I was listening to a report on NPR’s “Morning Edition” the other day when I heard an exchange that was both amusing and thought-provoking. The report, by national correspondent Tovia Smith, was about the growing number of schools that are trying to instill “grit” — perseverance in the face of adversity — in their students. Smith focused on one such school:

“Tom Hoerr leads the New City School, a private elementary school in St. Louis, MO, that has been working on grit. ‘One of the sayings that you hear around here a great deal is, “If our kids have graduated from here with nothing but success, then we have failed them, because they haven’t learned how to respond to frustration and failure,”‘ says Hoerr.

After years of focusing on the theory known as “multiple intelligences” and trying to teach kids in their own style, Hoerr says he’s now pulling kids out of their comfort zones intentionally.

“The message is that life isn’t always easy,” Hoerr says. His goal is to make sure “that no matter how talented [students are], they hit the wall, so they can learn to pick themselves up, hit the wall again and pick themselves up again, and ultimately persevere and succeed.”

“It is a major adjustment for everyone.”

Part of the adjustment involved seems to be the whiplash associated with following one set of educational trends (“multiple intelligences” and “learning styles”), then shifting gears to follow another set (“grit” and the “growth mindset”).

Grit, an idea developed by Penn professor Angela Duckworth, and the growth mindset, a concept researched by Stanford professor Carol Dweck, have more empirical support than the appealing-sounding but unproven theories of multiple intelligences and learning styles.

But even putting the question of educational trends aside, the experience of principal Tom Hoerr as documented in the NPR segment brings up a question that parents and teachers wrestle with all the time: Should we be making learning easier for kids — or harder?

The answer, according to research in cognitive science and psychology, is both.

First, let’s think about how and why we might make learning easier. This has to do with what psychologists call “cognitive load” — the amount of information we have to keep in mind as we solve a problem. Decades of research has shown that this capacity is quite limited.

Harvard psychologist George Miller famously said that we could keep seven pieces of information in our minds at a time (he called it “the magic number seven”), but he was looking at studies that used numbers or letters or simple symbols. When the pieces of information are more complex, like concepts or facts, the number of them that we can keep in mind goes down, to about four. And if we are actively manipulating or combining those pieces of information, as we do in most kinds of real-life problem-solving, the number of things our minds can hang on to drops even further, to maybe two or three.

The problem is that many tasks we ask students to do impose too great a cognitive load. They lose track of what they’re doing, they make mistakes, they get lost and give up. Even if they hang on long enough to solve the problem, they don’t have enough mental capacity left over to reflect on what they’ve done — and reflection is where learning really happens.

So: the way in which we should make learning easier is to reduce cognitive load, especially when we are introducing new or complicated materials. (Parents and teachers, who are already experts at this stuff, often don’t realize how much cognitive load they’re imposing on kids and other novices.) Slow it down. Break complicated ideas into smaller pieces, taking them one at a time. Offer lots of opportunities for practice with feedback. Avoid using jargon and other technical terms. Eliminate extraneous or distracting information and focus only on what the learner needs to know at this moment.

Makes sense. But what about making learning harder? Is there ever a reason to do that? Yes, and the reason is twofold. The first reason to make learning harder is to make it interesting. Learning something new and complicated is hard in itself, as we saw above. Lightening the learner’s cognitive load will allow her to learn more effectively without becoming frustrated or confused.

But once the learner has attained some degree of mastery, ratcheting up the difficulty will help her stay in her “sweet spot” of engagement, where the task is not too hard as to be frustrating and not so easy as to be boring. This is also the place where learners can practice encountering adversity and challenge and overcoming them, a key experience in the development of grit.

The second reason to make learning harder is that it makes learning work better. UCLA psychologist Robert Bjork has developed the idea of “desirable difficulties” — difficulties that we actually want to introduce into students’ learning to make it more effective. Bjork notes that many of the learning activities that make students feel competent and successful — like reading over a textbook passage several times so that it feels familiar — actually do very little to help them learn. What they should do instead is something like this: close that textbook and ask themselves to recall from memory what they’ve just read.

It won’t feel as good. They’ll struggle to remember the words that were just in front of their eyes. But this activity, known as retrieval practice (or simply self-testing) is an example of a desirable difficulty that will dramatically increase students’ learning.

Unlike learning styles, cognitive load theory and the theory of desirable difficulties have lots of research support. You won’t have to worry that another fad will come along to displace them in a few years. And in the meantime, you’ll have a good answer to the question of whether we should be making learning easier or harder. Remember: it’s both.