Home › Psychology
Cheat sheet
Psychology on one page
The summary, key terms and the trap to avoid for every lesson, in order. Print it, or read it, cover it, and say it back.
- Testing yourself makes memory do the work, and that makes it last.
- Rereading feels easier but fades faster (Roediger & Karpicke, 2006).
- Practice testing is one of two "high utility" methods (Dunlosky et al., 2013).
- Try hard before you peek; checking after a wrong try still helps.
- retrieval practice
- recalling information from memory, e.g. by testing yourself
- rereading
- going over material again; feels familiar but builds little lasting memory
- high utility
- Dunlosky's top rating for study methods that work widely
Watch out: Peeking too soon. If you look at the answer the moment it feels hard, you skip the struggle that does the work. Give yourself a real try first, even if you get it wrong. Getting it wrong and then checking still helps you learn.
- Memory fades fast after first learning; each spaced review slows the fade.
- Spaced study beats cramming across hundreds of experiments (Cepeda et al., 2006).
- Space your reviews out. A simple schedule is day 1, 2, 4, 8, 16. It's a rule of thumb, not a tested optimum.
- Make each review a self-test, not a reread.
- spacing effect
- learning lasts longer when reviews are spread out over time
- forgetting curve
- how quickly memory fades after learning, first measured by Ebbinghaus
- cramming
- studying everything in one long session just before a test
Watch out: Spacing by rereading. The gaps help most when each review is a test (see the lesson before this one). Spaced rereading is better than cramming, but spaced self-testing is better still.
- Interleaving mixes problem types, so you must choose the method each time.
- It feels harder and scores lower during practice, but wins on the later test (Rohrer & Taylor, 2007).
- Use it once you have met each type at least once.
- Studyfern's timed practice already shuffles questions for you.
- blocked practice
- doing many problems of one type in a row
- interleaved practice
- mixing different types of problem in a random order
Watch out: Mixing before you know the basics. Interleave once you have met each type of problem at least once. On day one of something brand new, a short block of practice is fine.
- Brief, occasional switches to a different task kept performance steady on a long, dull task (Ariga & Lleras, 2011).
- Work in short focused blocks with one clear goal each.
- Keep the phone out of sight.
- Rest away from screens in your breaks.
- Pomodoro technique
- 25 minutes of focused work, then a 5-minute break; a longer break after four rounds
- vigilance decrement
- the drop in attention during a long, steady task
Watch out: Spending the break on your phone. Scrolling keeps your mind busy and makes it hard to stop after five minutes. A break that rests your eyes and body is likely to rest you better than scrolling.
- A good plan says when, where and what.
- If–then plans are kept more often than goals alone (Gollwitzer & Sheeran, 2006).
- Start small enough to keep, then grow.
- Combine the path: test yourself, space, mix, focus.
- implementation intention
- an if–then plan: "If [situation], then I will [action]"
- habit
- a behaviour that starts automatically from a cue
Watch out: Planning too much. A plan for three hours a day that lasts one week likely does less than twenty minutes a day that lasts three months. Start small enough that you will keep it, and add more once it is a habit.
- A tutor prompt gives a role, a level, a task and one kind of feedback.
- Ask the AI to name the mistake and let you try again.
- Shadow native recordings, not the AI.
- Trust the AI on grammar and words, not on how you sound.
- Use made-up personal details.
- tutor prompt
- instructions that set the AI's role, level, task and kind of feedback
- corrective feedback
- being told about a mistake, ideally so you can fix it yourself
- shadowing
- speaking along with a native recording, a moment behind it
- voice-capable AI assistant
- an AI chat app you can talk to out loud
Watch out: Trusting the AI's verdict on your pronunciation. An AI assistant can be very good at grammar and vocabulary, but its feedback on sounds is hit and miss: