Our last articles
How to Automatically Award Partial Credit for a Partially Correct Answer in an Online Exam?
In an online exam, a partially correct answer can be scored automatically by awarding partial credit for correct elements, based on the scoring logic selected: all-or-nothing, even distribution of points, penalties for errors, or different weighting per answer.
This approach is useful when a strictly binary correction doesn’t accurately reflect a participant’s understanding. For example, someone might select some correct answers, add an incorrect choice, or identify a more difficult item that deserves greater weight in the scoring rubric.
To address these different assessment needs, Nexam now supports several automated partial credit strategies grouped under the concept of fractional weighting.
What Is Fractional Weighting?
Fractional weighting defines how a question’s points are distributed across its possible answers, rather than treating each question as a simple all-or-nothing outcome.
In practice, it enables more nuanced automated scoring for multiple choice questions, drop-down lists, association questions, and flexible content. This helps organizations align their scoring rubric with how they actually assess.
What Fractional Weighting Enables in Nexam
With fractional weighting, Nexam offers greater flexibility in how points are automatically awarded for a question.
Depending on the question type and scoring grid format selected, several correction strategies become available:
- Full success: award points only when the answer is entirely correct;
- Uniform distribution: distribute points equally across the expected correct answers;
- Uniform distribution with penalty: deduct points when an incorrect answer is selected;
- Weighting by answer: assign a different weight, positive or negative, to each answer.
The goal is to let the scoring rubric better reflect assessment intent. The same question can therefore be scored strictly, partially, with penalties, or in a more granular way, depending on what the organization actually wants to measure.
How Is the Score for a Partially Correct Answer Calculated?
To see how fractional weighting affects scoring, let’s use the same example across all four automated scoring strategies.
Multiple-answer question, 4 points, minimum 0 points
Which of the following animals are mammals?
A. Whale ✓
B. Bat ✓
C. Shark
D. Crocodile
Participant’s answer: A, B, and C
The participant found both correct answers, Whale and Bat, but also selected an incorrect answer, Shark. Depending on the strategy chosen, this same answer can lead to very different results.
Full Success
Full Success awards the maximum score only when the answer is entirely correct. Any partial, incorrect, or blank answer receives 0 points.
In our example: since the answer includes Shark, it isn’t entirely correct. The participant receives 0/4.
Uniform distribution
Uniform distribution awards an equal share of the maximum score for each correct answer selected. Incorrect answers are ignored, and the score never drops below the minimum.
In our example: Whale and Bat are each worth 2 points. The participant selected both. Shark has no effect. The participant receives 4/4.
Uniform distribution with penalty
Uniform distribution with penalty works like Uniform distribution, but each incorrect answer deducts an equivalent share of the score. The final result can never drop below the minimum score.
This strategy is especially useful when a question contains distractors, since it prevents a participant from selecting every answer without consequence.
In our example: Whale and Bat each award 2 points, but Shark deducts 2 points. The participant receives 2/4.
Weighting by answer
Weighting by answer is the most granular strategy. Each answer is assigned a percentage of the maximum score, defined directly on the answer.
The final score corresponds to the sum of the percentages of the selected answers. Correct answers should normally total 100%. An incorrect answer can also carry a negative weight, which reduces the score obtained without ever bringing it below the minimum.
In our example: if Bat is considered harder to identify than Whale, it could be worth 70%, versus 30% for Whale. If Shark deducts 50%, the participant obtains 30% + 70% – 50% = 50% of the maximum score, or 2/4.
| Key takeaway: The same answer can therefore receive different scores depending on the correction strategy chosen. This is precisely the value of fractional weighting: adapting automated scoring to assessment intent, rather than applying a single logic to every question. |
Which Scoring Strategy Should You Choose?
The choice of strategy depends primarily on the assessment objective. Here are a few guidelines for selecting the most suitable approach.
- Full success: when only an entirely correct answer should be rewarded. This strategy is well suited to questions where no error is acceptable.
- Uniform distribution: when every correct answer deserves recognition, without penalizing incorrect answers selected.
- Uniform distribution with penalty: when the question contains distractors and you want to discourage guessing.
- Weighting by answer: when some answers matter more than others, or when certain errors should have a greater impact on the result.
By offering several automated scoring strategies, Nexam lets you adapt scoring to your assessment objectives, rather than imposing a single way of correcting.
Which Strategies Apply to Which Question Types?
Not all weighting strategies apply to every question type. The table below shows the strategies supported in Nexam for each question format.
| Question Type | Full Success | Even Distribution | Even Distribution with Penalty | Weighted Answer |
|---|---|---|---|---|
| Single Choice | ✓ | |||
| True or False | ✓ | |||
| Multiple Choice | ✓ | ✓ | ✓ | ✓ |
| Dropdown | ✓ | ✓ | ✓ | ✓ |
| Association (darg-and-drop) | ✓ | ✓ | ✓ | ✓ |
| Flexible Content and Fill-in-the-Blank | ✓ | ✓ | ✓ | ✓ |
What If a Question Needs to Be Cancelled After an Exam?
Even with several automated scoring strategies, a question may sometimes need to be cancelled after an exam has been administered. A wording error, an ambiguity, or an issue discovered afterward can make a score adjustment necessary.
Nexam already allowed neutralizing a problematic question, for example by assigning it 0 points or awarding points for an answer deemed correct after the fact.
With the new Cancelled question format, every participant automatically receives a fixed score, regardless of their answer or lack thereof.
This approach simplifies the handling of exceptional situations while ensuring consistent treatment for all participants.
FAQ
Can a partially correct answer be scored automatically?
Yes. Depending on the correction strategy chosen, Nexam can award partial credit for correct answers, penalize certain errors, or apply a different weight to each answer.
Which scoring strategy should you choose?
It depends on the assessment objective. Full Success is suited to answers that must be entirely correct, while the other strategies allow for recognizing partial answers or applying penalties as needed.
Can a question be cancelled after an exam has been administered?
Yes. With the Cancelled Question format, Nexam automatically awards a fixed score to all participants, regardless of their answer or lack thereof.
Learn More
Fractional weighting offers greater flexibility to adapt automated scoring to your assessment objectives. Whether you want to recognize a partial answer, penalize certain errors, or assign a different weight to each answer, Nexam lets you choose the strategy best suited to your context.
For detailed configuration of each strategy and the compatible grid formats, see the article Add a Marking Grid and Correction Guide to a Question in the Nexam Help Center (login required).
Want to see these new options in action or discuss your digital assessment needs? Schedule a demo with the Solution Nexam team →