Which assumption is shared by parametric tests comparing group means?
Free Applied Statistics in Healthcare practice question on Parametric tests for comparing groups for BSN and Post-RN nursing students. Choose your answer, then open the answer and the explanation.
AThe dependent variable is measured on an interval or ratio scale
BThe dependent variable is measured on a nominal or ordinal scale
CThe sample size in each group is below thirty participants
DThe observations within each group are ranked before analysis
Show answer and explanation
✓ Correct answerA. The dependent variable is measured on an interval or ratio scale
Explanation
Parametric tests such as the t-test and ANOVA are built on a mathematical model of a normal curve, so they need data that behave like measurements on a continuous scale. This is why the dependent variable should be interval or ratio, the observations should be independent, and the scores in each group should be roughly normally distributed with similar spread. Non-parametric tests were developed for the situations where these conditions are not met.
Why each option is right or wrong
✓A. The dependent variable is measured on an interval or ratio scale — Parametric tests model continuous measurements, so an interval or ratio dependent variable is a core requirement of the method.
✗B. The dependent variable is measured on a nominal or ordinal scale — Nominal and ordinal data lack the equal spacing that the normal-curve model assumes, so they are handled by non-parametric methods instead.
✗C. The sample size in each group is below thirty participants — Small groups are not a violation; parametric tests can be used with modest samples when the data are reasonably normal.
✗D. The observations within each group are ranked before analysis — Ranking the observations is what non-parametric tests do, and a parametric test works on the raw scores rather than on ranks.
Requirement
Parametric test
Non-parametric counterpart
Scale of dependent variable
Interval or ratio
Nominal or ordinal
Distribution shape
Approximately normal
Any distribution
Variance across groups
Similar (homogeneity)
Not required
Typical example
Independent-samples t-test
Mann-Whitney U test
اردو وضاحت
پیرامیٹرک ٹیسٹ کیا ہوتا ہے؟
پیرامیٹرک ٹیسٹ ایسے ٹیسٹ ہوتے ہیں جو (normal curve) کے ریاضی کے ماڈل پر بنائے جاتے ہیں، جیسے (t-test) اور (ANOVA)۔ ان کو چلانے کے لیے ڈیٹا کا کچھ خاص قسم کا ہونا ضروری ہوتا ہے، ورنہ نتیجہ غلط آ سکتا ہے۔ سادہ الفاظ میں یہ ٹیسٹ فرض کرتے ہیں کہ ہمارا ڈیٹا ایک مسلسل پیمانے پر ناپا گیا ہے۔
ڈیٹا کس پیمانے پر ناپا گیا ہونا چاہیے؟
صحیح بات یہ ہے کہ (dependent variable) کو (interval) یا (ratio) پیمانے پر ناپا گیا ہو، کیونکہ یہ ٹیسٹ اسی مسلسل پیمانے والے ڈیٹا پر کام کرتے ہیں۔ اس کے ساتھ مشاہدات آزاد ہوں اور ہر گروپ کے سکور تقریباً (normal distribution) رکھتے ہوں۔ جو ٹیسٹ ان شرائط کے بغیر چلائے جاتے ہیں وہ (non-parametric) کہلاتے ہیں۔
Key terms
interval or ratio scale — A measurement scale with real, equally spaced numbers, such as weight in kg or temperature, where arithmetic makes sense. (Aisa paimana jismein asli aur barabar faasle wale numbers hon, jaise (weight) kilogram mein ya (blood pressure), jahan jama aur tafreeq ka matlab banta hai.)