Median
The median is the middle value of a set of measurements arranged in order, a useful summary when one trial is far off from the rest.
Definition
The median is the middle number when you line up your measurements from smallest to largest, with half the trials below it and half above. With an odd number of values it is the middle one, and with an even number it is the mean of the two middle values. In lab data it is preferred over the mean when a set contains an outlier, because it depends only on the order of the values and not on how extreme the outlier is. As the 50th percentile it is a robust estimator of central tendency, appropriate for skewed distributions and ordinal data, though it discards information the mean uses and is less efficient when errors are truly random and symmetric.
Example
Five bean plants measured $5$, $7$, $8$, $9$, and $21\ \text{cm}$ have a median of $8\ \text{cm}$, and the $21\ \text{cm}$ plant does not drag it up the way it would drag the mean. Four dissolving-time trials of $25$, $26$, $27$, and $58\ \text{s}$ (the last started late) have a median of $26.5\ \text{s}$, while the mean of $34\ \text{s}$ describes none of the trials. Germination times of $3$, $3$, $4$, $4$, $4$, $5$, $6$, $9$, and $14$ days have a median of $4$ days against a mean of $52/9 \approx 5.8$ days, because two late seeds pull the mean well above what most seeds did.
Key Insight
When one measurement looks wildly different from the rest, the median tells you what a typical trial looked like better than the mean does. If the mean and median are far apart, your data are skewed or contain an outlier, and that gap is a signal to look back at the raw trials before reporting anything. Robustness is a trade: the median ignores outliers, which protects against errors but also hides a real effect if those extreme values are genuine, so decide which risk matters before choosing the statistic.