Methodological Issues of Statistics

Introduction

The Union of International Associations (UIA) has prepared statistics on international organizations for many years – notably for inclusion in Volumes 1-4 of the Yearbook of International Organizations. The assembly of statistics in this section (new in 2001) has posed a number of challenges – beyond those of publishing standard tables following the standard form of more recent editions. This note endeavours to clarify some of the issues in achieving the larger set of tables made available for the first time in the statistics section.

Continuity

The data presented in this section draw on statistics presented throughout the 20th century in a variety of ways by the UIA. International organization has evolved to a very significant degree in number and kind during that period. Some of the challenges to continuity are described in Continuity between editions.

Criteria and data collection

Issues relating to types of organization and criteria are discussed separately in Type 1. Issues relating to coverage, sources (and their reliability), and methods of collection and verification procedures are discussed in Editorial practices.

Dating data

There are a number of issues affecting the year with which “annual” data are associated that can lead to confusion:

  • Detection delay: When an organization is created, this information is not necessarily widely publicized and is not necessarily detected in time for immediate registration in the Yearbook database. The information may first pass via other bodies with which the UIA is in contact and may only be confirmed after the UIA then questions these contacts on the existence of the new body. If the organization has little intention or motivation to make its existence widely known, the delay before any mention appears in the Yearbook may be a number of years.
  • Response delay: Contact with an organization may not necessarily result immediately in comprehensive information, especially if the organization has little motivation to be profiled in a widely circulated publication. An adequate profile, with significant details, may only build up over time as the organization accepts, and corrects, what is being presented about it. Since the details on some organizations are relatively static, no amendments may be received until significant changes provoke a response.
  • Cessation of activity delay: Organizations seldom report on their own cessation of activity. The tendency is to pass through a period of decreasing activity. This may be difficult to distinguish from relative inactivity between periodic meetings.
  • Editing delay: Although organizations are questioned annually, the delay between response to a questionnaire and incorporation of the information into the database may be up to 6 months.
  • Publication delay: Although the Yearbook has a very tight publication schedule (six weeks between final editing and bound book), the published book necessarily reflects a situation of many months past. During the period when the Yearbook was only produced every two years, using non-computerized systems, this delay was even more significant.
  • Date attribution: The date of receipt of information is printed with each individual profile. These dates will necessarily reflect a spread of months, if not years. Other dates that may cause confusion are: the year of publication of the volume, in contrast to the year of the edition as indicated on the title page. The latter has tended to indicate a two-year span of the currency of the volume in anticipation of the volume that will replace it. Thus the volume for 2004-05 is published mid-2004 and is replaced by the volume 2005-06, published mid-2005. This practice developed from an earlier period in which the volume was published every 2 years. This shows that data associated with the 2004-05 edition and associated with 2004 in the tables, at best effectively correspond to information from mid-2003 to early 2004. A deliberate effort has been made to clarify any confusion arising from this in the presentation of the tables.

Sources of statistical data

The intention of this volume is to bring together, and reconcile, statistical data produced in a variety of formats throughout the past century. The main sources have been:

  • Yearbook tables (since 1981): These tables have been produced and maintained in a relatively standardized format. Those of more recent years were available in electronic form. Those of earlier years required scanning.
  • Yearbook tables (1972-1980): These tables were produced in more varied form. They required scanning.
  • Yearbook-based tables (1950-1970): Prior to computerization in the early 1970s, tables were produced after publication of the volume and were published in the UIA’s journal Transnational Associations (or its predecessors). They required scanning.
  • UIA Studies: In the 1950s and 1960s the UIA undertook a number of studies of the evolution of international organization, using pre-war publications, in order to present overviews of the evolution of international organizations. These were then published in the UIA’s journal International Associations (or its predecessors) or in relevant volumes. They required scanning. For many years the tables published in the Yearbook have been generated directly from the relevant database. For the purpose of this volume they were converted into spreadsheet format, notably to facilitate generation of graphics.

Reconciliation

The main issues of reconciliation have been:

  • Geopolitical changes (emergence of countries, splitting of countries, merging of countries, name changes)
  • Verification of scanning and OCR
  • Rectification of typographical and other errors in old tables
  • Treatment of discontinuities due to change of structure of data set, notably emergent types of organization (discussed in Continuity between editions)
  • Classes of data omitted from tables of certain years, especially earlier years Integration of data from older UIA diachronic studies not directly related to the Yearbook structure