Data literacy · Grades 5–8
Two datasets can share an average and tell different stories
Keep the observations before calculating
A set of numbers needs a meaning. Suppose two fictional groups record minutes needed to complete a short puzzle. Group A records 4, 5, 5, 5, 6. Group B records 1, 2, 5, 8, 9. These invented values provide a teaching comparison; they are not evidence about real students.
Both groups have mean 5 minutes. A’s times cluster closely around five, while B’s are spread widely. If the question concerns predictability, the mean alone misses a central difference. Retain the observations, units, and collection method beside any summary.
The data desk begins with a dot plot because each observation remains visible. Repeated values are stacked rather than hidden behind one mark.
Find the mean with its units attached
Add the values and divide by the number of observations. For A, (4 + 5 + 5 + 5 + 6) ÷ 5 = 5 minutes. The unit remains minutes; the denominator counts observations rather than changing the measurement into another physical quantity.
A mean can fall between observed values. It need not be a value that any participant recorded. It summarizes a numerical balance, not a most common case.
Before combining values, check whether they measure the same thing in compatible units. Averaging a distance and a time does not produce a meaningful average merely because both are numbers.
Sort before finding the median
The median is the middle value after sorting. With an odd number of observations, one value occupies the middle position. With an even number, take the mean of the two middle values. For 2, 4, 7, 9, the median is (4 + 7) ÷ 2 = 5.5.
The middle position in the original entry order is not necessarily the median. The data desk sorts a copy for the calculation while displaying the entered observations so the transformation can be checked.
The mean and median can coincide, but they need not. Use both when the difference helps explain the shape of a dataset, rather than automatically reporting them as two names for the same statistic.
Explore an unusual value without silently deleting it
For 2, 3, 4, 5, 26, the mean is 8 and the median is 4. The large value raises the mean. That does not make the value invalid. It could be a real observation, a recording mistake, or a measurement made under different conditions.
Investigate the reason before removing it. If you show a summary with and without the value, label both versions and explain the decision. “It changes my conclusion” is not a sufficient reason to erase data.
Use the “add 30” example in the data desk to compare summaries. Then remove it and explain which quantities changed most. This is a demonstration of sensitivity, not a rule that all large values should be excluded.
Choose a display that preserves the question
A dot plot shows individual numeric observations and their distribution. A bar chart compares categories. A line graph can show a sequence over ordered time. Connecting arbitrary categories with a line suggests an order or continuity the data may not have.
Label axes and units. Consider how a truncated axis changes the visual impression. A graphic can be technically drawn from the values yet encourage a misleading comparison if its scale is not clear.
The plot here labels the minimum and maximum and positions values proportionally within the displayed range. It is a small exploratory display rather than a full statistical reporting system. An equal-valued dataset receives a padded range so every point remains visible.
Sampling determines who the summary describes
A summary of five puzzle times describes those observations. It does not establish the performance of an entire school. Selection, task difficulty, instructions, and missing responses all affect what broader claims are reasonable.
A large voluntary survey shared through one club can remain biased toward that club’s participants. Sample size and selection method address different issues. State the population of interest, who could participate, who did participate, and what groups may be absent.
When interpreting a report, trace the denominator. A percentage of respondents is not automatically a percentage of all people in the population. The arithmetic can be correct while the accompanying claim is too broad.
Write the story the numbers can support
Include the variable, units, source or collection method, a display, relevant summaries, and a bounded conclusion. Name both the pattern and a limitation. For the fictional puzzle data, a useful statement is that the means match while the observed spread differs; a claim about which group is smarter would not be supported.
For an extended task, invent two small datasets with the same median and different means. Explain how you constructed them and why the difference appears. Then use actual, appropriate classroom data only with educator guidance and without collecting unnecessary personal information.
Census Statistics in Schools provides official datasets and classroom resources for further data literacy. Keep the real source and definitions attached when using those materials.
Classroom use: print this guide, annotate the worked examples, and write your own explanation before checking the practice questions. Teaching guidance