1.3.5 · D5Basic Data & Probability
Question bank — Range — definition and calculation
This page is a trap course. Each item below looks simple, but hides a misconception about the range — the gap between the biggest and smallest value in a dataset. Read the prompt, cover the answer, decide out loud why, then reveal.
Before we start, one word we lean on constantly: dispersion means how stretched out the numbers are. Range is the crudest way to measure it — it looks only at the two endpoints and ignores everything in between.
True or false — justify
All values are identical, so range is undefined
False. If every value equals the same number , then and range . Zero is a perfectly valid answer meaning "no spread at all", not "undefined".
Range can be negative if the data has negative numbers
False. Range is always , and by definition, so the result is regardless of whether the individual values are negative.
Two datasets with the same range must have the same shape
False. and both have range , yet one clumps low with a lone high value and the other spreads evenly. Range describes only the endpoints, never the interior.
Adding a value in the middle of the data can change the range
False. A value strictly between the current min and max leaves both extremes untouched, so range is unchanged. Only a new value beyond an existing extreme moves it.
If you add to every data point, the range increases by
False. Both max and min rise by , so their difference is unchanged. Shifting all data slides it along the line without stretching it.
If you multiply every data point by , the range triples
True. Both endpoints scale by , so . Scaling stretches the whole line, endpoints included.
Range and Interquartile Range always give the same value
False. Range uses the absolute extremes; IQR uses the middle 50% and deliberately throws the extremes away, so IQR is usually smaller and far less shaken by Outliers.
A dataset with more data points always has a larger range
False. Range depends on the two extremes, not the count. A million values between and has range ; three values has range .
Range has no units because it is a "difference"
False. Subtracting two quantities in the same units (say marks marks) leaves those units intact. Range of test scores is "44 marks", not a bare "44".
Spot the error
A student writes "Range "
The subtraction order is flipped. Range is defined as precisely so the answer stays non-negative; here it should be .
For a student writes "Range "
They dropped the negative sign on the minimum. The min is , so range . Subtracting a negative adds.
A student reports "the range of the class heights is 30" with no unit
Missing units make the number meaningless — what? cm? students? It must read " cm" so the reader knows it is a length.
For a student states flatly "Range " and moves on
The computation is right, but is almost certainly an outlier. An honest report flags it: "Range , but dominated by one extreme; without it the range is ."
A student claims "the range tells us the average distance between data points"
No — range only measures the total span from smallest to largest. Average pairwise spacing is a different idea; measures like Standard Deviation and Variance capture typical spread using every point.
A student says "range must be a whole number"
False assumption. Range inherits the type of the data. For the range is ; decimals and fractions are fine.
Why questions
Why do we subtract instead of add the two extremes?
Adding gives a meaningless total magnitude; subtracting gives the gap between smallest and largest, which is what "spread" means. is a span, is nothing useful.
Why is range called the "simplest" measure of spread?
It uses only two numbers — the max and the min — and ignores every value in between, so it needs no averaging, squaring, or sorting of the interior.
Why is range so sensitive to Outliers?
It is defined entirely by the two most extreme points, so a single stray value that becomes the new max or min directly sets the answer, no matter how atypical it is.
Why might a large range not mean the data is "usually" spread out?
Range reflects only the endpoints. Most points could be tightly clustered with one lone extreme inflating the range, so it can overstate the typical spread — that's why Interquartile Range exists.
Why does range keep the same units as the data?
Because it is a difference of two quantities measured in those units; subtracting "marks minus marks" leaves "marks". Contrast this with Variance, which squares differences and so carries squared units.
Why does shifting all data by a constant leave the range unchanged?
A shift moves both extremes by the same amount, so their difference is untouched. Range measures stretch, not position, and a shift only changes position.
Edge cases
What is the range of a dataset with exactly one value, e.g. ?
The single value is both the max and the min, so range . A lone point has no spread.
Can the range equal one of the data values by coincidence?
Yes. For the range is , which happens to equal the max. Nothing forbids the difference from matching a value present in the data.
What happens to the range if the maximum and minimum are the only two values, repeated many times, e.g. ?
Repetition of the extremes changes nothing; range still equals . Range cares about which values are extreme, not how often they occur.
Is the range of zero or ?
Zero. All three values are equal, so and range . The values being negative does not make the range negative.
If every value is the same except one huge value, is the range trustworthy?
The formula is correct but the range is misleading — it is driven entirely by the lone extreme. You'd report it alongside a note that a more robust measure like Interquartile Range is better here.
How does range appear in a box plot from Data Visualization?
It is the total span from the end of the lower whisker (min) to the end of the upper whisker (max) — the full width of the plot — while the box itself shows the more robust IQR.
Does range complement the Measures of Central Tendency or replace them?
It complements them. Mean or median tells you where the data sits; range tells you how far it stretches. You need both to describe a dataset, since two sets can share a mean yet have wildly different ranges.