Not sure which lesson to do first? Find your level and get a plan →

Math · Statistics & Probability · Grade 6-8

Comparative Statistics

Comparative statistics uses statistical measures — like means and spreads — to compare two or more groups or data sets and judge whether they really differ.


Quick Answer

Comparative statistics uses statistical measures — like means and spreads — to compare two or more groups or data sets and judge whether they really differ. Use it whenever a question asks whether one group is bigger, better, faster or more consistent than another: two teaching methods, two battery brands, two classes' test scores. Compare a center (mean or median) and a spread (range or standard deviation) for each group, and switch to rates when the groups differ in size.

Why This Matters

Most real questions are comparisons: does the new medicine work better, is this route faster, did the change help? Looking only at averages can call two very different groups "the same," and treating every small gap as real can crown a winner that would lose on a rerun. Putting center and spread side by side is what makes a claim like "A beats B" believable.

Intuitive Explanation

Two basketball teams each have an average height of 180 cm. On Team A every player is between 175 and 185 cm; on Team B the players range from 160 to 200 cm. The averages match, but a coach would treat these teams very differently, because B has both the tallest and the shortest players. Comparing groups means comparing at least two things for each: where the values sit (center) and how far apart they are (spread).

Next, judge the size of a gap against the spread. Suppose Class A averages 78 on a test and Class B averages 74. If scores in each class run only from 70 to 85, a 4-point gap is a clear difference. If scores run from 40 to 100, a 4-point gap is small next to how much students vary within each class, and it could vanish on the next test. A gap between centers means more when the spreads are small.

Describing one group is not comparative statistics. "Class A's mean is 78" summarizes one group; it becomes a comparison only when another group's center and spread stand beside it. Correlation is a different tool again: it asks whether two measurements, like hours of sleep and test score, rise together for the same people, not whether two separate groups differ.

Think of it like this. Compare two archers. One archer's arrows cluster tightly, a little left of the bullseye; the other's land all over the target but average out to the center. Their average positions tell only part of the story; the spread of their arrows tells you whom to trust with a single shot.

Core idea. Comparative statistics uses summary measures to ask whether two or more groups truly differ, and by how much.


When to Use

Use comparative statistics when a question names two or more groups and asks which is higher, better, more consistent, or whether they differ. Compute the same center and spread for every group and compare them side by side; use proportions or rates if the groups differ in size. If there is only one group, you are describing it. If you have two measurements on the same individuals, use a scatter plot and correlation instead.

Tip. Ask: Am I weighing two or more groups against each other rather than describing a single group?

How to Recognize It

Count how many separate groups the data describe, and make sure you will measure the same center and the same spread for each one.

  1. Are two or more separate groups measured on the same thing, like the battery life of Brand X and Brand Y?

    Yes means compare a center and a spread for each group. Only one group means you are describing it. Two different measurements on the same people, like height and shoe size, call for correlation.

  2. Do the groups have similar centers but very different spreads?

    Yes means the means alone would hide the real difference: two classes can both average 75 while one has scores from 70 to 80 and the other from 40 to 100. Report the range, interquartile range or standard deviation beside the center.

  3. Are the groups very different in size?

    Yes means compare rates or percents, not totals: 30 wins in 40 games beats 45 wins in 80 games. Groups of equal size can be compared directly.

  4. Is the gap between the groups small compared with how much values vary inside each group?

    Yes means the difference may be just noise, and a second sample could reverse it. Deciding whether a gap is real is the work of signal vs noise, the step after you have compared the groups.

Comparative Statistics compared with Mean (descriptive), Signal vs noise, and Correlation

Every row summarizes data, but they differ in what is being related. Ask how many groups and how many variables there are: one group described (mean), separate groups side by side (comparative statistics), whether their gap is real (signal vs noise), or two variables on the same individuals (correlation).

Comparative Statistics

Meaning
Use this when the question asks whether two or more groups differ and by how much, not to describe one group alone.
Key test
Am I weighing two or more groups against each other rather than describing a single group?
Example
Method A scores average 80 (range 75–85); Method B averages 80 (range 50–110). Which method is more reliable?

Mean (descriptive)

Meaning
Summarizes the center of ONE group, with no comparison.
Key test
Use when describing a single data set, not contrasting two.
Formula
xˉ=∑xn
Example
Average of one class's scores

Signal vs noise

Meaning
Decides whether an observed group difference is real or random.
Key test
Use when you need to judge if a comparison's gap is meaningful, the next step after comparing.
Example
Is a 2-point gap beyond chance?

Correlation

Meaning
Measures how two variables move together within data, not how two groups differ.
Key test
Use when relating two measured variables, not comparing separate groups.
Formula
r
Example
Do study hours track with scores?

Formula & Notation

Worked Examples

Example 1 — Compare two study methods

Easy

Method A scores average 80 (range 75–85); Method B averages 80 (range 50–110). Which method is more reliable?
Show the steps and the answer
  1. Compare the centers: both methods average 80.

    Equal means say neither method gives a higher typical score, so the center can't decide.

  2. Compare the spreads: A's range is 85−75=10, B's is 110−50=60.

    Reliability means getting similar results every time, which is what a small spread measures.

  3. Size the difference: B's range is 6010=6 times as wide as A's.

    With B a student might score 50 or 110; with A every score lands within 5 points of 80.

Answer: Method A is more reliable despite the equal average

Takeaway: When centers tie, the spread decides which group is more consistent.

Example 2 — Describing one group

Standard

Method A's scores are 75,80,85. Report its typical score.
Show the steps and the answer
  1. Notice that only one group appears: Method A.

    There is nothing to compare against, so this is a description of one data set.

  2. Add the scores: 75+80+85=240.

    The mean needs the total first.

  3. Divide by the number of scores: 2403=80.

    Method A's typical score is 80; calling it good or bad would need a second group to compare with.

Answer: Mean =75+80+853=80

Takeaway: One group calls for a summary; it becomes comparative statistics only when a second group's center and spread sit beside it.

Example 3 — Find the mistake: a gap smaller than the spread

Application

Two brands of AA battery were each tested 5 times (hours of life). Brand X: 8, 10, 11, 12, 9. Brand Y: 9, 12, 8, 11, 12. Jordan writes: "Y's mean is 10.4 h and X's is 10.0 h, so Y is the better battery." Find the mistake and correct it.
Show the steps and the answer
  1. Check the means: X: 8+10+11+12+95=505=10.0 h; Y: 9+12+8+11+125=525=10.4 h.

    Jordan's 0.4-hour gap is computed correctly, so the problem is in what he concludes from it.

  2. Find the spreads: both brands run from 8 to 12 hours, a range of 4 h.

    Batteries of the same brand differ from one another by up to 4 hours.

  3. Set the gap against the spread: 0.4 h is one tenth of the 4-hour range.

    Swapping a single 8-hour battery for a 12-hour one would move a brand's mean by 45=0.8 h, twice the whole gap.

  4. Notice that the two sets overlap completely.

    When the values overlap this much, the honest conclusion is that this test shows no clear difference.

Answer: The data don't show that Y is better: a 0.4-hour gap is small next to the 4-hour spread within each brand.

Takeaway: A gap between means counts only when it is large compared with the spread inside each group; otherwise it may be noise.

Where You'll See It

  • Testing a new tomato fertilizer

    A gardener grows 20 tomato plants with a new fertilizer and 20 without. The fertilized plants average 4.2 kg of fruit, the others 3.9 kg. Before buying more, she checks the spread: if plants within each group differ by 2 kg, a 0.3 kg gap may be luck; if they differ by only 0.2 kg, the fertilizer clearly helps.

  • Choosing a route to school

    Route A averages 25 minutes and never takes more than 28. Route B averages 22 minutes but sometimes takes 40. On an exam morning, the route with the small spread may be the better choice despite its longer average.

  • Comparing schools' pass rates

    School reports compare pass rates, not numbers of passers, because schools differ in size. A school with 300 passers out of 400 students (75%) did better than one with 450 out of 750 (60%), even though the second has more passers.

Common Mistakes

  1. The slip: Comparing only the means and ignoring the spread

    Instead: two groups with equal averages can differ enormously in variability.

  2. The slip: Declaring a difference real just because the numbers differ

    Instead: a small gap may be noise, not signal.

  3. The slip: Comparing groups of very different sizes by raw totals

    Instead: use proportions or per-group rates to compare fairly.


Mini Practice

Try these on your own. Tap Reveal when you want to check.

  1. Group A scored 4,6,8,10,12 and Group B scored 7,8,8,9,8. Compare their means and ranges.

    Hint: Find one center and one spread for each group.

  2. Plants with Fertilizer 1 grew 10,12,11,13 cm. Plants with Fertilizer 2 grew 15,14,16,15 cm. Compare the medians. Which fertilizer looks better?

    Hint: Sort each list and find the middle; then check whether the groups overlap.

  3. Class A's heights have Q1=145 cm, median 150 cm, Q3=155 cm. Class B's have Q1=156 cm, median 160 cm, Q3=164 cm. At least what percent of Class B is taller than Class A's Q3?

    Hint: Where is Class B's Q1 compared with 155 cm?

  4. Class 1 has mean score 78 and Class 2 has mean score 80. Both have a standard deviation of 12 points. Is Class 2 clearly better?

    Hint: Compare the gap between the means with the spread inside each class.

  5. Question A: "What is the typical score in Mr. Kim's class?" Question B: "Did Mr. Kim's class score higher than Ms. Lee's class?" Which one needs comparative statistics?

    Hint: Count how many groups each question is about.

  6. Club A has 30 members and 12 came to a meeting. Club B has 80 members and 20 came. Ken says: "Club B had better attendance, because 20>12." Correct him.

    Hint: The clubs are different sizes, so compare shares.

Want the full set?

50 practice questions for this concept — free to try, every one with a complete worked solution showing the why, not just the answer.

Frequently Asked Questions

Should I compare the means or the medians?

Use medians when either group is skewed or has outliers. If one coffee shop's wait times are 2, 3, 3, 4 and 28 minutes, the mean is 8 minutes, though four of the five customers waited 4 minutes or less; the median, 3 minutes, is a better picture of a typical wait. When both groups are fairly symmetric, means work fine. Whichever you pick, use the same measure for every group.

What does it mean when two box plots overlap a lot?

It means many values from one group are as large as values from the other, so the groups are not clearly different, even if their medians are. A rough rule used in class: if the median of one group lies outside the box of the other, the difference is probably worth taking seriously. Heavy overlap means you would need more data before claiming one group is better.

Can I compare groups measured in different units?

Only after converting to a common unit or scale. One runner's times in minutes and another's in seconds must be put in the same unit first. If the groups took different tests, such as a quiz out of 20 and one out of 50, convert both to percents. Comparing raw numbers in different units is like comparing 3 feet with 2 metres and calling 3 bigger.

Learning Path

Comparative Statistics

You are here

Next →

Signal vs Noise
You need the Mean and Standard Deviation first, because a fair comparison uses both a center and a spread. Comparative statistics adds the side-by-side habit: the same measures for every group, gaps judged against spread, rates when sizes differ. Next, Signal vs Noise asks whether an observed difference between groups is big enough to be real rather than chance.

See Also