It's good that it's studied, but it does sound that they've "conclusively" proven something, and I'm not so sure. Small sample size too, like you said.
I have the same impression. Possibly it could be made less subjective by literally repping to failure, but I'd guess that has potential ethical implications.
Reps to failure can be done pretty darn safely with machines and incredibly low risk of injury if done on the higher rep ranges (say 20 per set). You’re not going to get rid or risk of injury but in order to even do this study, my opinion is you have to be at a normal level of injury risk or the study is pretty much flawed by design.
The wording is a little technical. They're saying that, if the subject thought they could do at least 2 more reps at the end of this set, they increased the weight on the next set. So if creatine does allow people to do more reps at a given weight, subjects on creatine should have lifted heavier weights. (To me, that doesn't seem like the right way to adjust for creatine, which is thought to provide extra energy, not the ability to lift heavier.)
Still, this cohort might suffer from the issue of lack of awareness about _how exertive_ the actual workout is. I feel this is also something you should learn.
https://www.mdpi.com/2072-6643/17/6/10812.3.4. Resistance Training Program
After the 7-day wash-in, both groups followed the same RT program that comprising 3 full-body sessions a week for 12 weeks (Supplementary Materials, Table S1). All sessions were supervised by tertiary qualified exercise physiologists and commenced with a standardised warm-up of dynamic flexibility exercises. Each session consisted of 5 exercises: 2 compound movements each for the upper and lower body, and 1 isolation movement for the upper body. Four sets were prescribed for all movements to ensure an adequate weekly training volume for hypertrophy [32]. Training intensities were 6 to 12 repetition maximums (RM) with 60 and 120 s of rest between sets and exercises, respectively. To adhere to the prescribed RM, an individual’s rating of perceived exertion (RPE) on a Likert scale of 1–10 was recorded. RPE corresponds to the number of repetitions an individual perceives they will be able to perform after the set is complete, where an RPE 5 equates to 5 reps more, RPE 6 is 4 reps more, RPE 7 is 3 reps more, and so on [33]. When a RPE of 8 or lower was recorded, the external load (kg) was adjusted on successive sets to ensure that subjects achieved the target RM. The RM method was used to ensure that training intensities were relative to the individual’s abilities while also standardising the training intensity across all participants [34].
I'm not quite sure this is clear enough for me, though it does somehow suggest that they were pushing the participants to do as much as they could. But like I say, unclear.
Small N as well - only a few dozen people.