A common metric used in research studies is "percentage of stuttered syllables," or %SS. To calculate that, you record yourself (or someone else) talking, make a note each time you suspect you hear a dysfluency, then divide that number by the total number of syllables, so: `number of stuttered syllables / total number of syllables spoken` Generally speaking, I'd say 2-5 %SS is mild dysfluency, 5-10% is moderate, and anything over 10 %SS is severe dysfluency. Interestingly, SLP's tend to score fluent speakers around .6 %SS; they often aren't told who does/doesn't have a stutter when they're listening to voice samples in a study. %SS is great for studies because it's a way to quantify fluency, but I wouldn't say it's everything, though. These studies often score "speech naturalness" too, and after intensive speech therapy, participants will usually go from 3-10 %SS down to 2 %SS, but their speech naturalness will actually be rated *worse* than when they were stuttering more. Plus there's the speaker's own subjective experience as they're speaking.