Grading and evaluation

Exam grading & evaluation for accurate, consistent results

Apply the right grading method to every question, automate scoring, standardize evaluation, and generate consistent results across all response types.

Book a demo Try for free
Multiple Choice
Which city is the capital of France?
A. London
B. Paris
C. Berlin
D. Madrid
Scoring configuration
Correct answer +100%
Wrong answer -25%
Unanswered 0%
Scored instantly
100 / 100 pts
Automatic grading

Get instant results for objective questions

Multiple choice, matching, true/false and other objective questions are scored instantly based on predefined rules. No waiting, no manual work.

Fine-grained scoring rules for every scenario

Each answer option can be assigned a percentage value between 100% and +100%, supporting both positive and negative scoring. Partial credit can be enabled so test-takers receive proportional credit for partially correct selections.

  • Assign percentage values between 100% and +100% per answer option
  • Apply negative multipliers to penalize incorrect answers and discourage guessing
  • Enable partial credit for multi-choice questions
  • Set minimum score thresholds so low-quality partial answers receive 0 points
  • Restrict the number of answer choices a test-taker can select
testinvite.com / app / question-editor / scoring
Multiple Choice

Which of the following best describes a primary key in a relational database?

Answer option Score % Status
Uniquely identifies each row
+100%
Correct
Allows NULL values
0%
-
Used for full-text search
-25%
Penalty
References another table
-25%
Penalty
Partial credit
Enabled
Min threshold
0%
Max selections
Unlimited
Rubric-based grading

Standardize open-ended evaluation with rubrics

Open-ended responses are evaluated against rubrics that define clear scoring standards. Each criterion carries a weight that determines its contribution to the final score.

Multiple rubric scoring approaches in one editor

Weighted Criteria

Each criterion row can carry a different weight, determining its contribution to the final score.

testinvite.com / app / rubric-editor
Rubric Editor
Marketing Strategy Essay - 10 pts
Preview
Save rubric
Criterion
Wt.
Exemplary
100%
Good
75%
Adequate
50%
Poor
0%
Content relevance
40%
Analysis depth
35%
Language quality
25%
Evaluator feedback
The analysis demonstrates strong understanding of market segmentation, however the competitive analysis section could benefit from...

Scoring Approaches

Assign fixed percentage values per level: Exemplary = 100%, Good = 75%. Or define custom intervals and allow manual input during evaluation.

Evaluator Feedback

Rubric structure also supports structured evaluator feedback. Evaluators can provide detailed written notes while scoring each response.

AI Rubric Generation

Generate structured scoring rubrics with AI

Generate evaluation rubrics instantly from assessment questions. AI proposes structured scoring criteria and performance levels that you can review, refine, and use.

Criteria
100%
75%
50%
25%
Text Comprehension
Textual Evidence
Argument Structure
Language Use
Generate rubric
Prompt
Describe the rubric you want to create, including the task, criteria, and performance levels.
CANCEL
GENERATE RUBRIC

Generate rubrics from assessment questions

AI analyzes the assessment question and proposes a structured rubric with evaluation criteria. Optionally, you can provide additional instructions to guide the generated rubric.

Support flexible scoring models

Generated rubrics can be used with both Level Selection and Score Entry scoring methods, supporting different evaluation workflows.

Preserve existing rubric settings

When regenerating an existing rubric, AI refreshes the evaluation criteria while preserving your current scoring configuration and rubric settings.

Review and customize before saving

Review and edit generated criteria, performance levels, descriptions, and scores before saving to ensure the rubric matches your evaluation standards.

AI-assisted grading

Scale open-ended evaluation with AI

Essays, long answers, spoken responses, video interviews, and coding tasks are evaluated using AI systems powered by large language models guided by your own scoring instructions.

Author instructions guide the AI, not just the response

While creating a question, authors provide evaluation instructions and scoring criteria. The AI then combines the participant's response with the question and those guidelines to generate a score. Different LLMs can be used depending on the question type.

  • Written responses
  • Audio responses (automatically transcribed and evaluated)
  • Video responses (automatically transcribed and evaluated)
  • Coding submissions
  • Flexibility to integrate with different AI models depending on assessment needs
Learn more about AI grading
app.testinvite.com/app/ai-grading
Written
Audio
Video
Coding
Audio response - Auto-transcribed

Describe a situation where you had to manage conflicting priorities. How did you handle it?

02:14
Transcription

"In my previous role I was simultaneously managing a product launch and a critical client escalation. I assessed impact first - the client issue had immediate revenue risk, so I delegated parts of the launch while staying across both threads until resolution."

AI Evaluation Instructions

"Evaluate based on: clarity of communication (40%), logical structure (35%), use of concrete examples (25%). Deduct if response lacks specifics."

Model: GPT-4o | Language: English
AI Score breakdown 79%
Clarity of communication
82%
Logical structure
71%
Use of concrete examples
88%
AI Feedback

"Strong situational awareness and concrete decision-making examples. Logical flow could be structured more clearly around the resolution. Communication is clear and confident throughout."

Graded automatically
Confidence: High - Flag for review
Rule-based & function-based grading

Automate grading for open-form answers without AI

Short answers and numeric inputs can be graded automatically using text matching rules or custom JavaScript functions, without requiring AI evaluation.

Text & Numeric

Rule-based grading

Responses are evaluated using predefined text rules: exact string matching, regular expressions, or semantic equivalents. Multiple acceptable answers and paraphrased variations can be defined so the system recognizes different ways of expressing the same idea.

  • Exact string matching
  • Regular expression matching
  • Semantic equivalents and paraphrased variations
  • Custom feedback triggered automatically when rules are not met
Matching rules
Exact "Paris" +100%
Regex /^paris$/i +100%
Equiv. "capital of France" +100%
Custom logic

Function-based grading

Responses can be evaluated using custom JavaScript functions written in a built-in code editor. Each function processes the test-taker's response and returns a score based on the defined logic.

  • Full credit, no credit, or negative scoring
  • Exact matching, partial matching, regex checks, character-level analysis
  • Timing-based adjustments and penalties for errors
  • Different evaluation functions for different question types or scenarios
grading-function.js
function grade(response) {
const val = parseFloat(response);
if (val === 42) return 100;
if (Math.abs(val - 42) <= 1) return 50;
return 0;
}
Manual grading

Apply human judgment where it matters most

Essays, written explanations, spoken answers, and video submissions can be reviewed and graded manually. Evaluators provide written feedback and evaluation notes while scoring. Manual grading is useful when subjective judgment, detailed interpretation, or nuanced feedback is required during the evaluation process.

Evaluator feedback

Evaluators can write detailed reviews to share with the test-taker, providing clear feedback on performance, strengths, and areas for improvement.

Multiple evaluators

Multiple evaluators can review the same response when needed, helping maintain fairness and reliability in subjective assessments.

testinvite.com / app / manual-grading
SK
Sarah Kim
Leadership Assessment, Essay Q3
Pending review
Candidate response

In my previous role, when our team faced conflicting deadlines between two major projects, I organized a priority matrix workshop with stakeholders to align on business impact. By establishing clear criteria and transparent communication, we successfully negotiated revised timelines...

Evaluation
Leadership qualities
80%
Communication clarity
85%
Total score 82.5%
Feedback to candidate
Your response demonstrates strong leadership initiative. Consider elaborating on how you measured the outcome of the decision and what you would do differently...
E1
E2
2 evaluators
Submit evaluation
Scoring model

A straightforward scoring methodology that stays flexible

The test score is calculated by dividing the sum of points earned by the total points of the test. Each question awards the test-taker a percentage score, which is then multiplied by the question's point value.

Test Score

Points earned ÷ Total points

Points earned
7
correct answers
÷
Total points
10
all questions
=
Test score
70%
final result

A test with 10 questions worth 1 point each, answered 7 correctly, produces a score of 70%.

Partial scoring

% score × question points

% score
50%
×
Q points
1 pt
=
earned
0.5

A 50% answer on a 1-point question earns 0.5 points. A 25% answer on a 4-point question earns 1 point.

Negative scoring

% score × negative multiplier

% score
−25%
×
multiplier
×1
=
penalty
−0.25

Incorrect answers can carry a negative multiplier to apply penalty points. Default is 0 — no penalty unless configured.

Questions marked as ineffective or ignored are excluded from both the total points of the test and the points earned.

Reporting outputs

Turn grading results into decision-ready reports

Once responses are scored, results can be presented through structured scorecards, dimension-level breakdowns, and branded PDF reports. This helps teams review outcomes, compare performance, and share consistent results with candidates, employees, or stakeholders.

  • Overall, section, question, and dimension-level scores
  • Branded PDF reports
  • Result visibility controls for admins and test-takers
Learn more about reporting
app.testinvite.com/app/results/scorecard
82%
PASSED
Overall score - Leadership Assessment
Section A
88%
Section B
74%
Dimension scores
Communication
79%
Leadership
86%
Problem solving
71%
Leadership_Assessment_Report.pdf
Download
Who it's for

Built for any scenario where evaluation must be accurate, consistent, and scalable

From hiring to certification, apply the right grading method for each assessment, automate where possible, standardize where needed, and rely on human judgment when it matters.

Recruitment and employee training

Evaluate employees and candidates with consistent scoring

Use AI, automated scoring, rubrics, and human review to evaluate candidates, employees, and trainees with the same level of consistency. Support hiring decisions, training validation, and skill-gap analysis with reliable results.

Certification and licensing

Ensure defensible and standardized scoring

Apply strict evaluation criteria to maintain fairness and consistency in high-stakes exams. Combine automated scoring with human review where needed, and ensure every result can be justified and audited.

Education and academic assessment

Evaluate at scale without losing grading quality

Grade essays, projects, and exams across large student groups using a mix of automated and manual methods. Maintain consistency between evaluators while reducing grading time.

English language assessment

Evaluate language skills with structured and scalable scoring

Assess writing, speaking, listening, and reading using a combination of AI evaluation, predefined scoring criteria, and human review where needed. Ensure consistent scoring across candidates while scaling language assessments efficiently.

Get started

Evaluation you can trust at scale

Apply consistent scoring across every candidate and scenario.

Book a demo Try for free