Guide to Social Science Data Preparation and Archiving

Guide to Social Science Data Prep= aration and Archiving

Supported = by The Robert Wood Johnson Foundation

Third Edit= ion, March 2002

Introduction=

<= /o:p>

Over the years, we have found that many investigators= are more than willing to make their data available to others, but they are sty= mied by the task of preparing data for outside use, particularly in terms of complete documentation. We also find that many of the datasets that we rec= eive have one or more problems that might have been prevented had the investiga= tor thought about the issues in advance. This guide is intended to help researchers document their datasets and prepare them for archiving. The guide is writt= en with the assumption that the reader is familiar with basic concepts of computerized data files, such as variables, fields, codes, and so forth. (= For those to whom these concepts are new, a glossary appears at the end of this document.)

While this guide is aimed at those engaged in the tas= k of preparing data for deposit in a public archive, most of the topics covered= here are of interest to anyone creating a dataset. ICPSR welcomes investigators= to contact us at any point in the data collection process to discuss their pl= ans with respect to the preparation of public-use datasets. The guide focuses = on ICPSR and its practices, preferences, and requirements, but we believe that the social science community would benefit if these recommendations were adopt= ed for data deposited with other archives, including the U.S. National Archiv= es.

The guide should not be read as a set of rules to pre= scribe procedures, but rather as a source of information and suggestions. Obvious= ly, every project has its particular needs and idiosyncrasies, and no single document can cover all contingencies. Also, we are mindful that the comput= ing environment changes rapidly and that new techniques for managing, document= ing, and analyzing data are continually being developed. Accordingly, this docu= ment appears in electronic form on ICPSR's Web site (http://www.icpsr.umich.edu/), where it will be updated between printings.= User comments and suggestions are welcome.

The Guide ad= dresses the following topics:

Introduction

v&n= bsp; Costs and benefits of archiving

Importance of a Data Mana= gement Plan

v&n= bsp; Documentation as part of the project plan <= /p>

v&n= bsp; Using integrated software

v&n= bsp; Data entry and documentation as part of pre-tests and pilot studies

Characteristics of "Good" Datasets

v&n= bsp; Dataset integrity

v&n= bsp; Accessibility

v&n= bsp; Codebooks

Ø Data Documentation Initiative (DDI)

Ø Codebook Elements

Technical Aspects of Data= sets

v&n= bsp; File structures

v&n= bsp; Variable names

v&n= bsp; Variable labels

v&n= bsp; Codes and coding

v&n= bsp; Missing data

v&n= bsp; Selecting missing data codes

v&n= bsp; A note on "not applicable" and sk= ip patterns

Data Management During Analysis

v&n= bsp; Master datasets and work files

v&n= bsp; Version numbers

v&n= bsp; Raw data vs. statistical system files

v&n= bsp; Backups

Confidentiality in Archiv= ed Data

v&n= bsp; The principles

v&n= bsp; The practice of protecting confidentiality =

v&n= bsp; How to handle indirect identifiers

Preparing Data for Archiv= ing: What ICPSR Requires

v&n= bsp; Format of data files

v&n= bsp; Documentation

v&n= bsp; Mode of transmission

v&n= bsp; ICPSR data deposit form

v&n= bsp; Collection evaluation and review

Related Publications and References

From The Inter-university Consortium for Politi= cal and Social Research (ICPSR).