Science fair judging that a committee can defend line by line.
LiveJudge is science fair judging software. You build the rubric your fair already uses, volunteer judges score at the booth from their own phones, and results normalize across a panel that was never calibrated.
Built for booth judging, not for rushing.
A science fair is a conversation at a table, not a stopwatch. Judges score on a color-graded slider per criterion, with the written guidance for that band sitting under it.
One element splits into as many criteria as it needs, which is the difference between a considered score and one gut number. And because everything runs on a single 1 to 10 scale with weighting behind it, judges never have to hold "up to 15 here, up to 5 there" in their heads. Fine increments capture the real gap between two strong projects that a 1 to 5 sheet would flatten into a tie.
No app, no passwords, no venue WiFi. A judge enters a phone number, taps the SMS code, and starts.
Controls
7.4Sound control, with one confound unaddressed.
Sample Size
5.8Enough trials to hint, not to conclude.
Drag any slider. The guidance and the score follow.
A bigger seat.
All the same features.
Hand a judge any tablet in landscape and LiveJudge opens a larger judging layout. Every element sits in a rail with its running subtotal, and the open element fills the screen with bigger sliders and room to read the guidance.
Turn it sideways
Any tablet in landscape gets the larger layout. Nothing to install or switch on.
All the same features
Slider and heads-up modes, written guidance, the same weighting and the same math.
Next, not back
Previous and Next buttons on every element, so heads-up judges never hunt for the list.
Preview it first
Tap Tablet View in Class Studio to see exactly what your judges will see.
A kind judge's 7.5 is not a grumpy judge's 7.5.
The judging screen shows 1-10 sliders for each evaluation criterion. Look at the difference between a bare 1-10 slider and a "judge guided" version of the same slider.
Without judge guidance, two judges evaluating the same thing might land on wildly varying scores.
Band 4 of 6Sound control, with one confound unaddressed.
Both judges are reading the same sentence, so both land in band 4. They still do not agree exactly, and closing that last gap is what normalization is for.
A fair recruits its judges by email and briefs them in a hallway twenty minutes before the doors open. Hand that panel a bare 1 to 10 scale and the same project comes back a 9 from one table and a 6 from the next, because a 7.5 means something different to a retired chemist than it does to a first-year teacher. Neither of them is wrong. Nobody ever told them what a 7.5 is.
So LiveJudge tells them, on the screen, while they score. Any criterion can carry six written bands in your words, which is where your fair says out loud what a controlled experiment looks like at a 4 and what it looks like at a 9. The guidance under the slider rewrites itself the moment the score crosses into the next band.
That pulls a volunteer panel together at the source, before a single algorithm runs. Normalization still settles the judge who marks hard. It simply has far less left to fix.
Six bands, your words
Write what a 4 looks like and what a 9 looks like once, when you build the rubric. Every judge scores against your fair's definition instead of their own.
It rewrites as they score
Move the slider and the sentence underneath changes to the band the score just landed in. Every criterion carries its own.
No briefing to run
Judges arrive at eight and are at a table by half past. The standard is on the screen while they score, so a volunteer who read nothing still marks to it.
Then the math runs
Anchored marks arrive closer together, so normalization and the high/low drop are correcting a smaller gap, not rescuing the result.
Judge 4 grades everyone hard. Every opinion counts equally here, which drags this contestant down.
Your judges are volunteers. They will not be calibrated.
A retired engineer and a first-year teacher will not mark the same project the same way, and on a paper scoresheet that difference quietly outweighs the science. Try it on the left: one project, four judges, and the fourth marks everyone hard. Switch the method and watch the result move.
You choose which method your fair uses from Raw average, High/low drop, Z-score, and Olympic-Z, and the raw marks are always preserved underneath, so you can show your work when a parent asks. Ties are detected automatically and every result carries a full audit trail and a CSV export.
Turn the names off.
Choosing a method settles the volunteer who marks hard. It does nothing about the volunteer who recognizes the school.
Turn on blind judging and the judging interface stops showing who an entry belongs to, so judges score the entry rather than the name on it.
Who has judged what, and what is still missing.
Build your judge pool once and assign across every category from the console. Decide what finished means: every assigned judge, or any three of eight. Name a head judge per category for oversight and ties. Set advancement once and it applies itself.
Scoring Admin
RecalculateExport Scores| Rank | Project | Judges | Round score Provisional |
Final score Provisional |
|---|---|---|---|---|
| 1 | #31Tidal MicroplasticsMaya Ellison | J. Croft88.50B. Adeyemi91.20P. Lund86.40 |
88.70 | 88.701 of 2 rounds in |
| 2 | #08Solar Kiln DryingDevon Park | J. Croft84.00B. Adeyemi84.00P. Lund84.00 |
84.00 | 84.001 of 2 rounds in |
| — | #22Wind Load on TrussesRuth Nakamura | J. Croft79.30 2 more judges needed |
Not scored | Not scored |
| — | #45Enzyme Cascade TimingIker Salas | Nobody yet 3 more judges needed |
Not scored | Not scored |
A printout in every student's hand.
When scoring closes, every entry already has an official report of each judge's marks, per round and per element.
Abstracts can be reviewed the same way before the fair opens.
| ELEMENT (weight) | JC | BK | AP | MK | AVG |
|---|---|---|---|---|---|
| Research Question10% | 8.40 | 7.20 | 9.00 | 8.60 | 8.30 |
| Design & Methodology25% | 21.00 | 18.50 | 22.50 | 22.00 | 21.00 |
| Execution25% | 22.00 | 19.00 | 23.00 | 22.00 | 21.50 |
| Written Presentation20% | 16.50 | 15.00 | 18.00 | 16.50 | 16.50 |
| Oral Presentation20% | 17.50 | 15.50 | 18.50 | 16.50 | 17.00 |
| JUDGE TOTAL | 85.40 | 75.20 | 91.00 | 85.60 | 84.30 |
The room votes too.
Every attendee gets a branded event page and a QR code, with no app to install: every project, profile, document and photo, and live standings if you choose to show them.
Open a People's Choice ballot and the floor votes from the same place. Set how many votes each person gets, geo-fence to a quarter mile so only people actually at the fair can vote, and put a sponsor on the ballot if one is funding it. Blend the audience result with the judges' at whatever weighting you choose and you have a composite award that reflects both.
Every project on the floor today. One vote each, from anyone at the fair.
You have 1 vote left on this ballot
The ballot your audience sees, on their own phones.
What fair directors ask us first.
Do volunteer judges have to download anything?
No. Runs in any phone browser. Nothing to install, nothing to update. A judge opens a link or scans the QR on the project's placard, signs in with an SMS code, and scores. There is nothing to install and no password to reset at 8am on fair day.
We use our own rubric and we are not changing it.
You should not have to. You define the categories, the elements inside each one, the weight of every element, and the score increments. LiveJudge honors the rubric you already use rather than bending your fair to fit a template. Send us your paper scoresheet and we will build it for you, free.
Our judges vary enormously. How is that handled?
Choose the normalization method from Raw average, High/low drop, Z-score, and Olympic-Z and a strict panel and a lenient one land on equal footing. The raw marks stay available underneath, and every result carries an audit trail.
Our volunteers know the schools, and sometimes the students. Can we hide that?
Yes. Turn on blind judging and the judging interface stops showing who an entry belongs to, so judges score the entry rather than the name on it. A volunteer scores the project rather than the name on it, and the committee keeps the full record underneath with names attached, so a result can still be defended line by line.
Not every judge gets to every project. Does that break it?
No, and this is the part most fairs need. Set the requirement per category: every assigned judge, or any three of eight. A result finalizes when the requirement is met, and the console shows exactly which projects are still short and by how many.
Can we review abstracts before the fair?
Yes. Run the paper round on its own rubric ahead of the event and the booth round on fair day, inside one event, with advancement between them if you want it.
What do students actually get?
An official score report per entry: every judge's marks, per round, per element, with the weight beside each. They print it from their own phone. For most students it is the only written feedback the fair produces.
Will the venue WiFi cope?
No dedicated network required. Judges, contestants and fans use their own devices and connections. A school gym with no usable network is the normal case, not the exception.
What does it cost?
Pay-as-you-go. No subscriptions, no per-seat license, and no credit card to start. Events with 10 contestants or fewer are free. Bigger ones pay $3.49 a contestant, never more than $299 an event, plus 2.5% of anything you sell through our checkout. Full detail is on the pricing page.
Other events on the same engine.
Every judged event is a rubric, a panel and a room. See what else LiveJudge runs, including pitch and academic competitions, or go straight to pricing and the FAQ.