How to assign judges to science fair projects

Seven steps from a registration list to route sheets

1. Count the seats. Multiply the number of projects by the judges each needs, and by the number of rounds. That is how many judge visits the day has to hold.
2. Collect every judge's details before the day. Ask each judge for their school or employer, the categories they can judge, the hours they are there, the most projects they will take, and any project or student they know.
3. Check the capacity. Add up the judges' maximums. If the total is below the seat count, some panels will be short whatever you do, so recruit or raise maximums now.
4. Decide the hard rules. Choose whether a category match is required or only preferred, and whether two judges from one school may sit on a panel.
5. Assign. Paste the projects and judges into the assigner and run it. Every conflict and maximum is kept, and any seat it cannot fill is listed with the reason.
6. Review by hand. Look over the grid, move anyone you know should be elsewhere, and lock the visits you are sure of before rebalancing the rest.
7. Print and hand out. Print a route sheet per judge, the judge list per project for the floor, and the check-in list for the desk.

Why the arithmetic comes first

Most judging days go wrong before anyone arrives, because the numbers never added up. A fair of 36 projects with 3 judges each needs 108 judge visits. If each of 12 volunteers will take at most 10 projects, the judges can cover 120 visits between them: enough, with a little room. Take away two judges and the same fair is short before a single conflict is counted. The category judging guide of one large regional fair (listed below) organises judges into teams of two to four, each seeing eight to twelve projects at about eight to ten minutes each, which is a useful check on what a maximum should be.

Worked example: the sample fair

The judge assigner ships with an invented sample fair: 36 projects from four schools in 6 categories, and 12 volunteer judges, each with two categories, a school or employer and a maximum of 10. The rules ask for 3 judges per project, a minimum of 6 projects per judge, category match preferred, no judge on their own school's project and no two judges from one school or employer on a panel.

The assigner filled 108 of the 108 seats. Every judge ended with between 8 and 10 projects. 4 visits went to a judge outside their categories. Some were unavoidable: with the one school rule and the declared conflicts, no arrangement that keeps every judge in their own categories fills more than 106 of the 108 seats. 21 visits matched a judge's expertise tags to the project. The independent check found 0 broken rules. The search ran 40,000 steps and, like every run, gives the same answer for the same fair and seed.

What to check before you print

Read the grid judge by judge. Look for anyone you know should not be with a particular project for a reason nobody entered, a judge whose projects are spread across two rooms, and panels that are all newcomers. Move them by drag or from the lists; a move that breaks a rule is allowed but stays flagged in red. Then lock what you are happy with and rebalance the rest, which keeps your locks and moves as few other visits as it can.

Common mistakes

• Collecting conflicts on the morning of the fair, when the schedule is already printed.
• Counting judges, not judge visits, and discovering at noon that the panels are short.
• Giving the most experienced judges twice the load of everyone else, so they rush the last projects.
• Keeping the only copy of the schedule in one person's head or inbox.
• Asking for student names you do not need; a project code does the job.

What the judging tools cost

These are the public prices and free plans of the tools organisers mention most, read on the dates shown below.