A non-parametric test is chosen instead of a parametric test mainly when the data are which of the following?
Free Applied Statistics in Healthcare practice question on Non-parametric tests and chi-square for BSN and Post-RN nursing students. Choose your answer, then open the answer and the explanation.
ASummarised by a mean and standard deviation that are close
BDrawn from a population with a very large sample size
CCollected from two independent groups of equal size
DMeasured on an ordinal scale or clearly not normally distributed
Show answer and explanation
✓ Correct answerD. Measured on an ordinal scale or clearly not normally distributed
Explanation
Non-parametric tests are distribution-free methods: they do not assume that the data follow a normal distribution and they work with ranks rather than raw values. They are therefore the standard choice when data are ordinal, when the sample is small and normality cannot be assumed, or when the data are markedly skewed with outliers. Parametric tests such as the t-test and ANOVA, by contrast, rest on the assumption of approximately normal distribution of the variable.
Why each option is right or wrong
✗A. Summarised by a mean and standard deviation that are close — A mean and standard deviation close together describes the spread of values, not the shape of the distribution that the choice of test depends on.
✗B. Drawn from a population with a very large sample size — A large sample size actually strengthens the case for a parametric test through the central limit theorem.
✗C. Collected from two independent groups of equal size — Equal group sizes are not a reason to abandon a parametric test; the t-test handles unequal groups as well.
✓D. Measured on an ordinal scale or clearly not normally distributed — Ordinal data and marked non-normality are exactly the situations where distribution-free, rank-based methods are required.
Feature
Parametric test
Non-parametric test
Distribution assumption
Approximately normal
Distribution-free
Data level
Interval or ratio
Ordinal or ranked
Typical example
Independent-samples t-test
Mann-Whitney U test
Sensitive to outliers
Yes
Less sensitive
EXAM HINT Ordinal or odd-shaped data → go non-parametric (ranks, not raw values).
اردو وضاحت
Non-parametric test کیا ہوتا ہے؟
یہ ایسے test ہیں جو data کے normal distribution ہونے کی شرط نہیں رکھتے۔ یہ values کی بجائے ranks پر کام کرتے ہیں، اس لیے جب data ordinal ہوں یا normal نہ ہوں تو یہی استعمال ہوتے ہیں۔
کس صورت میں non-parametric test چنیں؟
جب data ordinal scale پر ہوں یا clearly normal distribution نہ ہوں تو non-parametric test صحیح انتخاب ہے، کیونکہ یہ distribution کی شرط نہیں رکھتا۔ بڑا sample size ہو تو parametric test زیادہ طاقتور ہوتا ہے، اس لیے وہاں t-test یا ANOVA بہتر ہے۔
Key terms
ordinal scale — A ranking scale where order matters but the gap between values is not equal, like mild/moderate/severe pain. (ایسا پیمانہ جس میں ترتیب اہم ہو مگر وقفہ برابر نہ ہو، جیسے درد کا ہلکا، درمیانہ، شدید۔)