Tensions in Practice: Magnitude sensitivity in tension with a resistant center¶
Three invented readings with one large value
The same readings—0, 0 and 9—have mean 3 and median 0. The mean carries the size of the large value into the result; the median uses the middle position and resists changes to that extreme. Neither arithmetic answer is wrong. Choosing one makes a commitment about whether the large magnitude should influence this summary.
Carry every magnitude
Let each observation affect the total and its equal-weight average.
Resist an extreme value
Describe the middle ordered observation without tracking the size of the extreme.
Why these aims pull against each other
Resistance to an extreme comes from making some magnitude changes irrelevant. That is useful only when those changes are not the intended signal.
Choose an arrangement to see what changes and what remains difficult.
Arrows express the declared relations, not measured effect sizes. Examples and quantities are illustrative.
What this choice protects
What it costs
When it fits
Compare the arrangements
Use the mean
Add the three values and divide by three.
- What it protects
- Every magnitude contributes; the total can be recovered from mean times count.
- What it costs
- One extreme can pull the result away from most observations.
- When it fits
- Total contribution or the population mean is the relevant target.
Illustration note: The values are fixed; no change over time or claimed measurement error is introduced.
Use the median
Sort the same three readings and take the middle.
- What it protects
- The extreme’s size does not drive this center.
- What it costs
- The size of that extreme is absent from the median even if substantively important.
- When it fits
- A middle-position summary is appropriate and extreme magnitude is not the quantity sought.
Illustration note: The median is not a repair that recovers the mean more accurately; it is another statistic.
What this illustration does—and does not—establish
The source supplies the stated tension; the selected arrangements are bounded editorial illustrations. Costs and conditions remain part of the comparison.
- All values are invented; the example does not classify the 9 as an error.
- A single summary can be insufficient regardless of which rule is selected.
- No estimator is universally preferable without specifying the target and the process generating the observations.
Source entries
Aggregation
This source passage supplies the contextual tension. The concrete arrangements and schematic examples are editorial illustrations, not measured findings.
False objectivity
T3: False objectivity. An aggregation function appears mathematically objective: a mean is just arithmetic. Yet the choice of aggregation function—mean vs. median, sum vs. max, equal weighting vs. cap-weighting—is deeply normative. A mean is sensitive to outliers; a median is robust but discards magnitude information.
The source operation
Aggregation collapses many items into a unified form that retains chosen features while suppressing granular detail, formalized in classical statistics as the reduction of a sample to a summary statistic (Fisher, 1925).