Interpreting Medians and Quartiles in Boxplots

Reading the Center and Middle Half

A boxplot gives a compact picture of a quantitative distribution. The line inside the box marks the median, while the two ends of the box mark the first and third quartiles. Reading these three positions answers two useful questions: where is the distribution centered? and between which values does its middle half lie?

The median and quartiles are locations in the ordered data, not counts of units or distances along the axis. Their positions on the scale report values in the variable’s original units. For example, a median of 18 minutes describes a typical central travel time in minutes; it does not mean that half the trips take 18 minutes.

A boxplot can be drawn horizontally or vertically. In either orientation, locate the median line and the two box edges, then read each position against the numerical axis. For a horizontal boxplot, the lower-value edge is on the left; for a vertical boxplot, it is lower on the page. “Lower” and “upper” refer to the values, not necessarily to the physical direction on the screen.

Quartile boundaries divide the ordered observations into sections. Roughly one quarter of the observations lie at or below \(Q_1\), and roughly one quarter lie at or above \(Q_3\). The interval from \(Q_1\) to \(Q_3\) therefore covers the middle half by rank. If many observations have the same value, some may equal a boundary, so the groups are not always exactly 25%, 50%, and 25%. Read the box as a rank-based summary, not as a promise about exact percentages in every tied data set.

Interpreting the Median

The median is the value at the center of the ordered observations. It divides the data into two sides by rank: about half the observations are at or below it and about half are at or above it. With an even number of observations, a median can fall between two measured values; a boxplot may therefore mark a value that is not itself an observed measurement.

A contextual description should name both the variable and its units. If a boxplot of delivery times has its median at 32 minutes, a clear interpretation is: “The median delivery time is 32 minutes; about half of the recorded deliveries took no more than 32 minutes and about half took no less.” With tied times, “about half” is safer than claiming exactly half are strictly below and strictly above.

Interpreting the Box

The two box edges give the endpoints of the middle-half interval. The lower edge is \(Q_1\); the upper edge is \(Q_3\). A contextual interpretation names both endpoints and describes the interval as covering the central half of the observations by rank. It is not correct to say that every observation lies inside the box: the box represents the middle half, while observations outside it are part of the lower or upper portions of the ordered data.

The length of the box on the value axis is \(Q_3-Q_1\). This is a distance between values, so its units match the variable’s units. For example, if quartiles are measured in minutes, the IQR is also measured in minutes. The IQR is not a percentage, and it is not the number of observations inside the box.

Why the Middle-Half Description Works

The box’s interpretation follows from the way quartiles locate portions of an ordered data set. Imagine arranging all observations from smallest to largest. The median marks the center of this ordering. The first quartile marks a lower-quarter boundary, and the third quartile marks an upper-quarter boundary. Consequently, the observations between those boundaries occupy the central half of the ordering.

This is a statement about rank, not about equal spacing on the measurement scale. The middle half may span a short or long interval of values. For example, a box covering 10 to 20 units has width 10, while a box covering 10 to 80 units has width 70; both represent the middle half of their respective data sets.

This rank-based meaning is useful when comparing distributions measured on the same scale. The medians compare central locations; the boxes compare the locations and widths of the middle-half intervals. These are descriptions of the displayed data. A boxplot alone does not establish why groups differ or whether a difference is statistically meaningful.

Read Positions Carefully

To interpret a plotted box, trace each marked boundary to the numerical axis and use the axis scale. If the boundary lies between labeled ticks, estimate its value in keeping with the spacing of the scale. Do not assume that equal-looking sections of a drawing represent equal numbers of observations; the box is marked by quartile positions, not by a count scale.

A reliable description keeps three ideas separate:

• Median: the central value by rank, stated in the variable’s units.
• Quartile boundaries: the values \(Q_1\) and \(Q_3\) that mark the ends of the middle-half interval.
• IQR: the numerical width \(Q_3-Q_1\) of that interval.

These quantities answer related but different questions. The median gives one central location. The two quartiles give an interval. The IQR gives the interval’s width. Keeping them distinct prevents a common interpretation error: \(Q_3\) is not the width of the box; it is the upper boundary value.

Common Mistakes

• Calling the median the mean. The line inside the box marks the median, the central value by rank. It does not generally equal the arithmetic average.
• Calling \(Q_3\) the IQR. \(Q_3\) is an upper boundary value. The IQR is the difference \(Q_3-Q_1\).
• Claiming every observation is in the box. The box marks the middle half by rank, not the full range of observations.
• Ignoring units or context. “The median is 18” is incomplete when the axis measures minutes, centimeters, or another unit.
• Claiming exact quarter counts despite ties. Repeated values can lie on a quartile boundary, so use “about” or explain the rank-based meaning rather than asserting strict percentages.

A useful final check is to read the box from low to high: identify \(Q_1\), identify the median, identify \(Q_3\), then subtract the first boundary from the third if the IQR is needed. State each result in context and in the units shown on the axis.