Back to basics: Percentage agreement measures are adequate, but there are easier ways.

The authors looked at everyday data sheets. They asked: how much disagreement is too much?

They tested interval and trial records. They set a simple cut-off: 10% or less disagreement.

If you hit that mark after 50 observations, they say you can stop checking further.

The 10% rule caught the same bad data as longer formulas. No extra math needed.

Teams saved time. They still met journal standards for reliable data.

Jones et al. (1977) pushed a chance-corrected formula. Friedling et al. (1979) say skip it if you meet the 10% rule. The two papers seem opposite, but they serve different moments. Use the formula when disagreement is high; use the shortcut when it is low.

Goldman et al. (1979) published a table that turns percent agreement into a phi score the same year. C et al. say you can often ignore that step. Again, the clash is only on paper. The table helps when you need a fancy stat; the 10% rule helps when you just need to move on.

Lanovaz et al. (2017) give a phase-length rule to curb false positives. Like C et al., they trade a small bit of rigor for a big gain in speed. Both papers give you permission to stop collecting earlier if simple checks pass.

You can check reliability on the clinic floor without a calculator. Run your session, count disagreements, and if you are at or below 10% after 50 intervals or trials, keep going with your intervention. File the sheet and focus on your client, not on stats software. The rule keeps you honest while giving back precious therapy minutes.