Practical technology learning
Mean or Median? Check What a Typical Value Means in Uneven Data
A single summary number can make a dataset easier to discuss, but it can also conceal important differences between records. Compare the mean and median on a tiny synthetic dataset, check both calculations and then explain the decision each summary could support. These values are invented for learning and are not Mango Engineers business results.
Begin with six visible observations
Use the values 100, 100, 100, 100, 100 and 1000. They are already ordered. Their total is 1500, so dividing by six gives a mean of 250. For an even number of observations, the median is the average of the two middle values; here both are 100.
Calculate the result manually before using a spreadsheet or script. Keep the original list visible so you can check whether a formula range accidentally omits a record or includes an extra cell.
Ask what typical means for the question
The mean uses every value and relates directly to the total divided by the record count. The median describes the middle of the ordered observations. The larger value affects the mean strongly in this example while leaving the middle unchanged.
Do not conclude that one measure is always honest and the other is misleading. Start with the reporting question. A question about total activity per order may need a different explanation from a question about a middle observation.
Investigate an unusual value before changing it
Check whether all records use the same unit, time period and population. The value 1000 could be a legitimate large order, an entry error or a different kind of record. The number alone does not tell you which explanation is correct.
Keep a note of what you checked and what remains uncertain. Do not remove a record just to obtain a preferred summary. If you present a filtered analysis, state the filter and retain the unfiltered result for comparison.
Make the sample size visible
Six observations are useful for learning because every value can be inspected. They do not establish a general pattern for a business, a city or a course. Avoid presenting the exercise as evidence about real customers or learners.
If you add more synthetic records, recalculate both measures and explain why they changed. Use a different distribution rather than repeatedly changing only the highest value until the result looks attractive.
Build a repeatable check
- The source contains exactly six observations.
- The sum is 1500 and the mean is 250.
- The ordered middle pair is 100 and 100.
- The reported unit is consistent across records.
- The high value remains visible unless a justified rule excludes it.
- The summary includes the record count and the question it addresses.
Write a conclusion a reader can interpret
A useful practice conclusion is: 'For these six invented order values, the mean is 250 and the median is 100; the difference reflects one much larger observation.' This states the evidence without inventing a cause.
For guided learning, discuss Mango Engineers' data analytics path and the relationship between spreadsheet work, SQL, analysis and reporting. Confirm the current course scope and prerequisites with the team.