The research behind Second Nature

Why does a fact have to come in under 3 seconds?

A fact a child remembers comes back faster than a fact they work out. Counting up or using a strategy takes longer, so a short limit separates facts a child has memorized from facts they can still figure out. Second Nature draws that line at 3 seconds.

In the app: a bar under each fact empties over 3 seconds. A correct answer before it empties counts. A half of the square fills only on a counted answer given at least one full round after the child was last shown that answer, so an answer still in short-term memory never counts as learned.

Why does the app pick every card, and aim for about 85% correct?

Practice that is too easy teaches nothing new, and practice that is too hard turns into failure a child notices. A 2019 study of how learning systems train found the fastest learning when about 15% of tries fail.

In the app: of 20 cards, 3 to 7 work on new or shaky facts, more after a strong round and fewer after a weak one. New facts come easiest first and climb as answers land, and no more than six facts are in progress at once. Every rule aims at about 17 of 20 right.

Source: Wilson, Shenhav, Straccia & Cohen, 2019

Why show the answer after a miss?

Seeing the correct answer after an error is what produces learning. In a study of 258 people learning word pairs, giving the right answer after a wrong one raised retention a week later by 494%. Feedback after correct answers made little difference.

In the app: every miss or timeout shows the full fact with its answer for 1.5 seconds, hides it, and has the child type it from memory. The fact returns 3 to 5 cards later with the normal bar. When misses pile up at one level of difficulty, new facts there start on a New card with the answer shown first.

Source: Pashler, Cepeda, Wixted & Rohrer, 2005

Why do learned facts keep coming back?

Recall that is spread over time lasts longer than recall packed into one session. Karpicke and Roediger found that delaying the first test improved memory days later, and that evenly spaced tests beat gaps that grow longer when memory was checked two days on.

In the app: a newly filled fact comes back two rounds later, then after 3, 7 and 14 days. A fact that slips steps back and is asked again sooner.

Source: Karpicke & Roediger, 2007

Why mix addition, subtraction, multiplication and division?

When problem types are mixed, a student has to decide which operation each one needs, and that choice is part of what gets learned. In a classroom study, students who practiced math with mixed problem types scored higher on a later test than students who practiced one type at a time.

In the app: review cards mix every operation the child has started, so each card means reading the sign. A switch on Home can limit practice to addition and subtraction or to multiplication and division.

Source: Rohrer, Dedrick & Stershic, 2015

Why does 56 ÷ 8 follow 7 × 8 = 56?

Division is multiplication run backward, and subtraction is addition run backward. Children can learn to use that link as a shortcut. In a study of 81 children in kindergarten through third grade, a helper fact such as 7 + 7 = 14 stayed on screen while the child answered 14 − 7. Children trained this way beat two comparison groups at knowing the link and using it.

In the app: after a correct timed answer, the fact can turn into its partner in place: 7 × 8 = 56 becomes 56 ÷ 8, and the child finds the 7. The link then comes off. Only answers given without it count, and each partner has its own half of the square.

Source: Paliwal, 2013

Why are there no points, coins or prizes?

Tangible rewards can undercut a person's own interest in a task, and a review of 128 experiments found the effect was larger in children than in college students. Second Nature's rewards are the progress itself: a grid that fills and answer times that fall.

In the app: no coins, characters, leaderboards, notifications or prizes. Small visual and haptic feedback fires only on correct answers.

Source: Deci, Koestner & Ryan, 1999

Why does the app praise progress instead of the child?

Children praised for being smart persisted less and enjoyed the work less after a setback than children praised for their effort.

In the app: summary lines name what changed, such as "Six facts got faster this round," and never call the child smart or talented.

Source: Mueller & Dweck, 1998