacity for the minor movement, then movement capacity once you adjust for shared lanes and upstream blocking. None of that math runs on a vehicle count by itself. It needs the gap distribution.
Why the two get confused
Both studies happen at the same intersection, sometimes on the same morning, and both involve someone standing on a corner with a clipboard or a camera pointed at traffic. The field setup looks alike enough that a scope-of-work line item saying "traffic count" gets interpreted as whichever one the crew is used to running. But a volume count and a gap study are different data collection methods producing different inputs: one is an aggregate tally, the other is a timed sequence of individual gaps with accept/reject outcomes attached. You can't derive one from the other after the fact.
Where this leaves an uninstrumented link
Gap studies need the sequential timing data, which means a human or a camera watching continuously through the study period. There's no shortcut around that part. But the volume side of a capacity or signal-timing study, the turning-movement counts and approach totals you need alongside the gap analysis, doesn't require that same continuous field presence. A single scheduled pass over the corridor can produce a count by link and direction without laying tubes or closing a lane, which is useful on the links nobody has instrumented and nobody wants to send a crew to.
If the volume count is the piece you're missing for an uninstrumented link, that's the gap Vehicle Counting is built to fill.