Small experiments in reading a number

One worksheet, seven default values, and a stubborn question: what does an average actually summarize? Enter your own numbers below. Everything runs in this browser page.

13 Labs — Small Experiments

One outlier can move an average

A browser-local statistics worksheet. Enter 3–200 finite numbers, then compare the mean and median with and without one identified outlier removed. Optionally resample sample means with replacement. Everything runs in this page; your entries are not uploaded or stored.

Outputs are illustrative, generated from the numbers you type. They are not evidence about the real world, not measurements, and carry no significance claims, p-values or predictions.

This experiment needs JavaScript. Your browser is not running scripts, so the worksheet controls are disabled and no plot or table can be produced. The guides on averages and outliers, sample size and reading experiments are plain pages and still readable.

Worksheet

Dataset and resampling settings

Dataset — comma, space or newline separated, 3–200 finite numbers 2, 3, 3, 4, 5, 6, 100 Example default: 2, 3, 3, 4, 5, 6, 100 — arithmetic mean 123/7 ≈ 17.571, median 4.

Sample size — whole numbers 2–100, not more than the dataset length How many values each resample draws, with replacement.

Repeats — whole numbers 50–500 How many sample means are drawn. Only 50–500 is accepted.

Seed — whole number seed for the pseudorandom draw The same seed and settings reproduce the same sample plot.

Results

No calculation yet. Press “Run calculation” after the worksheet is enabled.

Summary of entered values

Summary statistics of the dataset as entered. Illustrative only.
Statistic	Value
Count	—
Mean	—
Median	—
Minimum	—
Maximum	—
Range	—

With and without one identified outlier

Comparison after removing one identified extreme value. Illustrative only; the removal is a judgment, not a rule.
Statistic	Full dataset	Outlier removed
Identified outlier	—	—
Count	—	—
Mean	—	—
Median	—	—
Range	—	—

Resampled sample means

Settings: none yet.

Summary of drawn sample means. Sampling with replacement; illustrative, not a confidence claim.
Quantity	Value
Samples drawn	—
Sample mean of sample means	—
Smallest sample mean	—
Largest sample mean	—

Assistive data table of every drawn sample mean

Each row is one drawn sample: its index and its mean. Illustrative only.
Draw	Sample mean
—	—

What this worksheet does not do

It makes no inferential significance claims and computes no p-values.
It makes no predictions about new or unseen data.
The resample range and count describe only the draws shown, not the population.
Removing an outlier is a judgment call; the comparison is offered so you can see the effect, not as a recommendation. See the guide on averages and outliers.
Small samples make approximate uncertainty statements less reliable; see sample size and reading the experiment.

Guides that go with the worksheet

One outlier can move an average

Why the mean and the median of the same values can disagree sharply, and why removing an extreme value is a judgment rather than a rule.
What a small sample can and cannot tell you

Approximate uncertainty formulas get shaky when the sample is small. What that means for reading a dot plot of resampled means.
Read the experiment without overclaiming

Plots and summary statistics help you notice unexpected patterns. They do not, by themselves, establish significance or predict anything.

Open the dedicated experiment page — the same worksheet, on its own.