Scored Assessment Template
A weighted scoring scaffold where the hard questions count double. Shown on spreadsheet skills, it adapts to any subject you grade.
Six questions, nine points: the tougher items are worth two. Answer all six and your weighted total, out of nine, decides the verdict.
Flat scoring pretends all knowledge weighs the same: knowing what CSV stands for counts as much as knowing how a lookup works. Real subjects are not flat, since some questions certify a skill while others only decorate the total. This template is the weighted answer: six questions worth nine points, where the load-bearing items count double, demonstrated on spreadsheet skills and built to be re-skinned onto whatever you assess.
How the weighting is wired. Every scoring rule carries its own point value: the add_var action on the SUM, VLOOKUP, and absolute-reference questions adds 2 to the score variable, while the vocabulary-grade items add 1. Three range rules read the nine-point total: 8 or 9 grades Advanced, 5 to 7 Working Knowledge, 4 or under Foundations. Each question states its weight in the description, which does more than keep things fair. Telling people what counts double changes how they think about the subject before they answer.
Why weighting, and what stays out. Weighting encodes your judgment about mastery. A candidate who answers both one-pointers but misses every two-pointer scores 3 of 9; a flat quiz would have called that 50 percent and hidden the thing you needed to see. In this spreadsheet demo the two-point items are the concepts people fake through interviews (functions, lookups, reference anchoring), and the one-pointers are the terms anyone absorbs in a week. Draw the same line in your subject and the score starts predicting performance. A few mechanics stay out. Negative marking makes respondents obsess over the penalty and leaves the endings' arithmetic fragile. Partial credit goes too: the engine supports any point value per rule, but "half-right" usually signals a badly keyed question. So does stretching past roughly ten items, since a weighted six-question core measures cleaner than a padded twenty.
Who assesses with it, and rescaling the weights. Trainers check whether the workshop stuck, hiring screens rank applicants beyond pass and fail, course creators gate module two behind module one's fundamentals, and managers baseline a team before deciding what training to buy. To rescale it, rewrite the six questions for your domain, then set each rule's point value in the Logic panel to match how much the skill matters; the pattern supports any weights, well beyond 1 and 2. Recompute your maximum, reset the three tier boundaries, and update the "worth N points" descriptions so the weighting stays visible. The name field pipes into every verdict, which makes results feel issued rather than generated, and the CSV export carries the weighted totals for your records. Assessing at volume? Skip the notification noise and add a webhook instead: every submission posts to your endpoint signed and with retries, so the results register stays complete even on the day your receiving tool has a bad one.
Frequently asked questions
How do I change how many points a question is worth?
Open the Logic panel and edit that question's rule; the add-to-score action takes any number. Update the question description and the tier thresholds to match your new maximum.
Do respondents know the weighting up front?
Each question states its value in the description. Visible weighting reads as fair and tells respondents which concepts matter most.
Can the result screen address people by name?
It already does: the name answer is piped into the ending copy alongside the weighted score, so every verdict reads like it was written for that person.
Where do the weighted totals end up?
On every response: the responses view shows the computed score with the full answer set, and the CSV export includes it as a sortable column for ranking or records.