Central Tendency: Mean, Median, and Mode
Why extreme tail values pull the Mean away from the typical patient while the Median remains anchored.
In this lesson you will learn
- 1Compare how the Mean, Median, and Mode behave under symmetric normal versus skewed distributions.
- 2Identify why right-skewed clinical variables (such as hs-CRP, ferritin, or hospital stay) require the Median.
- 3Recognise when bimodal distributions or trimmed means require subgroup stratification.
On this page▼
Distribution Skewness & Mean Drift
This section demonstrates how positive and negative skewness separate the Mean, Median, and Mode across clinical distributions.
Reporting Mean (SD) for a skewed biomarker in Table 1 misrepresents the typical patient and signals to reviewers that normality was not checked.
In clinical medicine, many inflammatory biomarkers (such as serum ferritin, high-sensitivity C-reactive protein, and troponin) and healthcare utilisation metrics (such as hospital length of stay or treatment cost) are strongly right-skewed.
Most patients cluster at low physiological values near the left boundary, while a small subset of acute patients extends far into the upper tail. Because the arithmetic mean sums every value in the numerator (∑x / n), those few high readings pull the mean upward away from the bulk of patients. The Median depends only on the middle rank order, remaining anchored to the typical patient.
Distribution Skewness & Mean Drift Lab
Drag the skewness slider to observe why skewed data corrupts the Mean and requires the Median
Watch how a single outlier drags the Mean significantly while the Median stays completely robust.
Mean, Median, and Mode are closely clustered. Parametric tests (t-test, ANOVA) are fully valid, and reporting Mean ± SD in Table 1 is mathematically sound.
Clinical Example
A surgical audit evaluates postoperative ward stay (days) across 50 appendicectomy patients; 47 recover in 2–4 days, while 3 develop complications staying 21, 28, and 35 days.
Interpretation: Reporting Median = 3.0 days [IQR 2.0–3.5] accurately informs clinicians that half of patients go home within 3 days, whereas Mean = 4.6 (SD = 6.1) is distorted by the 3 complicated cases.
Common Pitfall
Reporting only the arithmetic mean for a bimodal distribution (two distinct clinical peaks), where the average falls in the empty trough between subgroups.
Included with Any Analysez AI Study Package
This deep interactive chapter (Central Tendency: Mean, Median, and Mode) and its clinical simulation labs are unlocked for everyone who has bought any Analysez AI study package.