Nursing Education and Distractor Evaluation for Reliable and Valid Test, Assigning Grades to Evaluation Test

Distractor Evaluation in Test for Outcomes Evaluation, Revising Test Items for a Reliable and Valid Test, Assigning Grades and Principles of Test Evaluation, Absolute Scale for Test Grading, Relative Scale for Test Grading, Grading Standards, Grade Inflation, and Grade Indexing of Evaluation Test in Nursing Education.

Distractor Evaluation in Test for Outcomes Evaluation in Nursing Education

In addition to the evaluation of the correct answer to an item for a positive point biserial correlation, each distractor should be individually evaluated. Effective distractors should appeal to the non-learner, as indicated by negative point biserial correlation values.

Distractors with a point biserial correlation of zero indicate that students did not select them and that they may need to be revised or replaced with a more plausible option to appeal to students who do not understand the content. Distractors that were not selected increase the chances that a student could have obtained the correct answer by guessing.

One way to develop appealing distractors is to periodically ask open ended questions to determine the most common errors in thinking. Distractors for questions with numerical answers may need to be worked out by following the most typical mistakes that students make.

Revising Test Items for a Reliable and Valid Test in Nursing Education

Developing a valid and reliable test is an ongoing process. It is helpful to revise the test immediately after administering it while faculty can recall items and student responses to them. Item revision should be conducted after item analysis. One way to analyze items for revision is to use a test item analysis form.

The result of item analysis for each question is entered in the form. Those items falling outside the “ideal” range should be considered for revision. Items to be revised should include those with the following statistical characteristics:

1. Items with P values that are too high or too low (around 0.5 is ideal)
2. Correct answers with low positive or negative point biserial values (> 0.30 is ideal)
3. Distractors with highly positive point biserial values (negative values are ideal)

Other considerations for revising test items include those questions that were scored incorrectly, items not written clearly or noted by students who did not understand the question because of linguistic or cultural bias or were not matched with the teaching learning process of having clear learning outcomes, clarity of instruction, and opportunity for practice with feedback.

Assigning Grades and Principles of Test Evaluation in Nursing Education

Grades provide both feedback and motivation for students. In the academic setting, assignment of grades may be guided by the institution grading policy or scale. Many computer software programs are available to assist faculty with assigning grades accurately and efficiently. The two basic methods for assignment are the absolute and the relative (“curved”) scales (described in the following sections). Principles of good grading include the following (McDonald, 2013):

1. Informing students of the specific grading criteria at the beginning of the course (stated explicitly in the syllabus)
2. Basing the grades on learning outcomes (not factors such as attendance or effort)
3. Gathering sufficient data (amount and variety) for the assignment of a valid grade
4. Recording data collected for grading purposes quantitatively (e.g., 89%, not B +)
5. Applying the grading system equitably to all students
6. Keeping grades confidential
7. Using statistically sound principles for assigning grades

Absolute Scale for Test Grading in Nursing Education

An absolute grading scale rates performance relative to a standard (McDonald, 2013). The student’s earned points are compared with the total possible points and are expressed as a percentage earned. The standard should be included in the syllabus at the beginning of the course.

Theoretically, all students could receive an A (or an F) with this scale. In reality, the dispersion of the grades depends on the difficulty of the tests. See Table 24-8 for an example of an absolute grading scale.

Relative Scale for Test Grading in Nursing Education

A relative grading scale rates students according to their ranking within the group. To assign grades in this system, faculty record scores in order, from high to low. Grades may then be assigned using a variety of techniques. One method is to assign the grades according to natural “breaks” in the distribution. This method has the disadvantage of being subjective. A better method of assigning grades based on a relative scale is to use the test statistics to create a “curve”:

1. Decide whether to use the mean or the median as the best measure of central tendency. If the mean and median are approximately the same, use the mean. If the distribution is skewed, use the median.
2. Determine the SD. The C grade will be set as the mean plus or minus one half of the SD (encompassing 40% of the scores).

Relative grading scales may also be developed by using linear scores such as z scores or t scores (see the section on standardized test scores for directions for calculating these scores). The scores are more commonly used for grading purposes because there are no negative values in this system. The mean score becomes a score of 50. The z score and t score are figured for each raw score.

The faculty member decides what grade to assign to the ranges of t scores. Assuming a normal curve, a score of 50 would be assigned a grade of C. Computer grading programs that calculate grades according to absolute or relative scales are available. Many experts in assessment and grading do not recommend the use of relative grading scales (Haladyna, 2004; McDonald, 2013).

Grading Standards, Grade Inflation, and Grade Indexing of Evaluation Test in Nursing Education

Periodically, faculty, administrators, boards of trustees, or consumers raise questions about the relative meaning of grades and potential or actual grade “inflation.” Some possible causes for grade inflation are improved academic readiness of admitted students; student retention programs; competency-based assessments; competitive admission standards; a student population of older, mature, and career-directed students; and pass–fail grading systems.

Grade indexing involves indicating how many students in each course or section of a course received grades that equaled or exceeded the grade of a given student. This index may appear on the student’s transcript. Nursing faculty should review grading policies
and practices on a regular basis.

A consistent philosophy about grading and fair and equitable grading practices communicates concern to the students and competence to nursing’s varied publics. Faculty must also continually strive to administer valid and reliable tests that measure students’ attainment of course and program competencies.

Although developing, administering, and analyzing classroom tests may seem like a monumental task, the step-by-step approach presented in this Topic can be used as a guide to simplify this process. By following these guidelines, faculty can create written tests that can be used as effective measures of outcomes in the classroom. Assigning grades is the last step in this process.