Due to the fastidious efforts of Professor Harvey Mansfield, Harvard has confirmed the legend that "the hard part is to get in". Not only does it appear impossible to flunk out but according to the new revelation (link), the median grade given is A- and "the most frequently awarded grade at Harvard College is actually a straight A".
The last sentence can be interpreted in two ways. If "straight A" means As across the board, then he is saying a lot of graduates end up with As in all courses taken. If "straight A" is used to distinguish between A and A-, then all he is saying is that the median grade is A- and the mode is A. Since at least 50% of the grades given are A or A- and there are more As than A-s, there would be at least 25% As, possibly a lot more.
Note also that the median being A- tells us nothing about the bottom half of the grades. If no professor even gave out anything below an A-, the median would still be A-. If such were to be the case, then the 5th percentile, 10th percentile, 25th percentile, etc. would all be A-.
For full disclosure, Harvard should tell us what proportion fo grades are As and what proportion are A-s.
I'm very interested in the topic of grade inflation for various reasons.
The first is that it illustrates the fact that data by themselves is worthless without human judgement. It used to be that we'd say an A at Harvard is worth more than an A at a second-tier university, or vice versa, a B at Harvard might be equivalent to an A at a second-tier college. This relevation rubbishes that interpretation. We now should think that an A at Harvard could be worth less than an A at a second-tier college. We just couldn't tell because Harvard gives out too many As.
In the age of Big Data, we love to fuse data sets, whether the data come from different countries, different competing businesses, different states, etc. Even when the metric is identical, in this case, the Grade Point Average (GPA), it is dangerous to throw the data together without rescaling them.
Meritocracy is supposedly an important part of American culture. In some fields, this appears to be the case. For example, in sports and entertainment, we pay our top stars multiple orders of magnitude more than other athletes or actors. In business, it is controversial that we pay CEOs and other executives hundreds of times more than the rank and file. (The controversy surrounds the fact that management "performance" is not really measurable.)
By contrast, in education, meritocracy is not allowed. We must not differentiate between exceptional academic performance and average academic performance. Everyone must be above average - in fact, everyone must be exceptional. While Harvard is being singled out here, this "grade inflation" phenomenon is really prevalent across all top schools.
Except at Princeton, where the previous Dean imposed a "grade deflation" policy - limiting As to 35% of the total for large classes and 55% for small classes (typically junior/senior classes). As an alumnus, I'm very proud that the school has taken this brave stand, and shown the way forward.
What makes me sad is that Princeton has announced it is "reviewing" this policy. In the several years since Princeton made the change, no other prominent school has joined the fray. The administrators are under immense pressure to do away with the policy because students complain they are placed at a disadvantage and parents (donors!) complain a lot.
Why should American universities not be meritocracies? When these same graduates go out to the workplace, and they eventually become managers, they all grade their staff and often fire people based on performance (many of whom also went to Harvard, etc.). Why do they think that they themselves should not be subject to merit-based evaluation while at school?
A further irony. As a professor, I get evaluated by students every semester. I can tell you that students do not "grade inflate". I get comments ranging from critical to admiring. These are the same students who think they all deserve As.