I. Before the Exam: Designing for Curving
1. Define the Curve Type: Decide *before* the exam what type of curve you'll use. Common methods include:
* Percentage-based: A predetermined percentage of students receive specific grades (e.g., top 10% get A's, next 20% get B's, etc.). This is less dependent on the overall exam difficulty.
* Standard Deviation: Scores are curved based on the mean and standard deviation of the class. This is suitable when you expect a normal distribution of scores. A common approach is to assign grades based on standard deviation units from the mean (e.g., one standard deviation above the mean gets an A, etc.).
* Fixed Point Curve: You select a target average score (e.g., 75%) and adjust scores to achieve this average. Simpler to implement but less robust to variations in difficulty.
* Comparison to a Previous Year: Compare the overall performance to a similar exam from a previous year. This works best if the exams are very similar in content and difficulty.
2. Design a Comprehensive Exam: A well-designed exam itself minimizes the need for extreme curving. Ensure the exam:
* Covers the material comprehensively and fairly: Avoid overly difficult or misleading questions.
* Has appropriate difficulty: Aim for a level of difficulty that's challenging yet attainable for the majority of students who have studied adequately. Pilot testing a draft with a small group can be helpful.
* Has clear instructions and question wording: Ambiguity can lead to unfair scoring discrepancies.
* Uses a variety of question types: This can provide a more comprehensive assessment of student understanding.
3. Communicate the Curve (Optional): While not always necessary, informing students *before* the exam that a curve will be applied can help manage their expectations and reduce test anxiety. However, be explicit about the *type* of curve you plan to use.
II. After the Exam: Applying the Curve
1. Calculate Descriptive Statistics: After grading, calculate the mean, median, standard deviation, and range of scores. These statistics are crucial for applying most curve types.
2. Choose and Apply the Curve:
* Percentage-based: Rank the scores and assign letter grades based on the pre-determined percentages.
* Standard Deviation: Determine grade cutoffs based on standard deviations from the mean (e.g., mean + 1.5 SD = A, mean + 0.5 SD = B, mean - 0.5 SD = C, etc.). You may need to adjust these thresholds based on the distribution of scores.
* Fixed Point Curve: Adjust scores so that the average matches the target average. This involves adding a constant value to each score.
* Comparison to a Previous Year: Compare the distribution of scores with the previous year's distribution. Make adjustments to reflect the relative performance.
3. Review and Adjust (If Necessary): Before finalizing grades, review the results of the curve. Are the grade distributions reasonable? Does the curve fairly reflect student performance? Consider adjusting the curve if the results seem unfair or inconsistent.
4. Transparency and Communication: Clearly explain the method used to curve the exam and the rationale behind it. Provide students with their raw scores and curved scores.
III. Important Considerations:
* Ethical Implications: Curving can be helpful, but overuse can devalue the assessment's purpose. If a large portion of the class fails, consider if the exam or the teaching needs revision, rather than relying heavily on curving.
* Avoid Over-Curving: A curve should adjust for minor discrepancies, not drastically change the overall distribution. Excessive curving can inflate grades unrealistically.
* Document the Process: Maintain detailed records of the curving method, calculations, and rationale. This is important for transparency and accountability.
By carefully planning the exam and thoughtfully applying a chosen curving method, you can create a fairer and more effective assessment for your students. Remember that a well-designed exam that accurately assesses student learning minimizes the need for extensive curving.