My da|ra Login

Detailed view

metadata language: German English

Electronic Dialogue 2000 (ED2K)

Resource Type
Dataset : Survey and aggregate data
  • Price, Vincent (University of Pennsylvania)
  • Cappella, Joseph (University of Pennsylvania)
Other Title
  • Online Panel Research (Alternative Title)
Publication Date
  • Knowledge Networks, Inc. (Data Collector)
  • ZA:
    • Political Issues
    • Political Attitudes and Behavior
    • Political Parties, Organizations
    • Legal system, Legislation, Law
    • Economic Policy, National Economic Situation
    • Society, Culture
    • Education, School Systems
    • Communication, Public Opinion, Media
    • Public expenditures
    • Person, Personality, Role
  • CESSDA Topic Classification:
    • Mass political behaviour, attitudes/opinion
    • Elections
    • Educational policy
    • Information and communication
    • Psychology
    • Law, crime and legal systems
  • Abstract

    Since the early stages of public opinion research, nonresponse has been identified as an important threat to the degree to which our sample can represent the population we are interested in. Researchers have documented a trend of declining response rate over the years. However, the nonresponse rate becomes a concern only when it introduces error or bias into survey results. One way to estimate nonresponse bias is through imputation. Online panels, which maintain a pool of respondents who are invited to participate in research through electronic means, face unique opportunities as well as challenges with regards to nonresponses and their imputations. Using data from a nation-wide online panel, this paper hypothesizes that nonresponse bias may exist due to the common causes shared between response propensity and opinion placements. After testifying the common causes, imputations are made to estimate the missing values. Lastly, the differences between observed distributions on variables of interest and imputed distributions are made to show the scope of nonresponse biases. This paper finds that nonresponse biases may exist in online panels. First, the theoretical model of nonresponse bias was supported because the commoncause pattern was found in the dataset. In other words, response propensity and opinion items that are of interest appeared to share common causes including mostly demographic variables. Second, imputation analyses show that although most of the differences between imputed and measured opinions do not indicate serious biases, there were few cases in which the differences seemed to be critical. The limitations of this study, especially those of the imputation method, are discussed at the end of this chapter. Suggestions for future research are provided too.
Temporal Coverage
  • 2000-02
Geographic Coverage
  • United States (US)
Sampled Universe
U.S. Citizens, 18+
Sampling Procedure Comment: Probability Sample: Stratified Multistage Sample
Collection Mode
  • Self-administered questionnaire: CAWI (Computer-Assisted Web Interviewing)
Data and File Information
  • Number of Variables: 2487
C - Data and documents are only released for academic research and teaching after the data depositor’s written authorization. For this purpose the Data Archive obtains a written permission with specification of the user and the analysis intention.
All metadata from GESIS DBK are available free of restriction under the Creative Commons CC0 1.0 Universal Public Domain Dedication. However, GESIS requests that you actively acknowledge and give attribution to all metadata sources, such as the data providers and any data aggregators, including GESIS. For further information see
Alternative Identifiers
  • ZA5658 (Type: ZA-No.)

Update Metadata: 2020-10-21 | Issue Number: 21 | Registration Date: 2014-04-10

Price, Vincent; Cappella, Joseph (2014): Electronic Dialogue 2000 (ED2K). Version: 1.0.0. GESIS Datenarchiv. Dataset.