September 21, 2000

Who's Not Online

Methodology

About this report

This report is built on the findings of an innovative tracking survey of Internet activities, which was designed to get an accurate reading on the impact of the Internet on Americans’ lives. Running almost continuously between March 1, 2000, and August 20, 2000, the daily poll has asked thousands of Internet users not only about what they have ever done online, but also about what they did “yesterday.” Using a daily sample design, this approach measures the scope of Internet activities more accurately than conventional surveys because it focuses on activities that are fresh in respondents’ minds. It also provides new insights into the range of online behaviors that occur daily.

The general demographics of non-computer and non-Internet users are based on the aggregate data of the six months of telephone interviewing conducted by Princeton Survey Research Associates between March 1, 2000 and August 20, 2000, among a sample of American adults, 18 and over, numbering 12,751 respondents. This sample included 6,338 non-users of all types.  For results based on these non-Internet users, one can say with 95% confidence that the error attributable to sampling and other random effects is plus or minus 2.5 percentage points.

The survey results related to details of those who are not online, including the reasons why they are not online, as well as details about those who were online but are not any longer are based on telephone interviews conducted by Princeton Survey Research Associates among a sample of 2,503 adults, 18 years of age or older, in the continental United States during the period April 1-30, 2000. This sample included 1,158 non-Internet and non-computer users. The April survey as well as May and June were conducted using a rolling daily sample, with a target of completing 75 interviews each day throughout the month. The survey in the month of March had a target of 100 interviews each day through out the month.

For results based on an April survey which focused on non-Internet users, one can say with 95% confidence that the error attributable to sampling and other random effects is plus or minus 3 percentage points. In addition to sampling error, question wording and practical difficulties in conducting telephone surveys may introduce some error or bias into the findings of opinion polls

The data on privacy concerns of non-Internet users of all kinds are based on telephone interviews conducted by PSRA from May 19 through June 21, 2000. The survey involved 2,117 persons, 1,017 of whom are Internet users. For results based on these non-users Internet users, one can say with 95% confidence that the error attributable to sampling and other random effects is plus or minus 3 percentage points.

Throughout this report, the survey results are used to estimate the approximate number of Americans, in millions, who engage in Internet activities. These figures are derived from the Census Bureau’s estimates of the number of adults living in telephone households in the continental United States. As with all survey results, these figures are estimates. Any given figure could be somewhat larger or smaller, given the margin of sampling error associated with the survey results used in deriving these figures.

The questions about the non-Internet user population are available as a PDF on our website at http://www.pewinternet.org/