Intermittent Attention: Designing Review for ADHD

WordLinea’s design follows one principle: don’t ask for a stretch of uninterrupted attention. Reviews have no sets, every answer is saved as it’s given, and words come back when they fall due. The principle comes from ADHD research. Each of the four sections below takes one design and the evidence behind it.

Attention breaks at question 5

AnsweredNot answered

A set of 10Set unfinishedOne at a time4 saved

01No stretch of time required

In 1997, Barkley proposed a theory of ADHD that went on to spur a large body of research: the core problem is behavioural inhibition, the ability to put the brakes on a response, and executive functions such as working memory and the regulation of motivation and emotion depend on it. Reviewing the evidence of the time, he found the strongest support for deficits in inhibition, working memory, the regulation of motivation and motor control, and said the model needed more research to test it.⁠[1]

Nine years later, Castellanos, Sonuga-Barke and colleagues argued that the findings and their methods challenged inhibition as the root cause. They put the weight on variability within a person, for which the evidence was growing: fast one moment, slower the next.⁠[2]

In 2013, Kofler and colleagues pooled 319 studies in a meta-analysis. Children and adolescents with ADHD varied more in their response times than their peers, with an effect size of g = 0.76; for adults it was 0.46. Once that variability was accounted for, they were not slower on average.⁠[3]

The same average, more spread

ControlsADHDFasterAverageSlower

How big the differences are

Difference between people with ADHD and controls (effect size)

Variability · children and teensg = 0.76Variability · adultsg = 0.46Delay discounting · all agesd = 0.43

The problem is variation, not speed: a similar average, more spread, and no telling which moment will break.

WordLinea’s approachReviews have no sets, no “3 / 10” progress and no daily goal. Each answer is saved as it’s given; stop the next second and every answer so far is kept.

02Feedback that doesn’t wait

In 1992, Sonuga-Barke and colleagues had hyperactive and control children make the same choice over a series of trials: a small reward now, or a larger one after a wait. Under most conditions both groups earned as many points. The difference appeared when the session was a fixed 20 trials: there, choosing the small reward made the session end sooner, and hyperactive children chose it more often. When the session was a fixed 10 minutes, choosing early didn’t mean leaving early, and the two groups chose alike. The authors concluded that what these children cared about was not when the reward came but how long they waited in total.⁠[4]

Does the small reward let you leave?

Large reward: wait every timeSmall reward

Fixed 20 trialsEnds soonerFixed 10 minutesSame length

Sonuga-Barke later built this into a dual pathway model: poor inhibitory control and delay aversion are two independent characteristics of ADHD that can exist side by side.⁠[5]

This isn’t the result of one or two experiments. In 2016, Jackson and MacKillop combined 25 case-control comparisons from 21 studies, 3,913 people in all: people with ADHD discounted delayed rewards more steeply, valuing a reward less the longer they had to wait for it, with an effect size of d = 0.43. Children and adults, real and hypothetical money, gave similar results.⁠[6]

In 2005, Luman and colleagues reviewed 22 studies with 1,181 children in total. Rewards improved performance for both groups, somewhat more for children with ADHD, and children with ADHD preferred immediate rewards over delayed ones.⁠[7]

WordLinea’s approachEvery answer shows at once whether it was right and when the word will come back; a misspelling is marked letter by letter.

03Every review is a question

Memory Is Made in Recall showed that testing yourself keeps material longer than rereading it. Does that hold for people with ADHD?

In 2016, Knouse, Rawson, Vaughn and Dunlosky recruited 25 college students with ADHD and 75 without. Each learned two lists of 48 words: one by studying it eight times in a row, the other by alternating study and tests. Two days later they recalled what they could, and remembered more from the tested list: both groups showed a moderate testing effect, with d between 0.50 and 0.57.⁠[8]

One person, two lists

StudyTest

List ARecall 2 days laterList BRecall 2 days later

WordLinea’s approachInstead of asking you to grade yourself, every review is a question: recognise the meaning, put the word back in its sentence, or spell it from its sound. Answering is the act of recall.

04No streak count

Across seven studies, Silverman and Barasch found that when a log showed an intact streak, people were more likely to keep going; when it showed a broken one, they were less likely to, though what they had done was the same. The difference came from how the log presented it. The effect was stronger when people blamed themselves for the break, and weaker when a broken streak could be repaired.⁠[9]

Same work, shown differently

DoneNot done

Log 13-day streakLog 2Broken

The other side is Dai, Milkman and Riis’s fresh start effect: after a new week, a new month or a birthday, searches for “diet”, gym visits and commitments to goals all rise. The authors suggest these landmarks move past imperfections into a previous period.⁠[10] A fresh start can bring motivation, and it needs a past to put behind you.

WordLinea’s approachNo streak count and no daily goal. With no streak to break, there’s no fresh start to wait for. However long it has been, your words are waiting in the order they fall due.

05Four designs, one principle

Who was studied

Study	Children	Adults	General	Kind
Sonuga-Barke 1992				Experiment
Luman 2005				Review
Kofler 2013				Meta-analysis
Jackson 2016				Meta-analysis
Knouse 2016				Experiment
Silverman 2023				Experiment
Dai 2014				Field data
No participants of their own: Barkley 1997, Castellanos 2006, Sonuga-Barke 2002

Kofler and colleagues also found that adolescents and adults with ADHD were hard to tell apart from other clinical groups in how much their responses varied.⁠[3] Attention that comes and goes isn’t ADHD’s alone, and a design made for it suits more people.

Attention breaks, so WordLinea makes every answer count: saved as it’s given, answered at once, a recall each time, and no streak count.