§11.2.8

Scatter and Error Bars

A point has two readings—how are x and y recorded?

✅ NormativeSettled. The requirement holds as written.📄 Page ruleCan be applied from the page alone.Read §11.2.8 in the RFC ↗

The requirement

A named scatter point MUST be rendered as one row holding its x value and its y value. Marker series over categories are ordinary series columns. An error bar's centre is an ordinary value, and its whisker span MUST be given with it: the low and high ends read at the whisker caps. A box plot MUST give its median, first and third quartiles, and its whisker ends when they are drawn. Points with no printed name have nothing to locate them and are not rendered as rows.

Reading the examples. Each one pairs a source page with the output the rule expects. ✅ Sample: the rule applies. ❌ Counter-sample: it looks like it applies and does not, or a rendering got it wrong. A non-conforming output, in red, is shown only for contrast. … marks content left out of an excerpt.

Scatter

Each named dot contributes its two coordinates, each read against its own axis (see Values and Tolerances (§11.1.4)). A fitted trend line through the cloud is reference ink (see Multiple Charts and Reference Ink (§11.1.3)).

How scatter points, whiskers and boxes appear in the chart table.

A scatter chart MUST give one row per named point, a column for x and a column for y, each headed by its printed axis title.
An error bar's interval MUST be written in the centre's cell, after the centre, as centre (low–high) — 3.60 (3.47–3.73).
A box plot MUST give one row per box: its label, then the low whisker end, first quartile, median, third quartile and high whisker end in axis order, the whisker ends only where drawn, under the header row Low | Q1 | Median | Q3 | High (first cell: the printed label column's title, or empty). These header cells are protocol vocabulary, like the markers [blank] or [logo: …]: the page does not print them, and they are the one deliberate exception to never invent, so that the columns can be read without guessing. They are written exactly as given.

✅ SampleConstructedThe rule applies

Ten named points, one row each, x and y under the two printed axis titles`make_scatter`

The plotted values are whole numbers by construction; on a printed page they would be estimates, judged with a tolerance.

Source page

Expected output

**R&D intensity and operating margin, 2025**
	R&D spend (% of revenue)	Operating margin (%)
Alta	2	8
Borel	3	12
Cormo	4	9
Delta	5	16
Eiger	6	21
Fandor	7	14
Grus	8	24
Helix	9	18
Ionia	10	27
Jotun	11	22

Dot and marker series over categories

Dots drawn per category — markers above bars, one marker per gender per country — are ordinary series: one column per marker series, one value per mark.

✅ SampleThe rule applies

A bar series and two marker series over thirty-three countries`health-profile-europe-2024_p105`

A bar series and two marker series over thirty-three countries, nothing printed on the marks: 33 × 3 = 99 values, each an estimate on the 50-unit grid (tolerance 0.05 to 0.1 by readability). The highlighted EU27 bar is an ordinary row. The value-axis title follows the table.

Source page

Expected output

**Figure 3.9. Cancer mortality in European countries by gender, 2021 (or nearest year)**
	Total	Men	Women
Hungary	310	420	240
Croatia	303	428	225
Latvia	288	435	204
Slovenia	279	386	210
Slovak Republic	272	376	214
Denmark	266	322	228
Estonia	262	405	186
Poland	258	348	198
Lithuania	255	394	181
Czechia	252	334	200
Netherlands	250	306	214
Ireland	246	290	208
Romania	244	332	174
Greece	241	320	166
Germany	239	292	190
EU27	238	303	182
Bulgaria	233	314	168
Portugal	229	312	158
Austria	226	281	178
France	224	293	166
Italy	221	280	174
Belgium	218	279	172
Spain	215	296	146
Cyprus	212	268	162
Finland	208	265	166
Sweden	205	236	182
Luxembourg	201	268	154
Malta	197	254	151
United Kingdom	255	306	216
Norway	222	258	188
Iceland	214	243	196
Switzerland	190	240	153
Türkiye	168	245	106

Age-standardised mortality rate per 100 000 population

Source: Caldrivane Health Observatory, mortality database 2024.

Error bars

The centre is copied when printed and estimated otherwise. The whisker ends are read at the caps against the axis, like any estimate. A whisker whose ends cannot be read honestly is not given an interval.

A checker judges the interval on meaning: it accepts any output that conveys it — 3.60 ± 0.13, 3.60 (3.47–3.73), separate low and high columns, prose — within the tolerance of an ordinary estimate, and fails a centre given with no interval.

✅ SampleThe rule applies

Six bars printing their means`db242_p2`

Six bars printing their means, with footnote markers, and 95% confidence whiskers drawn but not printed. Each cell holds the printed centre, markers included, followed by the whisker ends read against the 0.50 grid. The dashed Adequate intake lines are reference ink: their labels follow the table as text. The whisker symbol before 95% confidence interval. is not text.

Source page

Source: CDC/NCHS, National Health and Nutrition Examination Survey, 2009–2012 — Figure 2 (public domain)

Expected output

Figure 2. Mean total water intake per day among adults aged 20 and over, by sex and race and Hispanic origin: United States, 2009–2012
	Non-Hispanic white	Non-Hispanic black	Hispanic
Men	^1,23.60 (3.47–3.73)	²2.92 (2.78–3.06)	3.33 (3.13–3.53)
Women	^1,22.85 (2.76–2.94)	²2.41 (2.29–2.54)	2.58 (2.49–2.67)

Total water intake (liters)

Adequate intake (3.7 liters)

Adequate intake (2.7 liters)

95% confidence interval. ¹Significantly different from non-Hispanic black, p < 0.05. ²Significantly different from Hispanic, p < 0.05. NOTE: Access data table for Figure 2 at: http://www.cdc.gov/nchs/data/databriefs/db242_table.pdf#2. SOURCE: CDC/NCHS, National Health and Nutrition Examination Survey, 2009–2012.

✅ SampleThe rule applies

Nothing is printed on any mark`climate-assessment-sdm_p7`

centres and whisker ends are all estimates on the 0.5 grid. A three-panel figure, each panel its own table headed by its axis unit (°C): panel (a) has a single bar with no category label; in panel (b), Solar and volcanic drivers and Internal variability are drawn on the zero line, so their centre is 0, with whiskers either side. Panel (c) shows its first three bars, then the rest; its two bracket captions follow its table as text. The page heading and the figure caption below the panels are elided.

Source page

Expected output · excerpt

…

**Observed warming** (a) Observed warming 2010–2019 relative to 1850–1900
	°C
	1.07 (0.88–1.21)

**Contributions to warming based on two complementary approaches** (b) Aggregated contributions to 2010–2019 warming relative to 1850–1900, assessed from attribution studies
	°C
Total human influence	1.07 (0.80–1.30)
Well-mixed greenhouse gases	1.50 (1.00–2.00)
Other human drivers	-0.40 (-0.80–0.00)
Solar and volcanic drivers	0 (-0.10–0.10)
Internal variability	0 (-0.20–0.20)

(c) Contributions to 2010–2019 warming relative to 1850–1900, assessed from radiative forcing studies
	°C
Carbon dioxide	0.79 (0.52–1.25)
Methane	0.51 (0.30–0.83)
Nitrous oxide	0.09 (0.05–0.15)
…

Mainly contribute to changes in non-CO₂ greenhouse gases

Mainly contribute to changes in anthropogenic aerosols

…

✅ SampleThe rule applies

Five bars with confidence whiskers, nothing printed`ac8b3538-en_p131`

five rows, negatives included, each with its interval. The whisker ends are read at the drawn caps against the printed axis. The GHG-intensive occupations bar sits almost on zero: a relative tolerance means little there, so a checker widens it (0.3) rather than tightening the reading.

Source page

Expected output

**Annex Figure 2.C.13. The incidence of job strain is lower in new and emerging and green-enhanced skills green occupations even when controlling for the job characteristics** Marginal effect of the job strain indicator on the share of green-driven occupations, controlling for main individual and job characteristics, 2021 for Ostrevane Union members and 2015 for Borvika
	%
Green-driven occupations	-8.7 (-15–-2.6)
Green new and emerging occupations	-11.5 (-17.2–-6.6)
Green-enhanced skills occupations	-25 (-32.1–-19)
Green increased demand occupations	17 (7.2–26.6)
GHG-intensive occupations	-1 (-9.3–7.3)

Note: The chart reports the point estimate (and 95% confidence intervals) of the percentage difference in the incidence of each type of green-driven occupation between job strained and not strained workers. Estimates come from a linear regression with the inverse hyperbolic sine of the share of each occupation type as dependent variable and a job strain dummy, occupation shares, education, gender, age, sector and country dummies as explanatory variables. The sample excludes Agriculture.

Box plots

A box's median, quartiles and drawn whisker ends are read from the geometry. Statistics printed beside the plot are text on the page and are copied as printed.

✅ SampleThe rule applies

Two box plots`charpy_verification_p24`

Two box plots, both over the category HH-103, each its own table under its own figure caption, which tells them apart. Each box gives five readings against its axis; the printed annotation box and the labels of the dashed outlier-limit lines follow the table as text. The individual test values (crosses) and the mean marker carry no name and are not rows.

Source page

Expected output

The same procedure was applied to the second set of specimens. As before, no single test value was excluded before the limits were computed.

**Figure 16 - Box-and-whisker plot for KV values from KLST specimens of HH-103.**
	Low	Q1	Median	Q3	High
HH-103	5.00	5.12	5.18	5.29	5.38

KV (J)

Upper outlier limit

Interquartile range: 0.1775 J 1.5 × Interq. range = 0.26625 J Upper outlier limit = 5.55625 J Lower outlier limit = 4.84625 J Number of outliers = 0

Lower outlier limit

**Figure 17 - Box-and-whisker plot for KV values from full-size specimens of HH-103.**
	Low	Q1	Median	Q3	High
HH-103	13.78	14.26	14.45	14.67	15.61

KV (J)

Upper outlier limit

Interquartile range: 0.42 J 1.5 × Interq. range = 0.63 J Upper outlier limit = 15.30 J Lower outlier limit = 13.62 J Number of outliers = 1

Lower outlier limit

Sample sources1

Public-domain U.S. federal government source material.

CDC/NCHS, National Health and Nutrition Examination Survey, 2009–2012 — Figure 2

Rule id charts-scatter-and-error-bars · charts types

Scatter

Dot and marker series over categories

Error bars

Box plots

Neighbouring rules, and how they differ

More