Presentation of Statistical Tables

1
Presentation of Statistical Tables

• Bratislava, 5-7 May 2003
• Angela Me
• Statistical Division UNECE

2
Presentation of Statistical Tables Summary

• Metadata
• Accuracy
• How to present gender differentials

3
Presentation of Statistical Tables

• We should satisfy the needs of all types of
  readers
• those who do not understand or are not interested
  in statistical issues
• those who understand statistical issues but are
  more interested in reading the message
• those who are interested in how the data were
  produced

4
Presentation of Statistical Tables

• The reader should be able
• to easily understand the message from the data
  and the graphics
• to find all the information on accuracy, sources
  and methodologies used

5
Presentation of Statistical Tables

• The reader should NEVER be mislead by the data

6
An example
7
An example
8
An example
9
Statistical Tables

• How can the presentation help the readers
  understand the message
• Simplicity
• Limited number of topics
• Limited number of cross-classification
• Presentation of both percentages (when relevant)
  and totals

10
Metadata

• Metadata
• Sources
• Institution
• Method of data collection title, year
• Publication
• Methodologies
• definitions,
• methods of calculation
• Explanations (graphics)
• Special cases (footnotes)

11
Accuracy

• the reader should get the message that the data
  presented do not have the same degree of accuracy
  and should not be equally trusted

12
Accuracy

• Include qualifiers
• ex strong, fair, weak
• present intervals
• highlight estimations
• Limit the decimals for small figures

13
Presentation of sex disaggregated data

• What data should be presented
• total numbers
• Percentages
• by categories within total women and men
• Sex distribution

14
An example Population ages 15 and over by
educational level in China 1990 Numbers in
1,000s, percentage distribution and sex
distribution ()
15
Presentation of sex disaggregated data

• How data should be presented
• Women and men should always be presented side by
  side
• Numbers and percentages should be rounded off to
  facilitate comparisons between women and men
• Graphics should start at zero at the y-axis for a
  quantitative variable

16
Presentation of sex disaggregated data

• How data should be presented
• Sex distribution within categories is best
  illustrated in a graph
• Pie charts should be avoided when distributions
  for several groups are compared

17
Presentation of sex disaggregated data

• More information
• B. Hedman, F. Perucci, P. Sundstrom Engendering
  Statistics a Tool for Change, Statistics Sweden.
• Contact Mr. Pehr Sundstrom (pehr.sundstrom_at_scb.se
  )
• Handbook for Producing National Statistical
  Reports on Women and Men, United Nations
  publications Sales No. E.97.XVII.10