Lesson 5 of 6 · 18 min
Comparing groups with summary statistics
Calculate the mean, median, range, IQR and standard deviation of each group with technology, choose the statistics that suit the data, and interpret the differences in context.
What you will be able to do
- Use technology to find the mean, median, range, IQR and standard deviation of two or more groups.
- Choose matching measures of centre and spread for a comparison: the median with the IQR, or the mean with the standard deviation.
- Interpret differences in centre and spread in the context of the data.
Jess lives in Southport and can get to work on the Gold Coast by car or by bus. She timed ten trips each way, in minutes:
| Travel time (minutes) | |
|---|---|
| Car | 22, 24, 25, 25, 26, 27, 28, 29, 30, 44 |
| Bus | 31, 32, 33, 33, 34, 34, 35, 35, 36, 37 |
The 44-minute car trip was the morning of a crash on the highway. Which way should Jess travel? A box plot is one way to compare. Summary statistics are the other, and they let you put a number on every difference.
Compare the summary statistics for the two groups. Then drag the 44-minute car trip down to 31. Which statistics change a lot, and which barely move?
| Statistic | Car | Bus |
|---|---|---|
| Mean | 28 min | 34 min |
| Median | 26.5 min | 34 min |
| Range | 22 min | 6 min |
| IQR | 4 min | 2 min |
| Standard deviation | 6.11 min | 1.83 min |
Centre. Both the mean and the median say the car is faster. The median car trip (26.5 minutes) was 7.5 minutes shorter than the median bus trip (34 minutes).
Spread. The bus was far more consistent: its standard deviation (1.83 minutes) and IQR (2 minutes) are both smaller than the car's (6.11 and 4 minutes).
Outlier. The 44-minute crash day is a possible outlier for the car (its upper fence is ). It drags the car's mean up to 28 minutes and more than doubles its standard deviation, but the median and IQR hardly notice it.
Key idea
Choose statistics that suit the data
Compare like with like, and pick the pair that fits the shape of both groups:
- Symmetric, no outliers: compare the means and the standard deviations. They use every value.
- Skewed, or any outliers: compare the medians and the IQRs. They are resistant to extreme values.
The range is quick but depends on only two values, so use it alongside the IQR or standard deviation, not instead of them.
Jess's car times include a possible outlier (the 44-minute crash day). Which statistics are most appropriate for comparing her car and bus trips?
Summary statistics for two groups with technology
Enter each group in its own list
For example, the car times in List 1 and the bus times in List 2.
Run one-variable statistics on each list
Choose 1-Var Stats (or the equivalent) for List 1, record the results, then repeat for List 2.
Record what you need
, , (not ), the minimum, , the median, and the maximum. Calculate the range and IQR from these.
Check
Check that is right for each group and that each mean and median lies between that group's minimum and maximum.
Worked example
Compare two filling machines with technology
A factory near Bundaberg fills packets of macadamias labelled 500 g. A sample of eight packets from each of two machines had these masses, in grams:
| Mass (g) | |
|---|---|
| Machine A | 498, 501, 499, 502, 500, 497, 503, 500 |
| Machine B | 492, 507, 495, 510, 499, 488, 504, 505 |
Use technology to compare the two machines. Which machine is better?
Step 1: Find the statistics for each machine
With one-variable statistics:
Median Range IQR Machine A 500 g 2.00 g 500 g 6 g 3 g Machine B 500 g 7.78 g 501.5 g 22 g 12.5 g For the IQR, the halves of A are 497, 498, 499, 500 and 500, 501, 502, 503, so and ; for B, and .
Step 2: Choose the statistics to compare
Both data sets are roughly symmetric with no outliers (A's fences are 494 and 506; B's are 474.75 and 524.75), so compare the means and standard deviations.
Step 3: Compare centre
Both machines have a mean of 500 g, exactly the labelled mass, so on average both fill correctly.
Step 4: Compare spread
Machine B's standard deviation (7.78 g) is almost four times Machine A's (2.00 g), and its range is 22 g compared with 6 g. Machine A is far more consistent.
Step 5: Answer the question
Machine A is better: both machines average 500 g, but A's packets are much closer to 500 g. Two of B's eight packets were 8 g or more under the label.
Try it yourself first, then check the working one step at a time.
Answer
Both machines have a mean of 500 g, but Machine A is much more consistent ( g compared with 7.78 g), so Machine A is better.
Method
Enter each group in its own list and run one-variable statistics on each.
Check the shape and outliers to choose mean and , or median and IQR.
Compare centre, then spread, with numbers and units.
Answer the question in context.
Watch out for
Reading (1.87 g and 7.28 g) instead of .
Concluding the machines are the same because the means are equal. Spread matters too.
Use technology to find the mean of Jess's car times without the 44-minute crash day: 22, 24, 25, 25, 26, 27, 28, 29, 30. Give your answer correct to one decimal place.
Worked example
Choose statistics to compare part-time earnings
A Cairns school surveyed 25 Year 11 and 25 Year 12 students with part-time jobs about their weekly earnings.
| Mean | Median | Min | Max | ||||
|---|---|---|---|---|---|---|---|
| Year 11 | $182 | $150 | $96 | $40 | $105 | $195 | $480 |
| Year 12 | $214 | $205 | $62 | $80 | $160 | $245 | $360 |
Decide which statistics are most appropriate for comparing the two groups, and use them to compare the weekly earnings.
Step 1: Check for outliers
Year 11: , upper fence , so the $480 is a possible outlier. Year 12: , fences and , so no outliers.
Step 2: Check the shape
Year 11's mean ($182) is well above its median ($150), and its maximum is far above : the data are positively skewed.
In skewed data, the mean is pulled towards the tail.
Step 3: Choose the statistics
Because Year 11 is skewed and has an outlier, compare the medians and IQRs. They are resistant to the $480.
Step 4: Compare centre
The median for Year 12 ($205) was $55 higher than for Year 11 ($150), so Year 12 students typically earned more.
Step 5: Compare spread
The IQRs are similar ($90 for Year 11 and $85 for Year 12), so the middle halves of the two groups were about equally spread out.
The standard deviations ($96 and $62) suggest Year 11 is much more variable, but much of that comes from the single $480 outlier.
Try it yourself first, then check the working one step at a time.
Answer
Compare the medians and IQRs, because Year 11's earnings are positively skewed with an outlier ($480). Year 12 students typically earned $55 more per week (median $205 compared with $150), and the spread of the middle half was similar (IQR $85 compared with $90).
Method
Check each group for skew (mean far from median) and outliers (fences).
If either group is skewed or has outliers, compare medians and IQRs; otherwise means and standard deviations.
Compare centre and spread with differences and units.
Watch out for
Comparing the means (a $32 difference) when one group is skewed, which understates the typical difference.
Concluding Year 11 is far more variable from the standard deviations alone.
Mean and standard deviation
- Use every value, so they make the most of the data.
- Best for symmetric data with no outliers.
- Sensitive: one extreme value can move both a lot.
- Standard deviation comes from technology.
Median and IQR
- Use only the middle of the ordered data.
- Best for skewed data or data with outliers.
- Resistant: extreme values barely move them.
- Match what a box plot shows.
Common mistake
Comparison slips
- Mixing families: comparing one group's mean with the other's median, or a standard deviation with an IQR.
- Reading spread as size: a larger standard deviation means more variable, not larger values.
- Ignoring the outliers: if either group has an outlier, say so and use the median and IQR.
- Numbers without meaning: "6.11 > 1.83" earns little. "Car trips varied much more than bus trips ( 6.11 minutes compared with 1.83 minutes)" earns the marks.
Summary
Find each group's statistics separately with technology, and read , not .
Compare centre (which group is typically higher, and by how much) and spread (which group is more variable or consistent).
Use the means and standard deviations for symmetric data without outliers; the medians and IQRs for skewed data or data with outliers.
Outliers pull the mean, range and standard deviation, but barely move the median and IQR.
Every comparison names the groups and the variable and quotes the statistics with units.
Finished the lesson?
Try the 2 quick checks above and the lesson completes itself, or mark it complete now.