I was recently told that in many schools in the US (maybe other countries too), the minimum grade for a graduate course is B or B−. Meaning that B is considered as a passing score. I would like to know whether this is true. Why should it be the case? Does this mean that grading in graduate school is easier?
In some sense grading may be "easier", but it's also just that the effective grading range is compressed, and that grades per se matter less than in undergraduate classes.
In many graduate programs, getting a grade lower than B (or even A-) is considered a warning that you need to seriously knuckle down and start doing better. To take a random example, UIUC has a campuswide minimum GPA of 2.75 (B-), but most departments set higher minimums (as high as 3.25, or B+) Here's a relevant quote from an old grad school blog post:
My first semester in grad school, my M.A. advisor laid it out for to me: "If you get a B in a grad school class, you should ask the professor whether you should consider dropping out of the program. If you get a C, don't bother asking."
What this means is that the range of grades in grad school is effectively compressed, going from A to C instead of A to F. Getting a C in a grad school class is, in many cases, practically equivalent to getting an F in an undergraduate class. This grade inflation probably has historical roots in a sort of academic politeness in which it became considered more and more insulting or embarrassing to give someone low grades.
However, it's not as significant as it may seem, because grades are typically much less important in grad school than in undergrad. In grad school (especially PhD programs) you are working much more closely with faculty and they monitor your progress at a much finer grain. It's not just about getting a good grade, but about conducting research, making progress on your degree milestones (e.g., writing an MA thesis or PhD dissertation), and so forth.
In grad school, your class grades are meant to be not just an evaluation of your performance in that class, but a signal as to whether you are developing the skills necessary to succeed in the program as a whole. It would be unusual (probably unthinkable) for a grad program to suddenly give someone the boot simply because their GPA dropped too low; rather, a series of low grades will lead to increasing concern from the faculty, with meetings, emails, etc., gradually escalating from "Is everything ok?" to "You need to start working harder" to "Get your act together or else".
I've known grad students who received multiple grades in the B range, and this is what happened to them; they were subjected to increasing pressure from the faculty, with increasingly more explicit suggestions that they could be kicked out of the program if they didn't improve their performance.
That said, it is true that in some cases grading in grad school can be "easier" than in undergrad. My personal experience has been that in some cases faculty members are willing to be somewhat more flexible on grades as related to the actual course content, as long as they are satisfied that you got what you personally needed from the class. For instance, if you are specializing in Topic X and you take a requred class in Topic Y intended to broaden your background, the professor may give you an A even if your paper in the class is only adequate, if the professor understands that that is not your specialization and you don't need to fully master it in order to succeed. The flip side of this is that if you take a class in Topic X (your specialty), the professor may hold you to a far more exacting standard, knowing that you really must be on top of that material.
I can tell you for a fact that people can get worse grades than B- in graduate courses. People can even fail in graduate courses. It happens pretty rarely, though, because the population of students in a graduate course are highly selected.
If you read other questions on this site, you'll see that graduate school admission is a strong filter: most people need to be both excellent and a little bit lucky to get in, because there are so many excellent students who want to go to grad school. That means most grad students are smart, hard-working, and ambitious enough to put in the work to deserve the grade. As for undergraduates in graduate-level courses: if they weren't willing and able to put in the work, why wouldn't they choose one of the easier options? Moreover, most schools allow students to drop a course without penalty early in the semester, and the people doing poorly are the ones likely to drop.
Now, grade inflation can certainly happen. But by the time you're dealing with graduate courses, usually the class is pretty much full of people who want to be there and are capable of making the grade.
Part of the answer is in disputing some assumptions in the question, as remarked upon to some degree in the other answers. That is, for one thing, graduate admissions are most often selective enough so that every grad student is probably able to do the work at a level that would, indeed, perhaps be a "B+" in "grading on a curve", if the population in the course were far less selective than it ... in fact... is. This is in contrast to the common conduct of undergrad courses, especially lower-division ones, where there is often no pre-filtering at all. The students self-select, and their own judgement may be flawed.
But then why might a "B+" be considered "a problem"? What is the "problem" if it's not "failure"? Well, it's not-at-all "failure" by undergrad standards, but that standard is far, far too weak for an apprentice professional. Unlike undergrad courses, where a "C+" may fulfill "a requirement", but absolutely not indicate competence, grad students need to be more-genuinely competent. A "B+" is an indication of some gaps in competence.
So, yes, those undergrad "C+"s are not evidence of "sufficient competence", at all. Indeed, that's why people who get that minimal "C+" in calc I invariably have terrible trouble in calc II, etc. The primary reason such grades are tolerated at all is that, in fact, much of the function of lower-division undergrad math is filtering, so that content mastery is nearly irrelevant in many cases.
Again, most often, grad students are not being filtered very much after admission, so the issue is genuine content mastery, not "getting by".
I think the idea is simply that people who get into graduate programs must have gotten A's or B's in nearly all of their undergraduate courses. So if graduate courses are graded in the same way as undergraduate courses, it makes sense that most of the students in them should get A's or B's. Meanwhile, since a C in an undergraduate course would be a strong strike against you for admission to grad school, it is also a strike against your continued enrollment in a graduate program.
It definitely depends on the program of study...I believe the grading system is harsher for people who study something career specific---AKA nursing, physical therapy, Occupational therapy, med school, law school, etc..These programs tend to have a lot more drop outs/people who are unable to maintain the 3.0 minimum...People who study something like a Phd in Sociology, philosophy, or an M.S in Biology usually get higher grades because these programs tend to put less emphasis on passing tests and put more emphasis on producing research... Which is very different from lets say law school, where there is a huge emphasis on grades and rankings...