STEM PRACTICE AND EVIDENCE WORKBENCH
Compare a planned question set with easy, medium, and hard percentage targets while preserving essential-item evidence.
Review output
Load the worked example to inspect realistic inputs, or enter anonymous classroom or laboratory evidence. Results remain in this browser tab until you copy or download them.
Transparent method
For each band, actual share = number of questions carrying that label divided by all entered questions times 100. Gap = actual share minus target share in percentage points. A band is within tolerance when the absolute gap does not exceed the declared limit. All three target percentages must total 100 before the plan is calculated. Targets within tolerance means the entered labels match the planning distribution, not that learners will experience those difficulties or that the test is valid. Difficulty is population dependent and should be examined with pilot or operational item evidence. Review standards coverage, cognitive demand, language and accessibility, item dependence, scoring, time, bias, security, consequences, and accommodations before use.
Confirm the learning question, authoritative source, units, notation, response coding, model assumptions, and review threshold. Keep an unchanged copy of the original evidence and remove learner names or identifiers.
Run the worked example first, inspect every intermediate row, and compare it with a hand calculation, diagram, representation, or approved subject method. Correct input evidence before interpreting a flag.
Record the method, assumptions, units, classifications, and flagged rows. Verify the editable report independently, connect the numerical signal with learner reasoning, and document any teaching or laboratory decision outside this tool.
How to interpret the result
Use during assessment blueprint review to see whether planned item counts approximate an intended challenge distribution. The second table preserves which items were marked essential so a balancing change does not silently remove a core objective. It does not estimate empirical item difficulty, measure cognitive demand, validate content coverage, recommend a universal mix, or make a high-stakes assessment fair.
The worked plan contains ten questions: three easy, five medium, and two hard. Those counts convert exactly to targets of 30, 50, and 20 percent, so all three bands fall within five percentage points. Three questions are marked essential. Changing an item label would immediately show the resulting percentage-point gap while retaining its essential flag.
For each band, actual share = number of questions carrying that label divided by all entered questions times 100. Gap = actual share minus target share in percentage points. A band is within tolerance when the absolute gap does not exceed the declared limit. All three target percentages must total 100 before the plan is calculated.
Targets within tolerance means the entered labels match the planning distribution, not that learners will experience those difficulties or that the test is valid. Difficulty is population dependent and should be examined with pilot or operational item evidence. Review standards coverage, cognitive demand, language and accessibility, item dependence, scoring, time, bias, security, consequences, and accommodations before use.
Do not confuse target percent with number of marks, label items only from surface complexity, use tolerance to hide a missing band, change an essential question solely to satisfy a ratio, assume hard means higher cognitive demand, or treat teacher predictions as observed facility. Keep blueprint purpose, evidence source, and later item analysis clearly separated.
Learning boundary
This browser-local workbench applies only the disclosed rules. It cannot diagnose a learner, certify an experiment, establish assessment validity, or replace teacher, laboratory, curriculum, accessibility, and safety judgment.
Review checklist
Practical questions
No. Processing occurs in the current browser tab. Avoid personal information and retain authoritative records only in the approved school system.
No. It describes entered evidence under one disclosed rule. Review reasoning, uncertainty, access, and several evidence sources.
Treat it as an editable calculation aid. Show your own reasoning, cite required sources, and follow course and academic-integrity rules.
YOUR NIRMION
