I get asked about the grade inflation that I’ve seen over the years. I know that there is anecdotal evidence of grade inflation from various sources. For example, the GCE A Level exams (based in the UK) had to introduce a new top grade (A*) because so many people were getting the previous top grade (A) that it was becoming somewhat meaningless. Likewise, as I mentioned in another post there are reports that over half of U.S. high school graduates have “A” averages. However from my side I can’t make any conclusions based on our admissions data. Here is why…
Over the past decade our admissions averages have been creeping upward. That’s not an indication of grade inflation however. We keep getting more applicants for every available space in our programs. We also tend to take applicants with the higher averages (although it’s not the only factor), and so the admission averages go upwards simply because of competition. This shows that there is more competition, but does not prove anything about grade inflation.
Also, our applicants are self-selecting and not a random sample. Presumably, applicants with low grades do not bother to apply to Waterloo Engineering, and so we never see those grades. Therefore we can’t draw any conclusions about grade inflation because the data we get is incomplete and not randomized. Engineers and scientists need statistically valid data to draw good conclusions, and this data is not satisfactory for answering the grade inflation question.
Some universities show class averages on their transcripts (Waterloo doesn’t). From there, I can see that many of those university courses have class averages in the range of 65 to 75% (C to B), which is typical for our first year courses too. This suggests that there isn’t much grade inflation at those universities in the STEM courses, but it’s just a small sample.
In Ontario, it would be interesting and useful for schools or school boards to publish high school course averages, especially for the key courses like the math, science and English courses (MHF4U, MCV4U, SCH4U, SPH4U, ENG4U). Then it would be clear for all to see over time if there are significant problems with grade inflation. Probably won’t happen though.
A Professor in Waterloo Engineering 
That would be an interesting data set! If they supply a course average, and not individual students(no names just numbers), would it still be possible to tell if there are problems? Since if one schools course average was higher than an other schools, it could be that they have a better quality of teaching styles. Since there can be a quality increase from school to school, there can be a quality increase over time from improved teaching experience. In order to combat this, an individual, or group, would have to monitor the teachers quality of teaching based on a per-determined scoring system.
Something else to consider may also be that the students are actually able to retain knowledge more or less efficiently. This can be from many factors such as modern technology.
All I’m suggesting are possible causes for grade inflation. Also how to combat the problems that can be faced due to high amounts of measurement bias. I do agree it is a concern, however, it would be very hard to measure properly, as you have also stated.
Lots of interesting things can be analyzed and assessed if good data is available. In the absence of data we can only speculate, which is a bit futile.
I have had various teachers have attest to grade inflation over the course of their time in the education industry (from students to teachers). Obviously the sample size is low and somewhat self-selecting (teachers who don’t observe any change are less likely to casually mention their observations than teachers who did), and as you said above, speculating on incomplete data is a fools errand.
What I find interesting though is how teachers might respond to perceived grade inflation. If individual teachers, convinced by their own experiences, think that a high grade is worth less – it may seem rational to combat the problem by adjusting the difficulty of their own class.
Mind you, I have effectively no evidence that teachers are actually doing this on any meaningful scale, and as such am not convinced of it as an issue. It’s just an interesting thought experiment.