It's important to view studies from the standpoint of "what did they actually examine?" and "what were they capable of measuring?" rather than immediately applying as a generalization. Which I'm not saying you did, of course.
Here's the first:
So, a group of 70-85 year olds received a total of 40 g whey protein per day, which isn't very much, and further was not an addition to an existing diet but a substitute for an isocaloric amount of composition not specified in the abstract, and underwent resistance training.
The measured average increase in muscle cross sectional area was somewhat greater for the WPC group: a person training with weights would probably rather experience a 4.6% increase than a 2.9% increase; for example, wouldn't we rather gain 16 lb over a course of time rather than 10 lb simply from adding one serving of whey per day? This is the same percentage difference, not that I'm saying that one can expect that.
But, the study was not statistically capable of resolving differences that "small," so therefore it is reported as being no result.
It does show something, not a great deal, but something, about what happens when you give 40 g/day whey protein to the elderly undergoing resistance training. It shows that they don't Hulk out, for example. The general applicability of the data is zero, when considering carefully what was looked at and what they were able to measure.
So, previously untrained young men were studied over a very short period of time. This group can gain muscle very rapidly almost regardless of nutritional status, with very high variability between individuals.
This study is sort of like trying to determine whether a particular VP Racing gasoline, say one designed for turbocharged engines, can produce more power than pump gas by enlisting a bunch of novice drivers, putting them in a widely varying fleet of rental cars, and seeing who can drive across New York City in rush hour the fastest. The gasoline is not what's going to make the difference.
Anyway, in terms of the results, the variability -- the plus and minus figures -- was so great that there could have been substantial differences in this group in terms of whether each individual would have done substantially better on one protocol versus the other, but not statistically detectable in terms of average. With high variability, it could be that one group happened to have more naturally-higher-gainers assigned to it than the other, with no way of knowing whether that happened but only a means of calculating the probability that it might have happened, and therefore nothing could be detected.
It always, in my opinion, would be good if they put into at least the article or preferably even the abstract how small an effect they would have been able to detect. This would save a lot of time for the reader. If seeing right away that even, for example, a 5 lb difference in muscle gain would have been "statistically insignificant" in a given study, then it could be ignored right away, instead of having to grind numbers oneself to find that out, or having to estimate by eye from the plus-or-minus numbers. But authors don't like to put out there directly just how little detection ability their study had. Almost never is this directly admitted.
I'd have to purchase the article to find the actual results and variabilities, which are not given in the abstract, but the situation is likely the same as the above study.
In general, it's an entirely different question whether an individual who has trained consistently and well for years and has reached a point of tending to stabilize at a given amount of LBM with further gains being difficult and slow, versus individuals as in the above studies. The former situation is easily personally evaluated, but difficult to do scientifically, and rarely if ever done. While it would be nice to find a study that even tries to look at that, I haven't seen where it's even been tried.
Attempted extrapolations to serous weight training from the elderly, or the untrained or typical (which is a low standard with regards to quality and consistency of training) college student getting one shake a day and so forth, will not necessarily be reliable.
Further, with regard to where many have discovered that one can personally observe differences, it's not from, for example, taking 40 g whey protein in a shake instead of having it in a meal. It's principally from protein shakes allowing substantially greater protein intake where the intake wasn't particularly high in the first place and/or substantially greater caloric intake than what the individual in fact does without the shakes. It's by no means a magic bullet where a relatively small amount does big things in a short time. Any study that could only detect such a thing will turn up nothing "to statistical significance."