
compute the mean, median, mode, variance, standard deviation, quartiles and interquartile range of a data set by hand and with a spreadsheet, and use the 1.5 × IQR rule to identify outliers.
Instruction
A display shows a distribution; summary statistics measure it. This lab builds every common summary from its definition, using St. Louis precipitation for the ten years 2015-2024 (inches): 61.28, 41.46, 36.69, 42.62, 54.02, 50.02, 42.51, 48.96, 33.18, 50.84.
Mean. x̄ = Σx / n = 461.58 / 10 = 46.158 inches. The mean is the balance point of the data.
Median. Sort the values: 33.18, 36.69, 41.46, 42.51, 42.62, 48.96, 50.02, 50.84, 54.02, 61.28. With n = 10 the median is the average of the 5th and 6th values: (42.62 + 48.96) / 2 = 45.79 inches. The median is the middle of the data.
Mode. The most frequent value. Here no value repeats, so the mode tells us nothing. For the full 75-year St. Louis annual mean temperature record the mode is 54.6°F (5 years). The mode is most useful for categorical data.
Variance and standard deviation. Spread is measured by how far values sit from the mean. The deviations x − x̄ are 15.12, −4.70, −9.47, −3.54, 7.86, 3.86, −3.65, 2.80, −12.98, 4.68. They always sum to zero, so we square them before adding: Σ(x − x̄)2 ≈ 641.14. The sample variance divides by n − 1:
s2 = 641.14 / 9 ≈ 71.24 in2, and s = √71.24 ≈ 8.44 inches.
Why n − 1? The deviations are measured from x̄, which was itself computed from the same data, so they are slightly smaller than deviations from the true mean would be. Dividing by n − 1 corrects for this, so s2 does not systematically underestimate the population variance. When the data are the whole population, divide by N and write σ2. The standard deviation is in the original units: a typical year at St. Louis differs from the mean by roughly 8 inches.
Quartiles and IQR. Q1 is the median of the lower half (33.18, 36.69, 41.46, 42.51, 42.62), so Q1 = 41.46. Q3 is the median of the upper half (48.96, 50.02, 50.84, 54.02, 61.28), so Q3 = 50.84. IQR = 50.84 − 41.46 = 9.38 inches, the spread of the middle half. Spreadsheets use a different rule: QUARTILE.INC gives 41.72 and 50.64 for these data. With small samples the methods disagree, so always say which one you used.
Outliers. The 1.5 × IQR rule flags any value below Q1 − 1.5(IQR) = 41.46 − 14.07 = 27.39 or above Q3 + 1.5(IQR) = 50.84 + 14.07 = 64.91. The very wet year of 2015 (61.28) is high but is not an outlier among these ten values. For all 75 years (Q1 = 32.52, Q3 = 43.70, IQR = 11.18), the upper fence is 60.47, and only 2015 (61.28 inches) is an outlier. The same value is unusual against 75 years and ordinary against ten, because a decade that contains several wet years makes a wet year look less strange. Always say what a value is being compared with.
Resistance. Suppose, as an illustration only, that 2015 had been a freak 100 inches instead of 61.28. The ten-year mean would jump from 46.16 to 50.03 inches, while the median would stay at 45.79. For the real 75-year record, dropping 2015 lowers the mean from 38.81 to 38.51 and the median from 37.67 to 37.53. The median and IQR are resistant: a single extreme value barely moves them. The mean and standard deviation are not. That is the rule for choosing: report median and IQR for skewed data or data with outliers, and mean and standard deviation for roughly symmetric data.
Spreadsheet check. =AVERAGE(), =MEDIAN(), =MODE.SNGL(), =VAR.S(), =STDEV.S() (sample) or =STDEV.P() (population), and =QUARTILE.INC(range, 1).
📖 Read for this lesson
The reading opens over the lesson; close it to come back to exactly where you were. The checks below draw on it.
Formative check
Work through these before moving on. They are not graded — they tell you whether the standard below has landed yet.
For the ten St. Louis values the mean is inches, the median is inches, and the sample standard deviation is about inches.
Why does the sample variance divide by n − 1 instead of n?
By the 1.5 × IQR rule, St. Louis’s 2015 total of 61.28 inches is an outlier among the 2015-2024 values.
Build a spreadsheet for all 75 years of St. Louis precipitation. Report the mean, median, sample SD, quartiles (state the method), IQR and fences, and name every outlier. Then compare your hand-computed ten-year SD (8.44) with the 75-year SD (8.42). They are almost equal even though the ten-year mean (46.16) is far above the 75-year mean (38.81). In two sentences explain what this tells you about the difference between a distribution’s center and its spread.
How confident are you that you can compute center and spread from their definitions and choose resistant measures when the data call for them?
Practice
Work these on paper or in your notebook, then open Check your answer. Aim for all of Fluency and Application; try at least one Challenge.
Printable version: this unit’s practice workbook (PDF)
Fluency
Build speed and accuracy with the core skill.
- For the data 3, 4, 11, 12, 26, 27, 28: find the mean, median, range and interquartile range. (Find the quartiles as the medians of the lower and upper halves, leaving out the overall median when the count is odd.)
Check your answer
Answer: mean = 15.86; median = 12.0; range = 25; IQR = 23.0 (Q1 = 4.0, Q3 = 27.0) - For the data 2, 3, 11, 15, 18, 21, 27, 30, 34: find the mean, median, range and interquartile range. (Find the quartiles as the medians of the lower and upper halves, leaving out the overall median when the count is odd.)
Check your answer
Answer: mean = 17.89; median = 18.0; range = 32; IQR = 21.5 (Q1 = 7.0, Q3 = 28.5) - For the data 2, 5, 6, 19, 23, 34, 39: find the mean, median, range and interquartile range. (Find the quartiles as the medians of the lower and upper halves, leaving out the overall median when the count is odd.)
Check your answer
Answer: mean = 18.29; median = 19.0; range = 37; IQR = 29.0 (Q1 = 5.0, Q3 = 34.0) - For the data 2, 3, 9, 10, 22, 24, 35, 37: find the mean, median, range and interquartile range. (Find the quartiles as the medians of the lower and upper halves, leaving out the overall median when the count is odd.)
Check your answer
Answer: mean = 17.75; median = 16.0; range = 35; IQR = 23.5 (Q1 = 6.0, Q3 = 29.5) - For the data 7, 8, 10, 15, 15, 16, 17, 19, 24, 24: find the mean, median, range and interquartile range. (Find the quartiles as the medians of the lower and upper halves, leaving out the overall median when the count is odd.)
Check your answer
Answer: mean = 15.50; median = 15.5; range = 17; IQR = 9.0 (Q1 = 10.0, Q3 = 19.0) - For the data 6, 7, 22, 28, 33, 34, 40: find the mean, median, range and interquartile range. (Find the quartiles as the medians of the lower and upper halves, leaving out the overall median when the count is odd.)
Check your answer
Answer: mean = 24.29; median = 28.0; range = 34; IQR = 27.0 (Q1 = 7.0, Q3 = 34.0) - Kansas City International Airport annual mean temperatures, 2020-2024 (°F): 55.7, 57.0, 55.7, 57.0, 57.1. Compute the mean, the deviations, the sum of squared deviations, the sample variance and the sample standard deviation by hand.
Check your answer
Answer: Mean 56.50°F; deviations −0.80, 0.50, −0.80, 0.50, 0.60; Σ(x − x̄)² ≈ 2.140; s² ≈ 0.535; s ≈ 0.731°F.Divide the sum of squares by n − 1 = 4. - For the same five values, what would the population standard deviation (dividing by n) be, and when would it be the right choice?
Check your answer
Answer: σ ≈ 0.654°F. Use it only when the five values are the entire population of interest.STDEV.P divides by n; STDEV.S divides by n − 1.
Application
Use the skill in context. Show your reasoning.
- Kansas City precipitation, 2015-2024 (NOAA, inches): 46.61, 48.67, 46.04, 40.78, 53.22, 34.10, 41.26, 34.57, 37.40, 34.67. Find the mean, median, sample SD, Q1, Q3 and IQR (median-of-halves method), and the 1.5 × IQR fences. Are there outliers?
Check your answer
Answer: Mean 41.73, median 41.02, s ≈ 6.68; Q1 = 34.67, Q3 = 46.61, IQR = 11.94; fences 16.76 and 64.52. No outliers.Sort the ten values; the halves are the lowest five and highest five. - Remove the wettest year from the Kansas City 2015-2024 list. How do the mean and the median change, and which is more resistant?
Check your answer
Answer: Mean 41.73 → 40.46 (down 1.27); median 41.02 → 40.78 (down 0.24). The median moves less, so it is more resistant.With n now odd, the median is the middle (5th) value. - A spreadsheet’s QUARTILE.INC gives different quartiles from your hand method for the same ten values. Which is correct, and what should your report say?
Check your answer
Answer: Both are legitimate conventions; state which method you used so readers can reproduce it.With small samples, quartile methods disagree slightly.
Challenge
Stretch problems. Expect to think before you write.
- Show that the deviations from the mean always sum to zero, and explain why this is why variance squares them.
Check your answer
Answer: Σ(x − x̄) = Σx − n·x̄ = Σx − Σx = 0. Unsquared deviations cancel, so their average says nothing about spread; squaring makes each deviation count positively.Use x̄ = Σx / n. - In the ten St. Louis values for 2015-2024 (mean 46.158 in, s ≈ 8.44 in), replace 2015’s 61.28 in with the mean, 46.158. Without recomputing everything, say whether s increases or decreases; then compute it.
Check your answer
Answer: It decreases, to s ≈ 6.58 in.The largest squared deviation (about 228.7) is replaced by 0, and the mean is unchanged.
Review
Keep earlier skills sharp.
- Solve: −6x − 2 = 16
Check your answer
Answer: x = −3Add 2 to both sides: −6x = 18. Divide by −6. - (Prerequisite: solving an equation) Four years at a station had 38, 41, 44 and 35 inches of rain. How much rain must a fifth year have for the five-year mean to be 40 inches?
Check your answer
Answer: 42 inches.(38 + 41 + 44 + 35 + x)/5 = 40 → 158 + x = 200.
The Same 75 Years, Sorted: Quartiles, Fences and One Outlier by Hand
NOAA NCEI Global Summary of the Year, St. Louis Lambert International Airport (USW00013994), calendar-year precipitation 1950-2024, sorted; retrieved November 10, 2026 · Public domain (works of the United States government)
Why this reading. The lab computes every summary by hand for ten years and then asks you to do the 75-year version in a spreadsheet and name every outlier. Here are the same 75 St. Louis totals sorted from driest to wettest, with their ranks, so that the quartiles can be found by counting, the fences checked, and the spreadsheet’s answer compared with the hand method. Every figure the lesson states for the full record is reproduced below.
The 75 years, sorted
Calendar-year precipitation at St. Louis Lambert International Airport (USW00013994), 1950-2024, NOAA Global Summary of the Year. Ties keep both years (33.50 in 1991 and 1992).
| Rank | Inches | Year | Rank | Inches | Year | Rank | Inches | Year |
|---|---|---|---|---|---|---|---|---|
| 1 | 20.59 | 1953 | 26 | 34.07 | 1999 | 51 | 42.51 | 2021 |
| 2 | 23.48 | 1976 | 27 | 34.44 | 1956 | 52 | 42.62 | 2018 |
| 3 | 25.69 | 1952 | 28 | 34.67 | 1962 | 53 | 42.69 | 2013 |
| 4 | 27.50 | 1980 | 29 | 34.72 | 1994 | 54 | 43.44 | 1977 |
| 5 | 27.63 | 1954 | 30 | 34.90 | 1986 | 55 | 43.45 | 2014 |
| 6 | 28.28 | 1965 | 31 | 35.33 | 2001 | 56 | 43.64 | 1998 |
| 7 | 28.35 | 1959 | 32 | 36.24 | 1970 | 57 | 43.70 | 1996 |
| 8 | 28.61 | 1989 | 33 | 36.35 | 1951 | 58 | 43.73 | 1969 |
| 9 | 28.65 | 1963 | 34 | 36.69 | 2017 | 59 | 44.83 | 1983 |
| 10 | 29.50 | 1979 | 35 | 36.86 | 1974 | 60 | 45.11 | 1990 |
| 11 | 29.94 | 2006 | 36 | 37.37 | 1958 | 61 | 45.56 | 1981 |
| 12 | 30.57 | 2007 | 37 | 37.39 | 2000 | 62 | 46.09 | 2003 |
| 13 | 31.24 | 1997 | 38 | 37.67 | 1950 | 63 | 47.17 | 1957 |
| 14 | 31.33 | 1955 | 39 | 37.74 | 1978 | 64 | 47.20 | 2011 |
| 15 | 31.80 | 1960 | 40 | 37.87 | 2005 | 65 | 48.96 | 2022 |
| 16 | 32.17 | 1964 | 41 | 38.43 | 1987 | 66 | 50.02 | 2020 |
| 17 | 32.33 | 2012 | 42 | 39.12 | 2010 | 67 | 50.75 | 1985 |
| 18 | 32.35 | 1966 | 43 | 39.85 | 1973 | 68 | 50.84 | 2024 |
| 19 | 32.52 | 1968 | 44 | 40.20 | 1975 | 69 | 50.94 | 2009 |
| 20 | 33.18 | 2023 | 45 | 40.98 | 2002 | 70 | 51.68 | 1984 |
| 21 | 33.50 | 1991 | 46 | 41.23 | 1961 | 71 | 54.02 | 2019 |
| 22 | 33.50 | 1992 | 47 | 41.32 | 1967 | 72 | 54.79 | 1993 |
| 23 | 33.73 | 1971 | 48 | 41.46 | 2016 | 73 | 54.99 | 1982 |
| 24 | 33.77 | 1972 | 49 | 41.66 | 1995 | 74 | 57.96 | 2008 |
| 25 | 33.95 | 1988 | 50 | 42.30 | 2004 | 75 | 61.28 | 2015 |
Center and quartiles by counting
With n = 75 the median is the 38th value: 37.67 inches (1950). The lower half is ranks 1 to 37, and its median is rank 19: Q1 = 32.52 (1968). The upper half is ranks 39 to 75, and its median is rank 57: Q3 = 43.70 (1996). So IQR = 43.70 − 32.52 = 11.18 inches.
Spreadsheets do it differently. QUARTILE.INC places Q1 at position 1 + 0.25(74) = 19.5, halfway between ranks 19 and 20: (32.52 + 33.18)/2 = 32.850. Q3 sits at position 56.5, halfway between 43.64 and 43.70: 43.670. The two methods differ by a few hundredths of an inch here, which is why a report has to say which one it used. Both give the same median.
Mean and standard deviation
The 75 values sum to 2910.99, so the mean is 2910.99 / 75 = 38.8132 inches. The squared deviations from that mean sum to 5245.48. Dividing by n − 1 = 74 gives s² = 70.885 in² and s = √70.885 = 8.42 inches. Dividing by 75 instead (treating the 75 years as the whole population) gives 8.36. A spreadsheet’s =STDEV.S gives 8.42 and =STDEV.P gives 8.36; if yours disagrees, the usual cause is a value typed wrong, and the sum check (2910.99) will find it.
The fences and the one outlier
1.5 × IQR = 16.770. Lower fence: 32.52 − 16.770 = 15.750, well below the driest year on record (20.59). Upper fence: 43.70 + 16.770 = 60.470. Rank 74, the 57.96 inches of 2008, is inside the fence. Rank 75, the 61.28 inches of 2015, is outside it. One outlier in 75 years. With QUARTILE.INC quartiles the upper fence is 43.670 + 1.5(10.820) = 59.90, and 2015 is still the only value above it, so the conclusion does not depend on the method here. It can in smaller data sets.
The same year, two verdicts
The lesson’s ten years, 2015-2024, have Q1 = 41.46, Q3 = 50.84 and an upper fence of 64.91. Inside that decade, 2015 (61.28) is not an outlier; against all 75 years it is. Nothing about the rain changed. Only the comparison group did.
Resistance, measured
Remove 2015 and recompute on the other 74 years: mean 38.51 (was 38.81), median 37.53 (was 37.67), SD 8.05 (was 8.42), IQR 11.12 (was 11.18). The mean and SD moved the most; the median and IQR barely moved.
Why the ten-year SD matches the 75-year SD
The lesson’s ten years, 2015-2024, have s = 8.44 inches; the 75 years have s = 8.42. They are nearly equal even though the decade’s mean (46.16) is 7.35 inches above the 75-year mean (38.81). Look at the ranks: the ten recent years sit at ranks 20, 34, 48, 51, 52, 65, 66, 68, 71, 75 of 75. Five of the ten are among the 11 wettest years in the record, so the decade is shifted wet, but the ten are spread from rank 20 to rank 75, nearly as widely as the whole record. Spread measures how far values sit from their own mean, so a whole decade can sit higher without being any more or less spread out.
What to do
- Enter the 75 values (or the year-ordered table from Lesson 1.1) into a spreadsheet and confirm 2910.99, 38.81, 37.67, 8.42, 32.850 and 43.670. State the quartile method you report.
- Compute the fences under both methods and name every outlier. There should be exactly one.
- Remove 2015 and recompute the mean, median, SD and IQR. Which two statistics barely moved? That is what resistant means.
- Write two sentences on why the ten-year SD (8.44) is almost the same as the 75-year SD (8.42) even though the ten-year mean is much higher, using the ranks above.
Sources: NOAA National Centers for Environmental Information, Global Summary of the Year (GSOY), station USW00013994 (St. Louis Lambert International Airport), data service query used here (ncei.noaa.gov)