The standard method of constructing multiple-choice questions is to write each item with one clearly correct answer and two to four plausible yet incorrect options. The standard method of scoring such tests consists of assigning one point for the correct answer and no points for the incorrect options. This dichotomous construction and scoring model assumes that examinees selecting incorrect options form a fairly homogeneous group. It fails to identify students with misinformation as well as those who have developed partial information about the item's knowledge domain but have not yet mastered the subject matter.

This study examined the feasibility and validity of creating and scoring multiple-choice items containing answer options with various degrees of "rightness" and "wrongness" in a test of drug administration for student nurses. The rationale behind developing this concept has been a desire to improve test efficiency and effectiveness. If more information than the binary one/zero score can be achieved, efficiency is improved. If more complex information about the knowledge domain can be evaluated, more effectiveness in measurement can be demonstrated as well.

One hundred and fifty-one senior and 71 freshman students in Associate Degree Nursing programs were administered a 45 item test. Students were instructed to select the best or "safest" option and the worst or the most "unsafe" option in each question. Partial information was clearly and significantly demonstrated in the total group scores. Seniors, more competent students and those with previous nursing experience had significantly higher scores on the "unsafe" or most wrong answers. Experienced and highly competent students also demonstrated significantly higher scores on the most right or "safest" options. These findings demonstrated the validity of this method of test item construction.