Skip to content

§23 · 25 tips

Which skills are worth learning

Which of welding, English and certificates really pay off, how do you learn without wasting time, and where do you apply for a professional title and is it worth it?

Tips from this section that apply in the United States.

Count schooling in the death-rate ledger too: each extra year of education lowers adult death risk by about 1.9%

Each extra year of schooling goes with an average of about 1.9% lower risk of death in adulthood. The effect is largest for people aged 18 to 49, about 2.9% lower per extra year, and only 0.8% lower for people 70 and over. People who can keep studying often had better circumstances to begin with, and it is not possible to tell how much of the difference comes from the schooling itself.

  • Moderate value
  • Evidence A
  • Buys life
  • Disputed

Turn “study or work” into a calculation: compare the three years of pay you'd earn by working now against the yearly income gap you'd gain over the next several decades by staying in school

Going to work now lets you earn a few years of wages sooner, but you collect that money once. Staying in school means earning somewhat more than you otherwise would every year, for 30 or 40 years in a row. So the two sums have to be put side by side. In Chinese cities, one more year of schooling went with roughly 4.0% to 10.2% higher yearly income, but those are figures from before 2001 and give only a rough idea.

  • High value
  • Evidence B
  • Saves money

Start from the baseline: the global average private return to one more year of education (the part that lands in your own income) is about 9% a year

Averaged across the world, one more year of schooling raises personal income by about 9% a year, and the return for women has consistently been higher than for men. That figure is about formal schooling; it can't be applied directly to signing up for a class. Use it as a yardstick: if a course asks for 30,000 (3万) CNY and half a year, ask whether the change in income it brings is worth as much as one more year of school.

  • Top value
  • Evidence A
  • Saves money

When choosing a skill, give priority to work that needs your hands and on-the-spot judgment; it is the hardest to automate away

By one U.S. estimate, about 47% of jobs were at high risk of being taken over by computers; later estimates made task by task came out much lower. There are three hurdles machines find hard to clear: manual dexterity, coming up with new things, and dealing with people. So work that is done on site, in non-standard surroundings, with judgment on the spot is harder to replace, and jobs with fixed procedures that mostly process information carry the highest risk.

  • Top value
  • Evidence B
  • Saves money
  • Disputed

For the same money and time, pick a short program that leads straight to a job

What policy encourages is “short, quick” training that leads to a job in a short time, plus micro-majors at vocational colleges. Choose a program by one test: when you finish, is there a specific position you can send a resume to? A program that takes a year or more and still tells you to “build up experience” afterward costs far more time than the listed price. Write down your exit condition before you start: how far along you will be when you stop if you still haven't found work.

  • Moderate value
  • Evidence C
  • Saves money

After studying, close the book and test yourself instead of rereading

Reading something a second time feels familiar, but it doesn't make it stick. Closing the book and quizzing yourself works clearly better: adding quizzes in class raised grades by a medium amount. In a ranking of ten common study methods, self-testing sits in the top tier. After finishing a section, write out the main points from memory first, then flip back to check.

  • High value
  • Evidence A
  • Saves time

Spread the same study time over several days instead of learning it all at once

The same amount of study time spread over several days is remembered better than the same time done at one sitting. 317 experiments tested this, and the conclusion was consistent. There is also a practical rule: the longer you need to remember something, the longer the gap between reviews should be. If the exam is tomorrow, review today; if it is a month away, go over it again every few days.

  • High value
  • Evidence A
  • Saves time

Don't make highlighting, rereading and summarizing your main way to study

What students use most are highlighting and rereading, yet in the ranking of ten common study methods both sit in the lowest tier. Writing summaries, the keyword mnemonic and picturing images for a text are in the lowest tier too. They are not completely useless, only that the conditions under which they work are too narrow, and the same time is better spent on self-testing.

  • Moderate value
  • Evidence A
  • Saves time

Mix several kinds of problems when you practice, instead of doing twenty of one kind in a row

Doing a long string of the same type of problem goes smoothly at the time, but afterward you may not be able to tell which method to use. Practicing several types mixed together works better in the long run, by a medium amount. But it depends on the material: it works best for learning to recognize images and categories, a bit less for math problems, and for memorizing vocabulary, practicing in blocks is actually better.

  • Moderate value
  • Evidence A
  • Saves time

Don't choose how to study by “I'm a visual learner, he's an auditory learner”

Many people believe each person has a learning style, such as visual or auditory, and that you learn best when taught to match it. Very few studies have tested this properly, almost none support it, and a few found the opposite. People do have preferences, but there is no evidence that teaching to a preference makes people learn better, and there is no need to pay for a style assessment.

  • Moderate value
  • Evidence A
  • Saves time

For things you have to memorize, quiz yourself in the way you will use them later, instead of reciting them as written and calling it done

Quizzing yourself with the book closed pays off beyond the exact words you memorized. In a combined analysis of 192 comparisons and more than ten thousand people, it also helped on questions in a different format, on application questions and on reasoning questions, by a medium-to-small amount. But two cases got almost nothing: swapping the question and the answer, and material that was only seen in the book and never tested.

  • Moderate value
  • Evidence A
  • Saves time

Before going on to a master's after your bachelor's, first work out the tuition and the two or three years of pay you give up, then see whether your field pays more

A master's costs tuition, and it also costs you two or three years of pay, which is usually the bigger part. In a UK study that removed personal differences, at age 35 women with a master's earned only about 2% more, and men about 2% less. Law, economics and management paid about 20% more, while art, English and philosophy paid more than 10% less, so work it out for your own field.

  • Top value
  • Evidence B
  • Saves money
  • Disputed

Before going on to a PhD after your master's, do it only if the job you are aiming at clearly requires one, and count the mental strain before you start

In the same UK study, after removing personal differences, at age 35 women with a PhD earned about 8% more than those with only a bachelor's, while men earned about 9% less. The mental strain of a PhD is also heavy: about a quarter of PhD students across countries have clear symptoms of depression, and about one sixth have clear symptoms of anxiety. So a PhD pays off only when the job you are aiming at hires only PhDs.

  • Top value
  • Evidence B
  • Saves money
  • Disputed

Mainland China only 12 tips

These tips describe mainland-China rules and do not apply in the United States; they are here for people living in China.