Statistics Explained

Tutorial:Table title guidelines - writing titles and short descriptions

This is the stable Version.

Revision as of 10:12, 30 October 2023 by Grinsai (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

This tutorial provides guidelines to statistical domain managers for writing titles of tables (datasets and main tables) available in Eurostat’s online database. Titles (and subtitles) of tables are important structural metadata that guide users through Eurostat's online database and are indispensable elements that explain table contents. Short descriptions of predefined tables are important reference metadata, providing basic information about the table contents. Clearer and more complete titles and short descriptions improve the accessibility to Eurostat data considerably.

Guidelines for table titles

Sources

Please note that these guidelines also take into account specific Eurostat needs, and the impact that the proposed changes would have on current practices.

Clarity

Titles should be short, precise and easily understandable for most users (abbreviations which are not common knowledge are avoided).

Context

Titles have to be seen in two contexts:

  • Within the context of the data navigation tree:
- tables are shown in the context of the domain to which they belong
  • Without the context of the data navigation tree:
- a table is opened by a browser (Data browser)
- a table is taken from the bulk-download

Titles should be understandable and meaningful in both contexts.

Content of the table title

Statistical indicator

The first information of a title should be the statistical indicator. Breakdowns and other secondary information such as periodicity should come in a second instance (see below for details).

If the table contains one indicator, this indicator should be in the title.

If the table contains several indicators, this should be made understood in the title. An appropriate generic term should be found; a listing of all indicators available in the table must be avoided.

Example:

Dataset: inn_cis8_spec
Indicators (within the dictionary INDIC_IN):
  • Enterprises that developed goods innovation (INPDGD)
  • Enterprises that developed service innovation (INPDSV)
  • Enterprises that developed process innovation by improving logistics, delivery or distribution methods (INPSLG)
  • Enterprises that developed process innovation by improving methods of manufacturing or producing (INPSPD)
  • Enterprises that developed process innovation by supporting activities for processes (INPSSU)
  • etc.
Title: Enterprises by specific types of innovation, NACE Rev. 2 activity and size class.

Important information to be provided to the user but too long to be included in the title should be reported in the metadata (ESMS); for predefined tables, such information might be included in the short description.

Breakdowns

Breakdowns (e.g. sex, age, marital status, economic activity, size class, etc.) are introduced in a title with ‘by’; the ‘by’ is only used once and not repeated in the title:

Wrong: Population by age and by sex
Correct: Population by age and sex

The most relevant breakdowns or those distinguishing one table from other tables with the same indicator(s) should only be mentioned. It is neither necessary nor meaningful to list all available breakdowns of a table in the title.

The positions of breakdowns should not be listed in the title as these are displayed when opening a table. Example: there is no need to mention 'single', 'married', 'divorced' for the breakdown ‘marital status’.

Ranking of items in a title

The way the title is written in terms of content and ranking of the information should fit the two following situations.

A title should first allow the user to understand easily what the table is about. Second, within a table list the reader should easily find the table he/she is looking for and see the differences amongst the tables.

To satisfy the first requirement, the title should be clear and comprehensive even if partial (in terms of dimensions for instance – dimensions are evident as soon as the table is opened with a browser).

In the second case, the eye of a reader should first catch the group of tables he/she is interested in and then the relevant distinguishing breakdown(s). Therefore it is recommended to put the main indicator in the first position, followed by the breakdowns and other information.

For big tables, the number of breakdowns should be adapted according to the context and within a reasonable length of the title (< 149 characters).

Examples:

Emigration by sex and age
Emigration by sex, age group and citizenship
Emigration by sex, age group and country of birth
Emigration by sex, age group and country of next usual residence

Make the titles easy to read, e.g. in a domain where several tables have very similar content and only one item distinguishes the titles, the distinguishing item should be put at the end of the title. This increase the readability within a list of tables.

Further detailed examples of ranking of information in a table title: see summary and examples.

How to express selected breakdowns: see selected breakdowns.

Unit

General principles for units

  • Start with lower case unless it is at the beginning of a sentence (exception in subtitles where it should start always with lower case). Grammar rules for lower/upper case in German have to be applied.
  • Use SI (Système international d'unités) symbols for international metric units: cm3, kg, km, km2, m3, mm.
  • Avoid other symbols/abbreviations, and especially if their meaning is not clear, the target public is not specialised or if the FR-DE equivalents are not commonly known.
  • Move information about seasonal adjustment to the title or the general description.
  • Keep units short.
  • Number in the context of the unit should be written in letters (e.g.: thousand). This also applies if the unit has to be inserted in the title in order to distinguish two datasets/tables which differ only for the unit:
Example: tin00059
EN: Number of mobile phone subscriptions – thousand
FR: Nombre d'abonnements aux services de téléphonie mobile – milliers
DE: Zahl der Mobilfunkteilnehmer - Tausend

Monetary units

  • Most of the cases the ISO code is to be used:
when indicating the main unit for a whole table (this is the typical case for table titles and subtitles),
Examples: EUR, thousand EUR, million EUR
Example: Amount in euros, Betrag in Euro, Montant en euros

Percentage

In general, the symbol ‘%’ has to be used in titles.

‘Percentage’ written out should only be used at the beginning of titles. This should be an exception as it is recommended to start titles with the statistical indicator and not with the unit (ex: 'Percentage of total something...').

Wrong: Tables by functions, aggregated benefits and grouped schemes - in percentage of the GDP
Correct: Social protection expenditure by functions, aggregated benefits and grouped schemes - % of GDP

Pay attention to the distinction between ‘%’ and ‘percentage point(s)’.

Other units of measurement

Units of measurement such as kg, km, etc. should be written without dots; they are not closed up to figures and take no plural.

For a list of units of measurement see abbreviations and symbols in common use in the Interinstitutional style guide of the Publications Office.

Please note: for subtitles, units of measurement might be spelled out in letters in order to increase the readability, e.g. tonnes.

  • Nominative units of measurement: “%”, not: “in %”
  • Use “billion” instead of “thousand million”
  • Use “thousand” in subtitles only when it literally means “one thousand”. Ex: “per thousand inhabitants”

Consistency

Titles should be consistent in their environment and hierarchy within the data navigation tree. i.e. they should be consistent with the titles of the folders to which they belong and with the titles of other tables within the same folder and domain. Titles should be consistent between national and regional tables.

The terminology used should be consistent with the one used in Euro SDMX Metadata Structure (ESMS).

Titles should – to the extent possible – be consistent with the terminology of code lists used in the tables concerned.

Consistency across domains should be ensured where relevant (e.g. wording of breakdowns applicable to all domains of social statistics).

Titles should be consistent in all three languages, i.e. the translations have to be precise and coherent in terms of structure, order of the words and terminology.

Length

Recommended: < 100 characters

Acceptable: [101 – 149]

To be avoided: > 150 (on the website the title will occupy the whole screen)


Wrong: Expenditure of selected health care functions (HC; types of goods and services produced) by financing agents in health care (HF; entities paying for purchase), in % share of current health expenditure
Correct: Expenditure of selected health care functions, by financing health care agents - % of current health expenditure

Separators

Only commas (,), hyphen (-), colon (:) and round brackets () are allowed in titles.

  • Comma
The comma is the separator that should be primarily used in titles. Where needed, i.e. for long titles, the hyphen, the colon and the round brackets might also be used.
  • Hyphen
For longer titles, the hyphen can be used in order to make the title more readable. Typical cases for using the hyphen are periodicity, country and unit. Lower case has to be used after the hyphen (exception in DE for nouns).
Hyphen should also be used to indicate methodology or survey.
Examples:
- Emissions of greenhouse gases and air pollutants induced by final use of CPA08 products - input-output analysis, ESA 2010
- Coverage rate of municipal waste collection by NUTS 2 regions - pilot project data
- Distribution of income by different income groups - EU-SILC survey
  • Round brackets (parentheses)
In titles, parentheses should primarily be used for the year(s) of data collection, if needed. Other uses: to give complementary information to a word or text.
  • Colon
The use of colon (:) should be avoided. In most of the cases it could be replaced by a hyphen (-).
Wrong: Credit institutions: balance sheet total
Correct: Credit institutions - balance sheet total
If the colon is used, the following rules apply by type of language:
EN: blank only after the colon; no caps after colon.
FR: blank before and after colon; no caps after colon.
DE: blank only after the colon; no caps after colon (unless a noun follows).

Capital and lower case

Use capitals sparingly. They are often employed to excess in commercial and administrative circles. When in doubt use lower case. In table titles in EN and FR, in general only the first word of the title should start with capital. In DE, nouns start with capitals. For further information see the Interinstitutional style guide.

Acronyms and abbreviations

Acronyms and abbreviations which are not common knowledge should be avoided. In cases where they are used, they should be spelled out in the short description on the first use (available for predefined tables only) and in titles if appropriate. No emphasis of the first letters should be made when providing the definition of the acronym.

Wrong: BERD (Business Expenditure on R&D)
Correct: business expenditure on R&D (BERD)

GDP (gross domestic product) is an example of an acronym that can be considered of common knowledge and does not need to be spelled out.

For acronyms related to international organisations and agencies, as well as for other main acronyms, please refer to the Interinstitutional style guide – Annex A4.

For acronyms linked to the EU Commission directorates-general and agencies, see Commission directorates-general and services: official titles.

Specific recommendations

  • Use of 'total' in titles

Do not use the word 'total' in the title if it carries no additional information.

Wrong: Total R&D personnel by sectors of performance (employment), occupation and sex
Correct: R&D personnel by sectors of performance, occupation and sex
  • 'Person' or 'individual'

Wherever the definition does not require the term 'individuals', 'persons' should be used.

  • Classification of economic activities - NACE

The breakdown for NACE activities has to be expressed as follows for NACE Rev. 2 for instance:

EN: by NACE Rev. 2 activity
FR: par activité de la NACE Rév. 2
DE: nach NACE Rev. 2 Tätigkeit

The revision of the NACE has always to be expressed. As general rule there is no need to express NACE positions in titles, unless this is needed to differentiate tables. In this case, NACE branches/divisions/sectors should be expressed between brackets. See also further examples here.

  • Ampersand

Do not use ampersand as a replacement of 'and' unless it is part of an official name, document, etc. (e.g. R&D).

  • Historical data

Datasets with historical data can be grouped in a separate folder. The title of this folder should follow the structure: Subject - historical data.

Wrong: Interest rates: historical time series (irt_h)
Correct: Interest rates - historical data (irt_h)

If necessary, additional information such as the period of reference can be added.

Example: Interest rates - historical data (1996-2005).

See also further information on writing 'time' in time items.

In the case of only one dataset with historical data, this will not require an ad-hoc folder and should be placed below all the datasets of the same domain/sub-domain.

  • Source of data

The source of data is a specific property in the Reference DB. It is an important metadata that it available in the data browsers and in the various outputs going along these data.

In terms of display, in the legacy data browsers (TGM and Data Explorer), the source is presented at the bottom right corner of the table. In the new Data Browser, the source displays in addition directly under the title both in the ‘information’ box and in the ‘action’ box.

In terms of writing, the source should start with the name of the institution/agency providing data (spelled out completely) and be followed by the acronyms reported in brackets. In case of multiple sources, Eurostat should come first followed by the other source(s).

Details of acronyms and correct spelled out names of institutions/agencies are available here: for acronyms related to international organisations and agencies, as well as for other main acronyms, please refer to the Interinstitutional style guide – Annex A4; for acronyms linked to the EU Commission directorates-general and agencies, see Commission directorates-general and services: official titles.

Subtitles

Subtitles are currently available for predefined tables only; they should merely include the unit of measurement. For three-dimensional predefined tables with breakdown by unit, the unit of measurement field will be empty. The nominative unit of measurement is to be used, i.e. ‘%’ and not ‘in %’. Units in subtitles should always start with lower case, even if at the beginning of the sentence (grammar rules for lower/upper case in German have to be applied).

Summary and examples

# Ranking Examples Separator Comments
1 Statistical indicator
  • Population
n/a
  • Titles should begin with the main information included in tables/datasets such as GDP, population, electricity consumption, etc.
2 Breakdowns
  • Population by age and sex
  • Population by age, males
  • Population by age, sex and legal marital status
  • Enterprise statistics on manufacturing
  • Population by age, sex and NUTS 2 regions
,
  • Not all breakdowns available in a table should be listed systematically. Rather, the title should contain the most important breakdowns, and particularly the breakdowns that distinguish one table from another.
  • Only the breakdown should be indicated; positions of the breakdown available in the table should not be listed in titles unless this is necessary to distinguish one table from another.
3 Periodicity
  • Population by age, monthly data
  • Population, activity and inactivity - annual averages
, or -
  • Most data available in Eurobase are annual data, therefore by default this information should not be indicated in the title. The periodicity is indicated via the TIME dimension which is mandatory for each table.
  • Reference to the periodicity should only be made where indicators are available with different periodicities (e.g. annual and quarterly data).
  • Hyphen can be used with long titles.
4 Country or geographical, economic, political entity
  • Participants by LMP interventions - Ireland


-
  • In general this information is not part of the title. The dimension GEO (or synonym) is mandatory for all tables.
  • Only the few datasets that contain data for one country (or similar entity) require this information.
5 Year
  • Job vacancy statistics by occupation and NUTS 2 regions - annual data (2000-2008)
  • Population by age and sex - monthly data, Austria (2004)
() or ,
  • In general this information is not part of the title. The dimension TIME is mandatory for all tables.
  • Information on year can appear in titles if the data series is limited (e.g. 2000-2008) or if data collection rounds have to be distinguished from each other (e.g. 2004).
6 Units in tables and datasets
  • %
  • Population by age and sex - monthly data, Austria (2004) - thousand persons
-
  • To be included in subtitles only
  • To be added at the end, preferably after a hyphen.

Guidelines for short descriptions

Short descriptions contain the methodological details required for the correct interpretation of the data. They are currently available for around 1 200 predefined tables. The guidelines are the following:

  • The basic characteristics of the indicator should be described (definition) including the description of the nominator and denominator if necessary.
  • Important exclusions of the data coverage should be mentioned.
  • The basic data source should be mentioned (primary and secondary sources).
  • The content should mainly deal with statistical methodology; however, where necessary, it might include specific information about the indicator(s) such as the context of and limits of use, alternative indicators, etc.
  • The short descriptions have to be available in EN, FR and DE and should not exceed 2000 characters (including FR and DE version). It is the responsibility of the domain manager to ensure that the text is correct in terms of grammar and orthography in all three languages.
  • The domain manager has to ensure that the short description is consistent with the terminology used in the table it refers to and that it is consistent between the three languages.

See also