For a classroom nursing exam, a Cronbach's alpha (or KR-20) of 0.70 or higher is the usual target. Above 0.80 is strong. Between 0.60 and 0.69 is borderline — fine for a low-stakes quiz, uncomfortable for a course grade. Below 0.60, individual student scores carry enough noise that you should be cautious about decisions that depend on a point or two.
What alpha actually measures
Alpha is a consistency number, not a quality grade. It asks: do students who do well on one part of this exam also do well on the rest of it? If yes, the items are pulling in the same direction and alpha is high. If an exam mixes a handful of items that top students miss and weak students get right, alpha drops, because those items are measuring something other than what the rest of the exam measures — usually confusion.
Two things move alpha that have nothing to do with how good your items are:
- Length. More items, higher alpha, all else equal. A 25-item exam will struggle to reach 0.70 even when every item is sound.
- Spread. If every student scores between 36 and 42 out of 50, there is very little variance for alpha to work with. Exams in programs that admit a narrow, high-achieving cohort routinely run lower alphas than the same items would in a broader group.
So a low alpha on a short exam given to a tightly clustered class is not a scandal. It is a reason to read the item-level numbers before drawing conclusions.
What a low alpha looks like at the item level
On an exam running around 0.55, the pattern is usually the same:
- A few items with a negative point-biserial — the top-scoring students missed them more often than the bottom-scoring students. These are the first place to look, and they are usually a keying error, two defensible answers, or a stem that reads differently than you intended.
- A larger group of items with point-biserial under 0.20. These are not broken, but they are not separating students who know the material from students who don't.
- Several items with P-values above 0.90 and distractors nobody picked. Each one is a free point. Free points don't hurt students, but they add nothing to reliability.
Removing or fixing the negative items alone will often lift alpha several hundredths. The "alpha if item deleted" column in your export tells you exactly which items are pulling the number down.
What to do before the next exam
- Check the key on every item with a negative point-biserial. Rescore for the current exam if the key was wrong.
- Rewrite stems that split your strongest students — ask two colleagues to answer the item cold, without the key.
- Replace distractors that drew zero responses. A distractor nobody chooses is a threeoption item in disguise.
- Keep hard items that discriminate well. A P-value of 0.40 with a point-biserial of 0.35 is a good item, not a bad one.
- Don't chase alpha by deleting every weak item. Content coverage matters more than the number; fix items rather than shrinking the exam.
Where MapMyExam fits
An item-analysis export tells you the numbers. The Exam Health Report tells you, item by item, what the numbers most likely mean and what to change — reviewed by a nurse educator before it reaches you. Exam Health Reports start at $199 for individual educators; see pricing.