Nudge Lab
← Nudge Lab

Class findings

What happened when students ran the experiment on themselves: whether reminders changed anything at all, and which style worked for how many of them.

These are invented example figures. They show what the page looks like with a class of 24 students in it. No real student data has been loaded — open Load real class data below and paste the codes your students sent you.
The headline

Reminded, students followed through 27 points more often than when nothing fired.

With a reminder
0.58

follow-through, 688 trials across 5 styles

Silent control
0.31

nothing fired at all · 137 trials

Difference
+27

percentage points — the case for nudging at all

Follow-through by reminder style

Pooled across every student. 1.00 means the work started at once and landed on time; 0.00 means the reminder was ignored.

Alarm Takeover 141 trials
0.71
Blocking Pop-Up 138 trials
0.66
Coach Text 136 trials
0.59
Daily Check-In 134 trials
0.54
Quiet Nudge 139 trials
0.40
Silent Control control — nothing fired 137 trials
0.31
0.00 — ignored1.00 — acted at once, on time
Whose reminder was whose

A style is only counted here once that student's evidence was strong enough to name it — 22 of 24 students so far.

Alarm Takeover 8 students
Blocking Pop-Up 6 students
Coach Text 4 students
Daily Check-In 3 students
Quiet Nudge 1 student

2 students are still gathering evidence — no winner named.

The same figures as a table
StyleTrialsFollow-throughNamed best for
Alarm Takeover 141 0.71 8
Blocking Pop-Up 138 0.66 6
Coach Text 136 0.59 4
Daily Check-In 134 0.54 3
Quiet Nudge 139 0.40 1
Silent Control (control) 137 0.31 0

24 students, 825 trials in total.

Load real class data

Each student copies their code from the app (the button, next to their name) and sends it to you. Paste them here, one per line. Codes carry counts only — no assignment titles, and this page never shows individual names.

Nothing leaves this page. The codes stay in your browser and are not saved anywhere.

How these numbers are made

Every time a reminder fires, that is one trial. A trial scores 1.00 if the student started within two minutes and the work landed on time, 0.80 if they acted but more slowly, 0.50 if they acted but the work was late or they snoozed twice, 0.20 if they dismissed it, and 0.00 if they ignored it.

Follow-through below is those scores pooled across every student and every trial of that style — total score divided by total trials, not an average of averages, so a student with more trials counts for more.

Silent Control fires nothing at all. It is the comparison that shows whether reminders do anything, rather than the student simply being motivated that week. Without it the whole experiment would prove nothing.

A style is only called a student's best when it beat every other style with at least five trials each and its confidence range cleared the runner-up's. Students still gathering evidence are counted in no winner yet.

Open the app · A Design Thinking project. Each student's homework stays on their own device; the only thing that travels is the summary they choose to send. These figures come only from codes a student chose to send.