The Graduate School of Political Management
Final IPDI report 2/11/04 5:44 PM Page a1
Final IPDI report 2/11/04 5:44 PM Page a2
Released February 5, 2004
Final IPDI report 2/11/04 5:44 PM Page a3
TABLE OF CONTENTS
Acknowledgments
Executive Summary
Table of Figures
Introduction
How We Conducted the Study
Defining Onli...
This study is a collaboration between the Institute for Politics, Democracy & the Internet and several partners who
provid...
2
EXECUTIVE SUMMARY Released February 5, 2004
A new community of citizens online is defining the 2004 presidential campaig...
Online Political Citizens are significantly more likely to donate money to
candidates. At this early stage of the 2004 cam...
RoperASW survey researchers pioneered the concept of “Influentials” based on
60 years of opinion research. They found that...
5
ex-hippies or, as one news story would have us believe, love-spurned youths who
have pulled up stakes to join the politi...
About one-quarter of the OPCs gave through the Internet (24%), compared to
2% of the general public.
ONLINE POLITICAL ACTI...
7
HOW WE CONDUCTED THE STUDY
The Institute for Politics, Democracy & the Internet commissioned parallel
surveys on the Int...
8
TABLE OF FIGURES
In the Executive Summary
Figure 1. Influentials
Figure 2. The 11 Influentials Questions: In the past ye...
9
INTRODUCTION
This report is about the cadre of politically active citizens who have used the
Internet to transform the 2...
10
Influentials share five major personality characteristics:
• They have an activist approach to life.
• They are connect...
11
HOW WE CONDUCTED THE STUDY
The Institute for Politics, Democracy & the Internet, together with Nielsen//
NetRatings and...
12
(a) made a contribution to a candidate or political organization online,
(b) received political e-mail,
(c) forwarded o...
13
Served as an officer for some club or organization.
Attended a political rally, speech or organized protest of any kind...
14
the ability to analyze problems and express positions in ways that are
persuasive to others.”iv
Influentials’ active mi...
15
One of our central findings is that Online Political Citizens are disproportionately
more likely to be Influentials. Ne...
16
acquired the experience or social standing to exert influence by holding office,
serving on local committees, or becomi...
17
both OPCs and Influential. We then compared them to the Influentials from the
telephone survey data (n = 118)—These are...
18
DEMOGRAPHIC CHARACTERISTICS
The “digital divide” is of abiding concern to researchers and policymakers
because of the p...
19
Online Political Citizens are a mix of old and new activists. However, a large
minority of OPCs have not been political...
20
Sources: U.S. Census data, 2002 estimates. Pew Internet & American Life Project Tracking Survey, March to May 2002.
Nie...
21
ARE THEY DISPROPORTIONATELY DEMOCRATS?
In the total sample of respondents to the online survey, there is a slightly gre...
22
Senator John McCain’s campaign stunned political observers in late 1999 by
collecting $5.6 million online in the month ...
Online Political Citizens are also more likely to donate money than Influentials.
Of Influentials in the general public, d...
24
INFLUENTIALS ONLINE
Influentials are trend-setters, and more likely to adopt new technologies first.
They also are disp...
25
Influentials are much more active online than the general public. They are signifi-
cantly more likely to have visited ...
26
OPCS ONLINE
As expected, Online Political Citizens are very active online. Their online political
activities far outstr...
27
Next, we can look at Online Political Citizens in the online survey data. This
provides a larger group of respondents a...
28
Perhaps most interesting about Online Political Citizens is their involvement in
activities that even other Internet us...
29
Online Political Citizens are heavy media consumers. In general, they do not
report very different media usage from oth...
30
We also find it noteworthy that the amount of time online is not very different
between Online Political Citizens and o...
31
Figure 16. Weekday radio use of Online Political Citizens and all other
online survey respondents
Figure 17. Weekday ne...
32
Figure 18. Weekday magazine use of Online Political Citizens and all other
online survey respondents
Figure 19. Weekday...
33
Finally, we conducted two statistical analyses using simple linear regression to
look at the relationship between media...
34
CONCLUSION
Online Political Citizens are a small group of highly politically involved citizens
whose influence, we beli...
35
Online activities take time and effort to muster. Web designers know that a site
needs time to get organized, debugged ...
36
The Online Political Influentials project was envisioned as parallel and comple-
mentary surveys. An extensive online s...
37
gather data and track trends. Therefore, there is little opportunity for multiple
follow-up calls, and the response rat...
38
Throughout this report we talk about relationships and compare different groups. In
each comparison, a statistical test...
39
NOTES
1
Ed Keller and Jon Berry, The Influentials (New York: The Free Press, 2003), book jacket.
Keller is CEO and Berr...
40
culled from the Nielsen//NetRatings survey were determined using three questions. This
depresses the overall percentage...
41
ROPERASW SURVEY (1,029 COMPLETED INTERVIEWS)
PERCENTAGES FOR THE ROPER DATA ARE WEIGHTED
Interviewer enter respondent s...
42
65 to 69 52 3.8
70 to 74 53 4.7
75 and over 92 7.6
Refused 55 4.3
What is the last grade of school you completed?
Less ...
43
I’m going to read you a list of things some people do about government or politics. As I read each
one, please tell me ...
44
NIELSEN//NETRATINGS SURVEY (1,392 COMPLETED INTERVIEWS)
What is your age?
18 to 24 178 12.8
25 to 34 296 21.3
35 to 54 ...
45
What is the total annual household income before taxes for all household members combined?
$0 to $14,999 87 6.3
$15,000...
46
Other third party 16 1.1
Don’t know 97 7.0
Thinking about an AVERAGE WEEKDAY, how much time do you spend with each of
t...
47
3 to 5 hours 163 11.7
More than 5 hours 205 14.7
Other than for e-mail, how often are you online?
Every day 1,020 73.3
...
48
Below is the list of Web sites from which Nielsen//NetRatings recruited respondents to the online
survey. Our intention...
49
MSNBC Web site politics page msnbc.com/news/politics_front.asp 2
Slate magazine slate.msn.com 1
Lost Angeles Times onli...
50
In several instances we conducted simple linear regressions, emphasizing only those relationships
significant at p < .0...
51
Regression analysis summary for media variables predicting whether respondent is an Online
Political Citizen (Nielsen//...
Institute for Politics, Democracy & the Internet
The Graduate School of Political Management
The George Washington Univers...
of 56

Political Influentials - IPDI report

Published on: Mar 4, 2016
Source: www.slideshare.net


Transcripts - Political Influentials - IPDI report

  • 1. The Graduate School of Political Management Final IPDI report 2/11/04 5:44 PM Page a1
  • 2. Final IPDI report 2/11/04 5:44 PM Page a2
  • 3. Released February 5, 2004 Final IPDI report 2/11/04 5:44 PM Page a3
  • 4. TABLE OF CONTENTS Acknowledgments Executive Summary Table of Figures Introduction How We Conducted the Study Defining Online Political Citizens How Many People Are Online Political Citizens? Online Political Citizens Are Influentials Demographic Characteristics Are They Disproportionately Democrats? Political Contributions Online Political Activities Influentials Online OPCs Online What Is Their Media Usage? Conclusion Methodology Notes Appendix A: Survey Results Appendix B: Web Sites Used for the Online Survey Appendix C: Linear Regression Information 1 2 8 9 12 12 11 11 18 21 22 24 29 34 36 39 41 48 50 24 26 Final IPDI report 2/11/04 5:44 PM Page a4
  • 5. This study is a collaboration between the Institute for Politics, Democracy & the Internet and several partners who provided funding and expertise. The Institute is the premier research and advocacy center for the study and promotion of online politics in a manner that encourages citizen participation and is consistent with democratic principles. For more information, visit www.ipdi.org. Nielsen//NetRatings provided technical support and expertise in the use of its WebIntercept survey research system, an online system that draws respondents from a panel of 1.5 million people. Nielsen//NetRatings is a global Internet audience measure- ment and analysis firm. For more information, visit www.nielsen-netratings.com. RoperASW provided technical expertise for a national telephone survey. RoperASW, an NOP World company, is a global research and consulting firm. For more information, visit www.roperasw.com. MSNBC.com and Slate.com provided funding. MSNBC.com, an Internet news site, is a joint venture, along with a 24-hour cable news channel, of Microsoft and NBC. Slate.com is an online news magazine covering news, politics and culture. For more information, visit www.slate.com. Joseph Graf, IPDI Project Director, and Carol Darr, IPDI Director, wrote this report. We welcome your comments or inquiries at jgraf@gwu.edu or darr@gwu.edu. Invaluable assistance was provided by Institute Deputy Director Julie Barko. Research Assistant Kevin Wells and Executive Assistant Glen Vierk helped analyze data. Research Assistants Laetitia DeWeirdt, Brandon Robinson, Diana Xiong and Matthew Zablud assisted throughout the project. We also thank Research Director Michael Cornfield for his suggestions on the survey design and the report, and Senior Advisor Robert Hoopes for advice on its public release. Jonah Seiger, a veteran online political strategist and visiting fellow at the Institute, was a consultant to the project and was instrumental in the design and funding of this study. We would like to thank Tom Acquaviva and Marc Ryan of Nielsen//NetRatings, Brad Fay of RoperASW, Cherylynne Crowther and Kyoo Kim of MSNBC, and Cyrus Krohn of Slate. We offer special thanks to Jason Levin, an Internet analyst at Nielsen//NetRatings, for his technical help. Textual and graphic design were created by Min Koh, with editorial assistance from Judith Dollenmayer, both of Reingold Inc. of Washington, D.C. The Institute is funded by a grant from The Pew Charitable Trusts and administered by The George Washington University Graduate School of Political Management. The principal investigator is F. Christopher Arterton, Ph.D., Dean of the school. 1 ACKNOWLEDGMENTS Final IPDI report 2/11/04 5:44 PM Page 1
  • 6. 2 EXECUTIVE SUMMARY Released February 5, 2004 A new community of citizens online is defining the 2004 presidential campaign. These citizens are Internet-oriented and politically energized, and they support their candidates by visiting their Web sites, joining Internet discussion groups, reading political Web logs and making political contributions over the Internet. Even before the first primary, they played a pivotal role in the campaign, and they may be harbingers of permanent change in American politics. WHO ARE THESE PEOPLE? To answer that question, the Institute for Politics, Democracy & the Internet of the Graduate School of Political Management at The George Washington University, together with RoperASW and Nielsen//NetRatings, conducted an innovative study of these people, whom we call Online Political Citizens (OPCs). We conducted two parallel surveys, an online survey to look in-depth at this unique community, and a national telephone survey to confirm the findings. Our research found that: Online Political Citizens are not isolated cyber-geeks, as the media has por- trayed them. On the contrary, OPCs are nearly seven times more likely than average citizens to serve as opinion leaders among their friends, relatives and colleagues. OPCs are disproportionately “Influentials,” the Americans who “tell their neighbors what to buy, which politicians to support, and where to vacation,” according to Ed Keller and Jon Berry, authors of the book, The Influentials. 1 Normally, 10% of Americans qualify as Influentials. Our study found that 69% of Online Political Citizens are Influentials. About 44% of Online Political Citizens have not been politically involved in the past in typical ways—they have not previously worked for a campaign, made a campaign donation or attended a campaign event. Online Political Citizens are twice as likely as members of the general public to have a college degree; they have higher incomes, are slightly younger, and are more likely to be white, single and male. Internet users in general lean in these directions, but the difference between OPCs and average Americans is more dramatic. Final IPDI report 2/11/04 5:44 PM Page 2
  • 7. Online Political Citizens are significantly more likely to donate money to candidates. At this early stage of the 2004 campaign, 46% have donated to a candidate or political party in the last two to three months, compared to 10% of the general public. E-mail is their lifeline: 87% receive political e-mail and 66% forward political e-mail to friends and colleagues. OPCs frequent political Web logs, political discussion groups and political chat rooms much more often than the general public. We estimate that Online Political Citizens comprise about 7% of the population. ONLINE POLITICAL CITIZENS ARE INFLUENTIALS The major finding of this study is the high percentage of Online Political Citizens who qualify as Influentials, or opinion leaders and trendsetters with their friends and neighbors. We found that 69% of OPCs are Influentials. This means that OPCs are disproportionately likely to exert a “multiplier effect” outward to the public at large. Keller and Berry call Influentials “canaries in the mineshaft for looming political ideas” who have predicted past elections. The authors argue that “If word of mouth is like a radio signal broadcast over the country, Influentials are the strate- gically placed transmitters that amplify the signal, multiplying the number of people who hear it.” 2 3 Sources: Data for the general public (and Internet users) from RoperASW national telephone survey Nov. 21-23, 2003 (n = 1,029). Data for Online Political Citizens from Nielsen//NetRatings online survey December 2003 (n = 491). Figure 1. Influentials Final IPDI report 2/11/04 5:44 PM Page 3
  • 8. RoperASW survey researchers pioneered the concept of “Influentials” based on 60 years of opinion research. They found that, while Influentials serve as opinion leaders across the board, their common denominator is involvement in civic and political activities. RoperASW devised a set of 11 questions to determine whether someone is an Influential. When we asked the Online Political Citizens this set of questions, 69% qualified as Influentials, and they significantly outpolled the general public on nine of the 11 indicators of civic and political participation. In the past year: 68% of OPCs wrote or called a politician at the local, state or national level, compared to 14% of the general public. 48% of OPCs attended a political rally, speech or organized protest, com- pared to 6% of the general public. 47% of OPCs attended a public meeting on town or school affairs, compared to 24% of the general public. 39% of OPCs wrote a letter to the editor or called a live radio or TV show to express their opinion, compared to 12% of the general public. These indicators suggest that Online Political Citizens have strong ties to their communities. They actively participate in local institutions, hold positions of responsibility and, like most Influentials, have strong opinions they do not hesitate to share. The data belie their reputation as isolated techies, aging 4 Sources: Data for the general public from RoperASW national telephone survey Nov. 21-23, 2003 (n = 1,029). Data for Online Political Citizens from Nielsen//NetRatings online survey December 2003 (n = 491). Figure 2. The 11 Influentials Questions: In the past year, have you … ? Final IPDI report 2/11/04 5:44 PM Page 4
  • 9. 5 ex-hippies or, as one news story would have us believe, love-spurned youths who have pulled up stakes to join the political equivalent of the Foreign Legion. 3 DEMOGRAPHIC CHARACTERISTICS Online Political Citizens are disproportionately male, younger, more highly edu- cated, richer, and more likely to be white. The population of Internet users gener- ally skews in these directions, but the differences between OPCs and the general public are even more dramatic: 62% of OPCs are male and 38% were female, compared to 49% male and 51% female in the general public. 36% are between the ages of 18 to 34, compared to 24% of the general public. 59% have college degrees, compared to 26% of the general public. 42% have incomes over $75,000, compared to 18% of the general public. 34% are single, compared to 28% of the general public. 86% are white, compared to 81% of the general public. Democrats are more likely to be Online Political Citizens than Republicans. While slightly more Democrats responded to our online survey (36% of the sam- ple compared to 33% who are Republicans), they were much more likely to be OPCs. About half of the Democratic respondents are OPCs (49%), compared to less than one-third of Republican respondents (29%). This is probably due to the greater activity in the Democratic campaigns during the primaries. Many OPCs are newcomers to political participation. Online Political Citizens are a mix of old and new activists, but a large minority (44%) have not been politi- cally involved in the past—they have never before worked for a campaign, made a campaign donation or attended a campaign event. 39% of OPCs have made a political donation in a past election, and 17% have donated for the first time in this election. 33% have attended a campaign event in a past election, and 8% have gone to an event for the first time in this election. 27% have worked for a campaign event in a past election, and 5% have worked for a campaign for the first time in this election. POLITICAL CONTRIBUTIONS Online Political Citizens are dramatically more likely to donate money to candi- dates. At this early stage of the 2004 campaign, 46% have donated to a candidate or political organization. Only about 10% of the general public have made dona- tions in the past two to three months. Final IPDI report 2/11/04 5:44 PM Page 5
  • 10. About one-quarter of the OPCs gave through the Internet (24%), compared to 2% of the general public. ONLINE POLITICAL ACTIVITIES E-mail is the lifeline of Online Political Citizens. More than half (57%) have signed up for political e-mail from candidates or political parties, and 87% receive it. Two-thirds of those who receive this e-mail then forward it to some- one else. About 21% of OPCs have received e-mail from Meetup.com, the Web site that arranges the local gatherings that have been a crucial element in the campaigns of Howard Dean and, to a lesser extent, Wesley Clark. About a quarter of Online Political Citizens view or post comments on political Web logs (27%) or visit political discussion groups (25%). Very few people in the general public do these activities (less than 4%). ONLINE POLITICAL CITIZENS ARE A SMALL BUT POWERFUL MINORITY We estimate that Online Political Citizens are about 7% of the population, based on a random national telephone survey and Web usage statistics. Sixty-nine percent of Online Political Citizens qualify as Influentials. It’s clear that candidates, political parties and issue advocacy groups who wish to reach the people who influence everyone else must look to the Internet. They will find there a high concentration of opinion leaders and political activists. 6 Sources: Data for the general public from RoperASW national telephone survey Nov. 21-23, 2003 (n = 1,029). Data for Online Political Citizens from Nielsen//NetRatings online survey December 2003 (n = 491). Figure 3. Who donates to political campaigns? Percent who have donated to a candidate or political organization in the past 2-3 months Final IPDI report 2/11/04 5:44 PM Page 6
  • 11. 7 HOW WE CONDUCTED THE STUDY The Institute for Politics, Democracy & the Internet commissioned parallel surveys on the Internet and by telephone. Nielsen//NetRatings and RoperASW, respectively, conducted the two surveys. Nielsen//NetRatings surveyed 1,392 Internet users who visited presidential campaign Web sites, other political sites (such as Web logs and party sites) and online news sites. At the same time, RoperASW conducted a nationwide telephone survey of 1,029 adults, seeking comparable data from a representative sample of the general population. The Nielsen online survey looks in-depth at this unique community, while the Roper national survey confirms those findings. For a thorough discussion of survey methodologies please consult the full report below, or at www.ipdi.org. The Institute for Politics, Democracy & the Internet is funded by a grant from The Pew Charitable Trusts and administered by The George Washington University Graduate School of Political Management. It is the premier research and advocacy center promoting the development of online politics in a manner that encourages citizen participation and is consistent with democratic principles. For more information, visit www.ipdi.org. Nielsen//NetRatings is the global standard for Internet audience measurement and analysis and is a premier source for online advertising research. Covering 70% of the world’s Internet usage, the Nielsen//NetRatings services offer syndicated Internet and digital media research reports and custom-tailored data to help companies gain valuable insight into their business. For more information, visit www.nielsen-netratings.com. RoperASW, an NOP World company, is a leading global marketing research and consulting firm. With headquarters in New York and offices in London, Manila, and throughout the U.S., NOP World is the seventh largest market research company in the U.S. and the ninth largest in the world. Bringing together some of the most renowned U.S. and European research firms in a unified global network, NOP World is a wholly-owned subsidiary of UK-based United Business Media plc. For more information, visit www.roperasw.com. MSNBC.com and Slate.com provided funding. MSNBC.com, an Internet news site, is a joint venture, along with a 24-hour cable news channel, of Microsoft and NBC. Slate.com is an online news magazine covering news, politics and culture. For more information, visit www.slate.com. Final IPDI report 2/11/04 5:44 PM Page 7
  • 12. 8 TABLE OF FIGURES In the Executive Summary Figure 1. Influentials Figure 2. The 11 Influentials Questions: In the past year, have you. . .? Figure 3. Who donates to political campaigns? Percent who have donated to a candidate or political organization in the past 2 to 3 months In the Full Report Figure 4. Influentials Figure 5. The 11 Influentials Questions: In the past year, have you. . .? Figure 6. Influentials in the general public compared to Online Political Citizens (OPCs) who are Influentials Figure 7. Demographic comparison of OPCs to Internet users and the U.S. population Figure 8. Who donates to political campaigns? Percent who have donated to a candidate or political organization in the past 2 to 3 months Figure 9. Online access of Influentials and non-Influentials in the general public Figure 10. Weekday Internet use by Influentials and non-Influentials in the general public Figure 11. Online habits of Influentials and non-Influentials in the general public Figure 12. Online habits: a comparison of OPCs and all other respondents (from RoperASW telephone survey) Figure 13. Online habits: a comparison of OPCs and all other respondents (from Nielsen//NetRatings online survey) Figure 14. Weekday media use among OPCs Figure 15. Weekday TV use of OPCs and all other online survey respondents Figure 16. Weekday radio use of OPCs and all other online survey respondents Figure 17. Weekday newspaper readership of OPCs and all other online survey respondents Figure 18. Weekday magazine use of OPCs and all other online survey respondents Figure 19. Weekday Internet use of OPCs and all other online surveyrespondents Final IPDI report 2/11/04 5:44 PM Page 8
  • 13. 9 INTRODUCTION This report is about the cadre of politically active citizens who have used the Internet to transform the 2004 presidential political campaign. Using cheap, easy and immediate communication, they have found political information, found each other, and found their own political voices. Most news stories written about this phenomenon have focused on the method they use to communicate—the Internet— and the candidates they support. This report is about the citizens themselves, whom we call “Online Political Citizens” (OPCs). The small band of Online Political Citizens who became involved in the 2004 presidential campaign followed in the footsteps of traditional activists, donating time and money to their chosen candidates. But this year they have followed their own path, using the Internet to an unprecedented degree to shape the course of presidential politics. They visit campaign Web sites, donate money online, join Internet discussion groups, and read and post comments on Web logs. These Online Political Citizens also organize local events through Web sites such as Meetup.com or donate money to their causes on sites such as Moveon.org or grassfire.org. OPCs use campaign Web sites as hubs, and they depend heavily on e-mail to stay in touch with the campaigns, receive news stories and muster support. The Institute is interested in the extent to which these people influence those around them and lead public opinion. To study this, we tapped a rich vein of political research about opinion leaders, a line of research dating back 60 years. The RoperASW market research firm coined the term “Influentials” in 1988 to describe a select group of people who influence their neighbors and colleagues in a variety of ways. According to the authors of the book The Influentials, Ed Keller and Jon Berry, Influentials are people who tell their neighbors “what to buy, which politicians to support, and where to vacation.” 4 As sources of news and information multiply, and word-of-mouth recommendations become more important, the authors argue that the huge degree of influence these individuals wield over their friends, relatives and colleagues only grows. Producers of goods and services are intensely interested in Influentials as tastemakers and trendsetters. They are the “holy grail” of marketing, according to Keller and Berry. 5 For the same reasons, people who sell political ideas covet the support of Influentials, whose endorsement is more valuable than the average voter’s. Not only are Influentials more likely to donate money, volunteer time, and write letters to the media, but they will persuade others and bring them along to the polls. Final IPDI report 2/11/04 5:44 PM Page 9
  • 14. 10 Influentials share five major personality characteristics: • They have an activist approach to life. • They are connected to many people and organizations. • They have active minds and diverse interests. • Their expertise on a wide range of subjects gives them an enormous amount of influence. • They are trendsetters. However, the common denominator that identifies virtually all Influentials is their active involvement in civic and political affairs. It is not surprising, then, that a disproportionate number of Online Political Citizens fit the profile of Influentials. It is, however, curious that the public per- ception of Online Political Citizens is so at odds with reality. They have been derided as a fringe group of political geeks or, as one news story suggested, love- spurned youths joining the political equivalent of the Foreign Legion. 6 They have been dismissed as politically insignificant because of their tiny numbers. In fact, all told, we estimate that OPCs comprise perhaps 7% of the adult population. Online Political Citizens have been key in the campaign of former Vermont gover- nor Howard Dean, whose supporters and staff formed the vanguard in seizing the Internet to promote a campaign. When the Dean online campaign began, Web logs publicized the first MeetUp gatherings, which drove curious Internet users to the campaign Web site, which in turn encouraged them to sign up and donate. As one commentator said, “Growth followed an exponential curve; Dean’s new support- ers contributed money, his piles of money won respect from the media, and the media attention pushed MeetUp numbers higher.” 7 Dean’s campaign raised more than $40 million in 2003. Most of the other campaigns have tried to follow suit, although their success has varied. Dean’s early triumphs encouraged followers of former general Wesley Clark to recruit others online with Web sites and discus- sion groups. On the strength of their online support, Clark entered the race. U.S. Representative Dennis Kucinich has also attracted support on the Internet, despite trailing most of the other candidates in the polls. Other candidates, as they have gained followers offline, have seen their Internet support increase. On the Republican side, the Web site of the Bush/Cheney presidential campaign draws a significant amount of Web traffic, which in some months has exceeded any of the Democratic candidates. Republican supporters are active online, and increasingly active in discussion groups, but to date they have not participated as much in MeetUp events or political Web logs. 8 We suspect that this lack of activi- ty may be because President Bush has no primary opponent. Nonetheless, we believe the influence of Online Political Citizens crosses the aisle, and that OPCs who are Democrats have shown greater involvement early in the campaign because of hotly contested Democratic primaries. Final IPDI report 2/11/04 5:44 PM Page 10
  • 15. 11 HOW WE CONDUCTED THE STUDY The Institute for Politics, Democracy & the Internet, together with Nielsen// NetRatings and RoperASW, assessed this new wave of political activism with two parallel political surveys. Nielsen//NetRatings surveyed a select group of people online who used news and political Web sites in December 2003. At the same time, RoperASW conducted a nationwide telephone survey asking comparable questions. The Nielsen online survey looks in-depth at this unique community, while the Roper national survey confirms the findings with a broad, nationally representative sample. Nielsen//NetRatings surveyed 1,392 people online who accessed political and news Web sites from Nov. 26 to Dec. 31, 2003. RoperASW interviewed 1,029 people in a telephone survey using random digit dialing techniques between Nov. 21 and 23, 2003. The Nielsen//NetRatings online survey lets us study people accessing news and politics online. In order to collect a large number of Online Political Citizens, the Nielsen//NetRatings survey was specifically designed to recruit respondents from a variety of news Web sites, including the political news pages of major news outlets and the Web sites of several political publications. The survey also polled respondents from political Web sites, including the official party sites, official sites for each candidate from the two major parties, and a handful of nonpartisan political sites. (For a full list of these sites see Appendix B.) The RoperASW national telephone survey helped test and validate our general conclusions and estimate the number of Online Political Citizens in the general population. Online Political Citizens are a group of people who have defined the early part of the 2004 campaign and will wield great influence in its outcome. OPCs are not merely people online reading the news, nor are they just occasional viewers of a political Web site. OPCs use the Internet in a variety of ways to be politically involved. We defined Online Political Citizens as people who have within the past two to three months (1) visited the Web site of a candidate or political party, and (2) taken part in at least two of the following online political activities: DEFINING ONLINE POLITICAL CITIZENS Final IPDI report 2/11/04 5:44 PM Page 11
  • 16. 12 (a) made a contribution to a candidate or political organization online, (b) received political e-mail, (c) forwarded or sent political e-mail, (d) visited or posted comments on a political Web log, (e) participated in a political chat room, or (f) visited a news Web site for news about politics and campaigns. (For a complete discussion of how we defined OPCs in the data, see Methodology, below.) In defining Online Political Citizens by their activities, we were able to exclude the occasional viewer of a political Web site, as well as those who just read the news online. We were also able to encompass uncommitted activists. Our definition of OPCs includes those who have not yet shown support for a particular candidate. By any estimate, Online Political Citizens comprise a small percentage of the population, and they engage in activities that hold little serious interest for most of the general population. At the end of 2003, relatively few people were following the election closely, but OPCs were intensely interested. As we expected, the Nielsen//NetRatings online survey captured a large number of Online Political Citizens, 491 individuals, or 35% of the sample. Also as expected, the nationwide RoperASW survey captured relatively few OPCs, only 83, or 7% of the sample. 9 We can project the RoperASW estimate to the total adult population in the United States with a margin of error of roughly plus or minus 3%. We consider 7% a rough estimate of the number of Online Political Citizens in the general population. 10 RoperASW researchers have developed a set of 11 questions that separate from the general public people who are more likely to influence their neighbors, recommend choices and lead public opinion. Those who respond affirmatively to at least three or four questions about their civic and political activities (depending on the survey method) are designated as Influentials. 11 Respondents are asked whether, within the last year, they have done any of the following: Attended a public meeting on town or school affairs. Written or called any politician at the state, local or national level. Served on a committee for some local organization. HOW MANY PEOPLE ARE ONLINE POLITICAL CITIZENS? ONLINE POLITICAL CITIZENS ARE INFLUENTIALS Final IPDI report 2/11/04 5:44 PM Page 12
  • 17. 13 Served as an officer for some club or organization. Attended a political rally, speech or organized protest of any kind. Written a letter to the editor of a newspaper or magazine or called a live radio or TV show to express their opinion. Been an active member of any group that tries to influence public policy or government. Made a speech. Worked for a political party. Written an article for a magazine or newspaper. Held or run for political office. RoperASW has found over the years that usually about 10% of the general public qualify as Influentials. How to Recognize An Influential The defining characteristic of “Influentials”—the 10% of Americans “who tell their neighbors what to buy, which politicians to support and where to vaca- tion”—is their activist orientation to life.i Influential individuals wield a huge amount of influence within their communities as leaders and opinion makers, according to Ed Keller and Jon Berry, authors of The Influentials. Influentials are not necessarily the ones you’d expect: They are not your stereotype of who runs the country. They are not the familiar faces portrayed in the mass media as the change agents in society and the marketplace. Many have achieved material success, but they are gen- erally not the richest Americans. They are well-educated but generally not the most educated Americans. They are accomplished in their careers but not at the top of industry. You won’t often see them on the front page of The New York Times or The Wall Street Journal or on the evening network news. They’re more likely to be people who live in your neighborhood.ii Influentials share five major characteristics: Foremost is their activism, which manifests itself through involvement in community civic and political activities. This common denominator cuts across the lives of all Influentials. Influentials are connected. They are the joiners and networkers. They have ties, usually strong ties, to more groups and organizations than the average American. When Influentials make a recommendation, they usually broad- cast it to a large network of friends, relatives and acquaintances. Influentials have influence. “Because they know many people and soak up a large amount of information, Influentials stand out as smart, informed sources of advice and insight. They know a lot about some things and some- thing about a lot of things, and if they don’t know the answer, they probably know someone who does.”iii Others look up to them for advice and opin- ions. Half of Influentials have graduated from college, and one in five has done postgraduate work. As the authors put it, “The Influentials’ high level of education is probably a significant factor in their influence, giving them Final IPDI report 2/11/04 5:45 PM Page 13
  • 18. 14 the ability to analyze problems and express positions in ways that are persuasive to others.”iv Influentials’ active minds seek out information—”In some ways, they are like human parallel processors, taking in information, rating it, sorting it, sifting out the element that’s useful or interesting, and storing it away for future purpose.”v They are also opinionated. They don’t hold back when they find something they like or don’t like. Influentials are trendsetters, often tending to be two to five years ahead of the rest of society. Over the years, RoperASW has found that Influentials have been “predictive of changing political winds, and were, for example, early critics of Richard Nixon.”vi Of particular interest to our study is how Influentials’ trendsetting applies to politics. The picture that comes through is that, in terms of the orthodoxies of the established political parties, Influentials are mavericks. Many are loyal to their political party, but there’s a streak of independent-minded- ness. Large numbers seem to question their parties’ stance on issues.vii Large numbers of Influentials—larger than in the public as a whole— think that traditional interest groups on both sides of issues have too much power in the society today … In Influentials’ thinking, traditional groups, with their traditional thinking and vested interests, often stand in the way of progress. It’s smarter and better to become involved with a group outside the mainstream, they seem to be saying; you can make more of a difference there.viii Finally, being an Influential is fluid. A person may participate in civic affairs early in his or her career, then drop out for a while to care for small children or aged parents.ix As the authors acknowledge, “Just as the U.S. Army has its ‘citizen soldier,’ the nation, it seems, has its ‘citizen leader,’ who becomes involved for a period of time then retires to the pleasures of garden and grandchildren.”x Given Influentials’ strong interest in politics, their maverick streak, and the fact that they are usually educated and technologically sophisticated, we thought it very likely that this profile would fit many participants in online politics in the 2004 primaries. In fact, it fit them to a tee. i Ed Keller and Jon Berry, The Influentials (New York: The Free Press, 2003), book jacket. ii Keller and Berry, 29-30. iii Keller and Berry, 147. iv Keller and Berry, 15. v Keller and Berry, 33. vi Keller and Berry, 136. vii Keller and Berry, 68. viii Keller and Berry, 119. ix Keller and Berry, 199-200. x Keller and Berry, 77. Final IPDI report 2/11/04 5:45 PM Page 14
  • 19. 15 One of our central findings is that Online Political Citizens are disproportionately more likely to be Influentials. Nearly 70% of the OPCs we sampled online quali- fied as Influentials. The individuals involved in online presidential campaign activities are dispropor- tionately likely to exert a “multiplier effect” outward into the general public. Influentials tend to broadcast their recommendations to a large network of friends, relatives, and acquaintances who, because of the Influentials’ persuasive ability and expertise, are likely to pay attention. Moreover, Influentials are asked their opinions about government and politics more than any other topic. 12 Online Political Citizens are more likely than the general public to have taken part in nearly all the civic and political activities that define Influentials. They are seven times more likely than the general public to have attended a political rally, speech or protest in the last two to three months. They are nearly five times more likely to have contacted a politician, three times more likely to have written a letter to the editor, and three times more likely to belong to groups trying to influence public policy. These activities suggest that OPCs are much more heavily involved in politics and civic life than the general public. Many of these activities are quickly and easily accomplished. It takes only a few minutes to e-mail a letter to the editor or e-mail an elected representative. This suggests participating in politics through the Internet facilitates spur-of-the- moment involvements that require less time and fewer personal resources. This may be attractive to younger people and political novices who have not yet Sources: Data for the general public (and Internet users) from RoperASW national telephone survey Nov. 21-23, 2003 (n = 1,029). Data for Online Political Citizens from Nielsen//NetRatings online survey December 2003 (n = 491). Figure 4. Influentials Final IPDI report 2/11/04 5:45 PM Page 15
  • 20. 16 acquired the experience or social standing to exert influence by holding office, serving on local committees, or becoming officers in community organizations. Online Political Citizens are only slightly more likely than the general public to engage in the kinds of civic activity that require roots in a community and a more long-term time commitment. In particular, there is less difference between OPCs and the general public with respect to serving on a local committee or holding or running for political office. This type of political involvement is more likely done by people who are established in their careers and community, who are typically older and with children. This is a slightly different demographic than that of the OPCs, who are younger and more likely to be single. Similarly, there is only a small difference between Online Political Citizens and the general public with respect to writing an article for a magazine or newspaper, or making a speech. These activities bespeak a certain level of professional prestige and authority. In no instance, however, do Online Political Citizens score lower than the general public on any of the 11 civic and political activities. OPCs are more likely than the general public to take part in every category of political involvement. 13 Not all Online Political Citizens are Influentials. Likewise, there are Influentials who are not OPCs (some Influentials are not even online). In order to more closely examine how OPCs are politically involved, we wanted to compare them to Influentials. To do that, we first pulled out from the online survey data those OPCs who are also Influentials (n = 339)—these are the 69% who are Sources: Data for the general public (and Internet users) from RoperASW national telephone survey Nov. 21-23, 2003 (n = 1,029). Data for Online Political Citizens from Nielsen//NetRatings online survey December 2003 (n = 491). Figure 5. The 11 Influentials Questions: In the past year have you … ? Final IPDI report 2/11/04 5:45 PM Page 16
  • 21. 17 both OPCs and Influential. We then compared them to the Influentials from the telephone survey data (n = 118)—These are Influentials from the general public. Our intention is to compare the political activity between OPCs who are Influentials, and a group of more typical Influentials. When the two groups are compared, OPCs are more likely to have contacted politicians, attended a politi- cal rally or worked for a political party. In other words, political involvement differs between people who are active in politics online and people who are politically active by other means. Influentials in the general public are more likely to have engaged in activities that require roots in a community, such as serving on committees, as officers of local organizations, or by attending local public meetings. 14 Moreover, Influentials in the general public are significantly more likely to have made a speech or written an article for a magazine or newspaper—activities that imply authority and pro- fessional standing. It is important to remember that Online Political Citizens are highly involved in all sorts of political activities. Nonetheless, the political activities of OPCs tend toward activities that are part of the current presidential campaign and require less time commitment and fewer personal resources. OPCs seem less likely to engage in activities associated with local involvement. OPCs also seem more inclined to “political” than “civic” involvement. It remains to be seen whether these activists will move on to political involvement that binds them to their local communities in the long term, or if they will stay focused on national and political interests. Sources: Data for the general public from RoperASW national telephone survey Nov. 21-23, 2003 (n = 118 Influentials). Data for Online Political Citizens from Nielsen//NetRatings online survey December 2003 (n = 339 Influentials). Figure 6. Influentials in the general public compared to Online Political Citizens who are Influentials Final IPDI report 2/11/04 5:45 PM Page 17
  • 22. 18 DEMOGRAPHIC CHARACTERISTICS The “digital divide” is of abiding concern to researchers and policymakers because of the positive impact of the Internet on people’s lives. People who are unable or unwilling to use the Internet are disadvantaged economically, socially and politically. They may have fewer opportunities to find jobs and greater diffi- culty learning about and procuring government services. They may have less access to information about health, personal finance and other information that affects quality of life. And they do not enjoy the opportunities for social interac- tion made easier by online communication. The digital divide crosses age, race and educational boundaries. Older Americans are less likely to be online than young people, and minorities are less likely to be online than white people. Those with jobs, higher incomes and better educations also tend to be online. The gap in Internet access between the highly educated and the less educated is particularly large. In the past two years, people who use the Internet have spent more time online doing more things, but the number of people going online for the first time has slowed considerably. About 63% of American adults use the Internet, a propor- tion that barely increased in the last half of 2002. About a quarter of all Americans do not ever access the Internet. 15 Not surprisingly, Internet users are younger, better educated and earn more money than the general public. 16 Compared to the general population, Online Political Citizens are disproportion- ately male: Almost 62% of our sample is male, compared with 50% of Internet users and 50% of Influentials in the general public. (Keller and Berry report that Influentials are gender balanced, about half male and half female.) 17 Online Political Citizens are much more likely to have attained higher education. Nearly 60% say they have a four-year college degree, compared to about a quarter (24%) of all Americans and 37% of all Internet users. There are very few OPCs with only a high school diploma or less (6%). Online Political Citizens also earn considerably more than the general public or Internet users. Sixty-four percent report earning more than $50,000 a year, compared to 44% of Internet users and 32% of the general public. Final IPDI report 2/11/04 5:45 PM Page 18
  • 23. 19 Online Political Citizens are a mix of old and new activists. However, a large minority of OPCs have not been politically involved in the past. Forty-four per- cent of OPCs say they have never before worked for a campaign, donated to a campaign or even been to a campaign event. Fifty-six percent report that they have done at least one of those three things in a prior election. About 39% of OPCs say they have made a donation in a past election, 33% say they have attended a campaign event in a past election, and 27% say they have worked for a campaign in the past. About 17% of OPCs say that in the past two to three months they have donated money for the first time. Eight percent have attended an event for the first time, and 5% have worked for a campaign for the first time. Age is clearly a factor. Survey respondents who say they are just getting involved for the first time tend to be younger people. While Web logs and discussion groups are filled with anecdotes from people who claim they are interested in politics for the first time, empirical evidence has been lacking. Newly engaged citizens may not remain engaged, and this sort of social change takes place over a long time, sometimes years. Because of that, it is hard to determine whether the Internet is attracting more new people to politics than generally enter by more traditional means in an election year. Another way to consider the question is to note that 64% of all the Online Political Citizens who have donated money in the past two to three months also donated in a prior election. That means 36% of the donors are new donors. Of all OPCs who attended a campaign event in the past two to three months, just over half have done so before (55%). And of those who have worked for a campaign in the last two to three months, two-thirds (66%) have done so before. In summary, Online Political Citizens are even more highly educated and wealthy than Internet users. Most significantly, about 60% of OPCs have college degrees, more than double the rate for the general public. OPCs are more likely to be single, although the portion who are married is about the same, and they are more likely to be white. (The relative youth of OPCs means fewer are divorced or widowed.) A majority of OPCs have had some past involvement with politics, but 44% have never before worked for a campaign, made a campaign donation, or been to a campaign event. Final IPDI report 2/11/04 5:45 PM Page 19
  • 24. 20 Sources: U.S. Census data, 2002 estimates. Pew Internet & American Life Project Tracking Survey, March to May 2002. Nielsen//NetRatings online survey December 2003 (n = 491 Online Political Citizens). Some columns do not total 100% because some respondents declined to answer. Figure 7. Demographic comparison of Online Political Citizens to Internet users and the U. S. population Final IPDI report 2/11/04 5:45 PM Page 20
  • 25. 21 ARE THEY DISPROPORTIONATELY DEMOCRATS? In the total sample of respondents to the online survey, there is a slightly greater percentage of Democrats (36%) than Republicans (33%). The sample also includes 20% independents. The remainder are from third parties or did not divulge their party affiliations. However, a much greater proportion of Online Political Citizens claim to be Democrats than Republicans. Half of the OPCs we sampled report they are Democrats (50%), 27% are Republicans and 16% independents. So, despite the fairly even political distribution of the whole sample, a greater proportion of Democrats online qualify as OPCs. This makes sense because of the much more active Democratic presence online early in the campaign, and the greater number of political events directed to likely Democratic voters. A higher percentage of Democrats have donated to a campaign (32% of Democrats compared to 23% of Republicans), donated online (16% compared to 3%) and attended campaign events (14% compared to 2%) in the past two or three months. Democrats were are also more likely to be politically active online, such as by visiting campaign Web sites (60% compared to 43%), visiting political Web logs (15% compared to 8%) and receiving political e-mail (58% compared to 43%) in the past two or three months. Final IPDI report 2/11/04 5:45 PM Page 21
  • 26. 22 Senator John McCain’s campaign stunned political observers in late 1999 by collecting $5.6 million online in the month after McCain’s upset win in the New Hampshire primary. 18 Since then, campaign organizers have tried to harness the Internet’s potential for quickly and easily collecting campaign donations. No campaign has been more successful in 2003 than the Howard Dean campaign. Although the Federal Election Commission is not required to collect information about whether donations have been made online, strategists for Dean report that about half of his donations were collected online. 19 The great promise of online fund raising lies in its low transaction costs. This makes it feasible for political fund-raisers to look to average people for funding. Much past fund raising has focused on big donors because they can write thousand- dollar checks. In the 2000 presidential election, donations of $750 or more provided 75% of the funding for George W. Bush’s campaign and 66% of the funding for Al Gore’s. 20 Online fund raising raises the possibility that candidates can effectively appeal to donors who are more representative of the general public. Online Political Citizens are dramatically more likely than the general public to donate money to candidates. By the end of 2003—before the 2004 primaries were in full swing—about 46% had already donated to a candidate or political organi- zation in the last two to three months. One-half of those donations were made online. By comparison, RoperASW found that about 10% of the general population has donated money to a candidate or political party in the last two to three months, and 17% of those donations were made online. Online Political Citizens are four times more likely than the general public to donate money to political candidates and organizations, and one-quarter of those OPC donations are made online. POLITICAL CONTRIBUTIONS Final IPDI report 2/11/04 5:45 PM Page 22
  • 27. Online Political Citizens are also more likely to donate money than Influentials. Of Influentials in the general public, drawn from the RoperASW telephone survey (n = 118), 31% have made a donation in the past two to three months. At this early stage of the campaign, Democrats are slightly more likely than Republicans to report they have donated to a campaign (32% compared to 23%). Democrats are much more likely, however, to donate money online. Among all donors, nearly half of the Democrats (49%) say they donated online, compared to 11% of Republican donors. 21 23 Sources: Data for the general public from RoperASW national telephone survey Nov. 21-23, 2003 (n = 1,029). Data for Online Political Citizens from Nielsen// NetRatings online survey December 2003 (n = 491). Figure 8. Who donates to political campaigns? Percent who have donated to a candidate or political organization in the past 2-3 months Final IPDI report 2/11/04 5:45 PM Page 23
  • 28. 24 INFLUENTIALS ONLINE Influentials are trend-setters, and more likely to adopt new technologies first. They also are disproportionately white collar workers, who are more likely to have computers and Internet access at both work and home. In the general public, Influentials are more likely to be Internet users, although it is noteworthy that 16% of “typical” Influentials are still not online at all. 22 Influentials are more likely to have either dial-up or high speed Internet access at home, and far more likely to use the Internet at work. As we would expect, they also spend more time on the Internet: almost 60% say they spend at least 30 minutes per day online. ONLINE POLITICAL ACTIVITIES Source: RoperASW national telephone survey Nov. 21-23, 2003 (n = 118 Influentials, 911 non-Influentials). Figure 9. Online access of Influentials and non-Influentials in the general public Figure 10. Weekday Internet use by Influentials and non-Influentials in the general public Source: RoperASW national telephone survey Nov. 21-23, 2003 (n = 118 Influentials, 911 non-Influentials). The question asked was, “Thinking about an average weekday, how much time do you spend on the Internet?” Final IPDI report 2/11/04 5:45 PM Page 24
  • 29. 25 Influentials are much more active online than the general public. They are signifi- cantly more likely to have visited news sites, political sites, discussion groups or chat rooms, and Web logs. In particular, Influentials are sharply more likely than the general public to have received and then forwarded political e-mail to others. In order to further explore the relationship between online political activity and personal influence, we used simple linear regression to look for predictors of whether someone is likely to be an Influential. The biggest predictor in both sets of data is whether someone had sent or forwarded political e-mail to some- one else within the past two to three months. Simply receiving political e-mail was also a significant predictor. 23 This result appeared in both sets of data, so we consider it fairly robust. (Appendix C contains full results of linear regressions.) Forwarding political e-mail is the single greatest predictor of whether someone is an Influential, and it suggests a means by which their influence spreads. A lot of political e-mail is deleted without being seen. However, forwarded e-mail has greater credibility and is read more often than e-mail that comes directly from the source. Political campaigns can encourage e-mail recipients to forward mail, and they can foster forwarding with jokes, photos or short videos. Ultimately, though, the habit of forwarding e-mail lies largely outside a campaign’s control. Source: RoperASW national telephone survey Nov. 21-23, 2003 (n = 118 Influentials, 911 non-Influentials). Questions about online habits refer to the past 2-3 months. Figure 11. Online habits of Influentials and non-Influentials in the general public Final IPDI report 2/11/04 5:45 PM Page 25
  • 30. 26 OPCS ONLINE As expected, Online Political Citizens are very active online. Their online political activities far outstrip those of both the general public and other Internet users. In particular, the use of e-mail is dramatically higher, apparently a key part of their online activity. In addition, OPCs use political Web logs and discussion groups, which everyone else largely ignores. 24 We can make two sets of useful comparisons here. First, the RoperASW data, which can be projected to the general public, shows that Online Political Citizens use the Internet for political purposes much more than others. Almost nine in 10 have visited a news Web site (89%) or received political e-mail (89%). More than half of OPCs have forwarded political e-mail (57%). We also find that almost half have visited a political Web log (47%) and a third have visited a political discus- sion group or chat room (34%), very rare activities among the rest of the public. Source: RoperASW national telephone survey Nov. 21-23, 2003 (n = 83 Online Political Citizens and 946 other survey respondents). Questions about online politi- cal activity refer to the past 2-3 months. Figure 12. Online habits: a comparison of Online Political Citizens and all other respondents Final IPDI report 2/11/04 5:45 PM Page 26
  • 31. 27 Next, we can look at Online Political Citizens in the online survey data. This provides a larger group of respondents and a broader set of online activities to study. The comparison of Online Political Citizens across both surveys is in some cases very close. Between 80% and 90% in both telephone and online surveys report they visited a news Web site for political news or received political e-mail in the last two to three months. In both surveys, roughly 60% say they sent or forwarded e-mail. Visiting Web logs, political discussion groups and chat rooms remain activities for few people other than Online Political Citizens. Less than 5% of non-OPCs in either survey say they visit Web logs or chat rooms. 25 About two-thirds (66%) of Online Political Citizens in the online survey have sent or forwarded political e-mail. By contrast, of all remaining online survey respon- dents, only 15% say they have sent political e-mail on to others. (The results from the Roper telephone survey were largely the same.) 26 Source: Nielsen//NetRatings online survey December 2003 (n = 491 Online Political Citizens and 901 other online survey respondents). Questions about online political activity refer to the past 2-3 months. Figure 13. Online habits: a comparison of Online Political Citizens and all other respondents Final IPDI report 2/11/04 5:45 PM Page 27
  • 32. 28 Perhaps most interesting about Online Political Citizens is their involvement in activities that even other Internet users are unlikely to participate in—visiting Web logs and discussion groups or chat rooms that address politics or current events. Neither of these activities is widespread among the general public. Only between 5% and 7% of all American adults report visiting Web logs or discussion groups, and these numbers can be generalized to the entire adult population. Even among the Internet users we surveyed online who are not OPCs, only between 2% and 3% have visited political Web logs, discussion groups or chat rooms. This is surprising. Even people who were surveyed online (and thus clearly are Internet users) were not more likely to visit these sources of political information. The use of political Web logs, discussion groups and chat rooms clearly separates Online Political Citizens from all others. Even others going online are no more likely to use Web logs and chat rooms than members of the general public. Final IPDI report 2/11/04 5:45 PM Page 28
  • 33. 29 Online Political Citizens are heavy media consumers. In general, they do not report very different media usage from other Internet users we surveyed. In addi- tion to being heavy users of the Internet, our online survey respondents tend to report more use of television and radio than of newspapers and magazines. About 43% of OPCs report they watch at least two hours of television on an average weekday, compared to 48% of the rest of the Internet users. Much of this may be multitasking—or leaving the television on while surfing online or check- ing e-mail. For many people, if they spend three or four hours an evening on the Internet they likewise spend nearly the same amount of time with the television on. Online Political Citizens are simply heavy consumers of media. They are polit- ical junkies, and political junkies tend to be news junkies. Source: Nielsen//NetRatings online survey December 2003 (n = 491). The question asked was, “Thinking about an average weekday, how much time do you spend with each of the following media?” Figure 14. Weekday media use among Online Political Citizens WH AT IS THEIR MEDIA USAGE? Final IPDI report 2/11/04 5:45 PM Page 29
  • 34. 30 We also find it noteworthy that the amount of time online is not very different between Online Political Citizens and other Internet users. This is interesting because of the different ways each group appears to use the Internet. As we saw above, OPCs are accessing Web logs and discussion groups, while other Internet users generally are not. Yet both groups spend roughly the same amount of time online. Almost a third of the Online Political Citizens spend three or more hours a day on the Internet. About 31% report spending three or more hours each weekday online, compared to 24% of the remaining survey respondents. Three-quarters of all OPCs are online each weekday for at least one hour. Online Political Citizens spend more time with newspapers than the other Internet users. About a quarter of all OPCs (28%) say they spend at least an hour with the newspaper each weekday, compared to 20% of other Internet users we surveyed. Source of Figures 15 through 19: Nielsen//NetRatings online survey December 2003 (n = 491 Online Political Citizens, 901 all other online survey respondents). Figure 15. Weekday TV use of Online Political Citizens and all other online survey respondents Final IPDI report 2/11/04 5:45 PM Page 30
  • 35. 31 Figure 16. Weekday radio use of Online Political Citizens and all other online survey respondents Figure 17. Weekday newspaper readership of Online Political Citizens and all other online survey respondents Final IPDI report 2/11/04 5:45 PM Page 31
  • 36. 32 Figure 18. Weekday magazine use of Online Political Citizens and all other online survey respondents Figure 19. Weekday Internet use of Online Political Citizens and all other online survey respondents Final IPDI report 2/11/04 5:45 PM Page 32
  • 37. 33 Finally, we conducted two statistical analyses using simple linear regression to look at the relationship between media use and whether someone was either an Online Political Citizen or an Influential. (This could only be performed with questions in the Nielsen//NetRatings data.) Of course, using the Internet is a pre requisite for being an Online Political Citizen. However, we also found that newspaper use is a strong positive predictor of whether someone is an OPC, and television is a strong negative predictor. In other words, when we look at all five media together—radio, television, newspapers, magazines and the Internet—the more time someone spent reading the newspaper, the more likely he or she would qualify as an OPC. The more time spent watching television, the less likely that person would qualify as an OPC. In fact, watching television was the strongest indicator, albeit a negative one, among all types of media use—including the amount of time spent on the Internet. Finally, we looked at the relationship between types of media and whether someone was an Influential. We found that newspaper use is the single strongest predictor of being an Influential, and television use was the strongest negative predictor. These relationships are fairly typical findings in communication research. However, use of the Internet generally did not predict whether someone would be an Influential. Despite the availability of information and ease of communication online, spending time with the daily newspaper (NOT Internet use) remains the strongest media indicator of whether someone will be an Influential. As we said, this is also true if we examine only Online Political Citizens. (Results were deter- mined by simple linear regression. Appendix C shows full results.) Final IPDI report 2/11/04 5:45 PM Page 33
  • 38. 34 CONCLUSION Online Political Citizens are a small group of highly politically involved citizens whose influence, we believe, stretches far beyond their numbers. They are dispro- portionately Influential and highly communicative. There is strong evidence in this study that these people are not out of touch with mainstream political activism. They participate in many of the same community and neighborhood activities as other Influentials. A majority have been involved in politics in the past. At the same time, this year’s political involvement is clearly different. It is more focused on the campaign and online communication, and less focused on local political involvement and long-term commitments to the community, such as running for local office or serving on a local committee. We do not want to blur a key point: Online Political Citizens are dramatically more involved than typical citizens. At the same time, when we compare them to Influentials in the general public, they have weaker ties to their local community and to long-term obligations. Part of this is probably due to the relative youth of OPCs and to the 2004 campaign. Online Political Citizens are a little younger than the general public, but a lot more highly educated and wealthy. They are also more likely to be male. The key activity for them is e-mail—getting it and sending it—but they also do lots of other political activities online. A few of these activities, such as using political discussion groups or Web logs, are things that few other people do. At this stage in the campaign, Democrats are more likely than Republicans to be Online Political Citizens. This is not unexpected. The Democratic candidates have a contested primary election; the Republicans do not. We know from 2000 and 2002 that millions of Americans will rush online in the last few weeks of the fall campaign. In the 2002 midterm election, 40% of Internet users turned to the Internet for political information. Most do so for convenience, but one-third (37%) say they do not find the information they want in the traditional media. 27 We estimate that as much as 75% of Internet users and half of American voters will look for political information online in the 2004 campaign. Most will seek sources other than the news media, and most will look in the last two weeks before the election. Final IPDI report 2/11/04 5:45 PM Page 34
  • 39. 35 Online activities take time and effort to muster. Web designers know that a site needs time to get organized, debugged and attract an audience. Web logs and online discussion groups need time to find their focus and to gather a critical mass of participants. You can build a Web site overnight, but technology is only one part. You cannot create an online campaign overnight. And online campaigns matter. We think that people chatting online today will vote in November. Online fund-raising has made it possible for a relatively unknown governor from a small state to quickly raise tens of millions of dollars, and a former general to jump into the race at the last minute. People who have donated online early in the campaign are, we believe, not likely to go back to sending a check in the mail. They will again donate online. Finally, the fact that so many of the Online Political Citizens are Influentials means that if candidates, parties and issue advocacy groups want to reach the people who reach others, the place to find them is the Internet. These citizens are active and engaged members of their communities, which suggests that their involvement may not be short-lived. This is a hopeful sign for reinvigorating the American political process, and that, for us, is the true promise of online politics. Final IPDI report 2/11/04 5:45 PM Page 35
  • 40. 36 The Online Political Influentials project was envisioned as parallel and comple- mentary surveys. An extensive online survey could take advantage of new tech- nology to gather a select sample of people we suspected were highly involved in politics, while at the same time a telephone survey could confirm those results with a nationally representative sample. Nielsen//NetRatings was commissioned to use its WebIntercept online survey system to poll respondents from a select group of Web sites. Respondents were culled from 1.5 million members of a panel Nielsen has recruited that is represen- tative of the online population. The Institute selected a group of Web sites based on politically-oriented content and submitted the list to Nielsen//NetRatings. Sites chosen were broadly representative, so political sites from all viable political par- ties and presidential candidates were included, along with the most popular news sites in the United States. We also selected some Web sites of political publications and nonpartisan political Web sites that we thought would attract Online Political Citizens. (A full list of the sites and the number of respondents culled from each is in Appendix B.) Nielsen began the operation of WebIntercept Nov. 20 and closed the survey on Dec. 31, 2003. As Internet users accessed the sites, a popup survey appeared on their screens to ask if they would take the survey. As an incentive, respondents were entered in a contest to win a television. During the time the survey was live, 33,367 surveys were viewed and 1,392 were completed, for a completion rate of 4.2%. One reason Online Political Citizens interest researchers is the same reason they are poorly studied. It is an exclusive group, but unless we know how exclusive, it is difficult to judge its impact. Because of that, a parallel survey was proposed and RoperASW enlisted as a partner. Roper was able to use conventional telephone survey methods to gather a sample much closer to a random representation of all American adults. Our goal was to offer some estimate about how many Online Political Citizens there would be in the general public and compare online behavior in a sample representative of all U.S. adults. This would give us greater confidence that our findings represent the public at large. RoperASW used a random digit-dialing probability sample of all telephone house- holds in the continental United States. Interviewers conducted interviews using computer-assisted telephone interviewing with 1,029 people Nov. 21 to 23, 2003. The questions used by the Institute were part of the weekly RoperASW omnibus survey. Omnibus surveys are conducted over a single weekend in order to quickly METHODOLOGY Final IPDI report 2/11/04 5:45 PM Page 36
  • 41. 37 gather data and track trends. Therefore, there is little opportunity for multiple follow-up calls, and the response rate is lower than typical national surveys. RoperASW estimates the response rate of its weekly omnibus surveys at between 7 and 12%. The data are weighted using a formula that takes into account age, sex, education, race and geographic region. In all cases in this report we use the weighted percentages, but the actual number of respondents differs. Roper interviewed 1,029 respondents, weighted to 1,000. There are some important considerations about the data. First, RoperASW identifies Influentials differently depending on whether the surveys are done in-person or over the telephone. When the survey is done in person, RoperASW defines an Influential as anyone who answers Yes to three of the 11 Influentials questions. Interviewing people in-person was popular 50 years ago, when door-to-door salesmen still appeared, but today it is much more difficult and expensive. Instead, survey researchers use the telephone to reach more people more easily. Unfortunately, opinion survey researchers have found that people are more likely to inflate their responses over the phone. A respondent answering questions over the phone may feel a need to report socially desirable responses, such as reporting they voted when they did not. In addition, respondents seem more likely to answer Yes to a series of questions given over the telephone instead of when they are allowed to view a list of activities in a face-to-face or online survey. RoperASW researchers have learned over the years that they must change the means of deter- mining Influentials, and when they use telephone survey data they require four affirmative responses, instead of three. Online surveys, however, do not impose on the respondent the same types of personal pressures to respond. The respondent is at home or in the office—not responding to another person—and can examine the full list of options, rather than hearing them read aloud. Because of this the Nielsen//NetRatings online sur- vey uses only three questions instead of four to determine who is an Influential. In comparing Influentials in both surveys, we determined Influentials using four questions in the telephone data and three questions in the online data. While we believe this correctly represents the data, it also makes the difference greater in the number of Influentials in each set of data. Nonetheless, our conclusions in the report remain the same regardless of how Influentials are calculated. We defined Online Political Citizens with several indicators. First, a respondent had to report that he or she had visited the Web site of a candidate or political party within the last two or three months. We believed that respondents must have taken an early interest in the campaign online in order to qualify as OPCs. Second, they had to be politically involved online in a variety of ways. To be defined as OPCs, respondents had to indicate that, within the last two to three months, they had participated in any two of the following activities: donated money using the Internet to a candidate or political organization, visited a news Web site for news about politics and campaigns, visited a political discussion group or chat room online, visited a Web log that discusses politics or current events, received any kind of political e-mail, and sent or forwarded to someone else any kind of political e-mail. Final IPDI report 2/11/04 5:45 PM Page 37
  • 42. 38 Throughout this report we talk about relationships and compare different groups. In each comparison, a statistical test of significance has been applied and the relation- ship found significant at p < .001. Comparisons where this is not the case are noted. Final IPDI report 2/11/04 5:45 PM Page 38
  • 43. 39 NOTES 1 Ed Keller and Jon Berry, The Influentials (New York: The Free Press, 2003), book jacket. Keller is CEO and Berry is a vice president of RoperASW. 2 Keller and Berry, The Influentials, 147-148, 2, 14. 3 Samantha M. Shapiro, “The Dean Connection,” The New York Times Magazine, 7 December 2003. 4 Keller and Berry, The Influentials, book jacket. 5 Keller and Berry, The Influentials, book jacket. 6 Shapiro. 7 Gary Wolf, “How the Internet Invented Howard Dean,” January 2004, <www.wired.com/wired/archive/12.01/dean/html> (18 January 2004). 8 Source: Nielsen//NetRatings data and Meetup.com. 9 This is a weighted percentage. 10 We believe the 7% estimate is very close, though perhaps slightly inflated. First, we know that telephone survey respondents are more likely to respond affirmatively to these ques- tions (see the methodology for a discussion of this). We also know that surveys of “socially desirable” activities such as political participation and voting tend to get high estimates of participation—people say they voted when they did not. Second, Nielsen//NetRatings estimates that the number of people accessing presidential campaign Web sites was about 4.3 million in November and December 2003. One of our defining characteristics of Online Political Citizens is that they have accessed a campaign Web site in the past two to three months. We can also estimate the number of individuals in other online activities. In December 2003, the number of individuals in campaign discussion groups on Yahoo! was about 54,000, although there are many other discussion groups outside the “campaign” designation that discuss politics. The number of people signed up to receive notices from MeetUp about upcoming presidential campaign events was 275,000. MoveOn.org claimed 1.7 million members in December. All of these numbers must be viewed as rough estimates. For example, calculating Web site visitors across numerous sites runs the risk of counting a single user more than once, and similarly a single user may sign up more than once in discussion groups or as a member of Meetup. Nonetheless, we believe that altogether this community numbers several million people. The Roper estimate of 7% of the adult population, with a margin of error of three per- centage points, would be between 8.6 and 21.5 million adults. 11 The number of affirmative answers depends on whether the survey is conducted over the telephone, on the Internet, or in person. See the Methodology section for a full discussion of this. 12 Keller and Berry, The Influentials, 147, 53. 13 In comparing OPCs with the general public, the differences are statistically significant at p < .001 in nine of 11 activities. For “writing an article for a magazine or newspaper” and “holding or running for public office” the differences are statistically significant at p < .05. 14 We believe that the results from the general public Influentials reflected in Figure 6 might be slightly inflated. There are two possible reasons for this. First, Influentials drawn from the RoperASW survey were determined using four questions, whereas the Influentials Final IPDI report 2/11/04 5:45 PM Page 39
  • 44. 40 culled from the Nielsen//NetRatings survey were determined using three questions. This depresses the overall percentages for the OPCs and increases the same for Influentials from the general public. Second, RoperASW interviewed a sample Nov. 21 to 23 that reported taking part in the Influentials activities to a great degree. For most questions, political involvement was greater than what is typically found (see Keller and Berry, The Influentials, 41). This may simply be an anomaly of the data. Nonetheless, due to the large differences between the two groups (23% in one case) we are confident of our observations about the difference in political activity between the groups. 15 Amanda Lenhart, “The Ever-Shifting Internet Population: A new look at Internet access and the digital divide,” (Washington, D.C.: Pew Internet & American Life Project, 2003), 3. Mary Madden and Lee Rainie, “America’s Online Pursuits: The Changing Picture of Who’s Online and What They Do,” (Washington, D.C.: Pew Internet & American Life Project, 2003), 3. Both reports are available online at <www.pewinternet.org>. 16 Madden and Rainie. 17 Keller and Berry, The Influentials, 31. 18 Jeff Glasser and Betsy Streisand, “Virtual campaign pays off,” U.S. News & World Report, 6 March 2000. 19 Jon Sawyer, “Dean’s savvy Web pros are changing political campaigns,” St. Louis Post-Dispatch, 11 January 2004. 20 Participation, Competition, Engagement: How To Revive and Improve Public Funding For Presidential Nomination Politics. (Washington, D.C.: The Campaign Finance Institute Task Force on Presidential Nomination Financing, 2003), 30. Since passage of the 2002 Bipartisan Campaign Reform Act, permissible donations have increased from $1,000 to $2,000 per individual. 21 For this analysis, n = 160 Democrats and n = 105 Republicans who donated in the past two to three months. 22 This is taken from the RoperASW telephone survey Nov. 21 to 23, 2003, conducted for this project. 23 In the Nielsen//NetRatings online survey data, the strongest predictor was sending or forwarding political e-mail, followed by having visited a party or candidate Web site, and then receiving political e-mail. In the Roper telephone survey data, the strongest predictor was sending or forwarding political e-mail, followed by receiving political e-mail. 24 While research has shown that about a quarter of all Internet users say they have participated in chat rooms or online discussions (see Madden and Rainie), we asked specifically about discussion and chat rooms about politics within the last two to three months. The question asked was “Which of the following have you done in the past two to three months? Visited a political discussion group or chat room online. Visited or post- ed information to a Web log that discusses politics or current events.” 25 We find it unusual that more of the OPCs in the telephone survey report using Web logs than those in the online survey (47% compared to 27%). This is also true for the ques- tion about discussion groups or chat rooms (34% compared to 25%). This may be due to the greater willingness of respondents to respond affirmatively in telephone surveys, or to the many more options available in the online survey (we discuss both issues in the Methodology section). 26 In the RoperASW telephone survey, 57% of Online Political Citizens say they forwarded some political e-mail to someone else. Of the non-OPCs, 5% say they forwarded some e-mail. 27 See Madden and Rainie. Final IPDI report 2/11/04 5:45 PM Page 40
  • 45. 41 ROPERASW SURVEY (1,029 COMPLETED INTERVIEWS) PERCENTAGES FOR THE ROPER DATA ARE WEIGHTED Interviewer enter respondent sex Male 496 48.1% Female 533 51.9% Now for classification purposes only, would you tell me are you … Single 244 28.4 Married 540 49.1 Separated, widowed, or divorced 212 20.0 Don’t know/Refused 33 2.5 Are you yourself currently employed? Full-time 489 47.7 Part-time 136 14.2 Not employed 372 35.8 Don’t know 4 0.4 Refused 28 2.7 Do you ever use the Internet at work? (Answered by 621 who are online.) Yes 309 46 No 312 54 What is your age? 18 to 20 56 7.7 21 to 24 55 5.3 25 to 29 78 10.2 30 to 34 60 8.3 35 to 39 81 10.3 40 to 44 87 10.8 45 to 49 95 7.2 50 to 54 107 7.9 55 to 59 92 7.2 60 to 64 64 4.6 APPENDIX A: SURVEY RESULTS Final IPDI report 2/11/04 5:45 PM Page 41
  • 46. 42 65 to 69 52 3.8 70 to 74 53 4.7 75 and over 92 7.6 Refused 55 4.3 What is the last grade of school you completed? Less than high school graduate 82 16.9 High school graduate 327 27.9 Some college 236 26.9 Graduated college 218 16.0 Post graduate school 113 8.3 Other 12 1.0 Don’t know/Refused 38 2.9 Would you please tell me which of the following categories most closely represents your annual household income? Under $15,000 89 10.4 $15,000 to less than $20,000 70 8.4 $20,000 to less than $25,000 53 6.0 $25,000 to less than $30,000 56 5.8 $30,000 to less than $40,000 99 9.7 $40,000 to less than $50,000 85 8.2 $50,000 to less than $75,000 141 13.2 $75,000 to less than $100,000 96 9.1 $100,000 to less than $125,000 42 3.3 $125,000 to less than $150,000 24 2.1 $150,000 and over 25 1.8 Don’t know/Refused 245 22.1 Are you Hispanic? Yes 57 11.8 No 972 88.2 Are you … White 825 68.1 Black 78 11.7 Asian 13 1.1 Other/Don’t know/Refused 109 19.1 Which of the following products and services do you or does any member of your household currently own or use at home? Internet access or e-mail by dialup service 369 34.0 High speed Internet access, such as cable 265 25.7 modem, DSL or digital set top box Final IPDI report 2/11/04 5:45 PM Page 42
  • 47. 43 I’m going to read you a list of things some people do about government or politics. As I read each one, please tell me if you happened to have done any of these things in the past year. How about … Written or called any politician at the state, local or national level 169 13.8 Attended a political rally, speech, or organized protest of any kind 80 6.2 Attended a public meeting on town or school affairs 258 24.2 Held or run for political office 18 1.9 Served on a committee for some local organization 136 11.8 Served as an officer for some club or organization 141 11.3 Written a letter to the editor or a newspaper or magazine 131 11.6 or called a live radio or TV show to express an opinion Signed a petition 304 27.2 Worked for a political party 62 4.9 Made a speech 94 8.6 Written an article for a magazine or newspaper 108 11.4 Been an active member of any group that tries to 114 9.2 influence public policy or government Thinking about an average weekday, how much time do you spend on the Internet? None 370 37.0 Less than 15 minutes 76 7.6 16 to 30 minutes 91 9.1 31 to 59 minutes 68 6.8 1 hour to less than 2 hours 167 16.7 2 hours to less than 3 hours 80 8.0 3 hours to less than 5 hours 52 5.2 5 hours or more 77 7.7 Don’t know/refused 19 1.9 Were you able to vote in the last election for your U.S. representative in Congress, in November 2002? Yes 650 65.0 No 340 34.0 Refused 2 0.2 Don’t know/refused 8 0.8 Which of the following have you done in the past 2 to 3 months? Visited a news Web site, such as CNN.com or MSNBC.com, 248 24.8 for news about politics and campaigns Received any kind of political e-mail 191 19.1 Donated money to a candidate or political organization 102 10.2 Donated money using the Internet to a candidate or political organization 24 2.4 Visited the Web site of a political party or candidate 98 9.8 Sent or forwarded to someone else any kind of e-mail about politics 90 9.0 Visited or posted information to a Web log that discusses 67 6.7 politics or current events Visited a political discussion group or chat room online 45 4.5 Attended a presidential campaign event of any kind 45 4.5 Final IPDI report 2/11/04 5:45 PM Page 43
  • 48. 44 NIELSEN//NETRATINGS SURVEY (1,392 COMPLETED INTERVIEWS) What is your age? 18 to 24 178 12.8 25 to 34 296 21.3 35 to 54 649 46.6 55 to 64 191 13.7 65 and older 78 5.6 What is your gender? Male 863 62.0 Female 529 38.0 Which of the following best describes the highest level of education you have achieved? Grammar school 3 .2 Some high school 23 1.7 High school graduate 126 9.1 Some college 374 26.9 Associate degree 138 9.9 Bachelor’s degree 422 30.3 Post graduate degree 306 22.0 Are you Spanish/Hispanic/Latino? No, not Spanish/Hispanic/Latino 1,298 93.2 Yes, Mexican, Mexican/American or Chicano 34 2.4 Yes, Puerto Rican 13 .9 Yes, Cuban 11 .8 Yes, other Spanish/Hispanic/Latino 36 2.6 What is your race? African American 56 4.0 White 1,172 84.2 Asian or Pacific Islander 29 2.1 American Indian or Eskimo or Aleut 19 1.4 Other 22 1.6 Are you currently employed? Full-time 938 67.4 Part-time 172 12.4 Not employed 282 20.3 What is your marital status? Single 420 30.2 Married 816 58.6 Separated 12 .9 Divorced 117 8.4 Widowed 27 1.9 Final IPDI report 2/11/04 5:45 PM Page 44
  • 49. 45 What is the total annual household income before taxes for all household members combined? $0 to $14,999 87 6.3 $15,000 to $19,999 45 3.2 $20,000 to $24,999 45 3.2 $25,000 to $29,999 55 4.0 $30,000 to $39,999 130 9.3 $40,000 to $49,999 153 11.0 $50,000 to $74,999 314 22.6 $75,000 to $99,999 217 15.6 $100,000 to $124,999 154 11.1 $125,000 to $149,999 72 5.2 $150,000 + 120 8.6 Here is a list of things some people do about government or politics. Have you happened to have done any of these things in the past year? Which ones? Written or called any politician at the state, local or national level 610 43.8 Attended a political rally, speech, or organized protest of any kind 361 25.9 Attended a public meeting on town or school affairs 502 36.1 Held or run for political office 30 2.2 Served on a committee for some local organization 238 17.1 Served as an officer for some club or organization 287 20.6 Written a letter to the editor or a newspaper or magazine 333 23.9 or called a live radio or TV show to express an opinion Signed a petition 841 60.4 Worked for a political party 173 12.4 Made a speech 175 12.6 Written an article for a magazine or newspaper 113 8.1 Been an active member of any group that tries 260 18.7 to influence public policy or government Were you able to vote in the last election for your U.S. representative in Congress, in November 2002? Yes 1,145 82.3 No 207 14.9 Don’t remember/Not sure 40 2.9 Were you able to vote in the last election for president, in November 2000? Yes 1,202 86.4 No 173 12.4 Voted but not for president 7 .5 Don’t remember/Not sure 10 .7 What is your political affiliation? Democrat 504 36.2 Republican 454 32.6 Independent 278 20.0 Green Party 23 1.7 Libertarian Party 20 1.4 Reform Party 0 0 Final IPDI report 2/11/04 5:45 PM Page 45
  • 50. 46 Other third party 16 1.1 Don’t know 97 7.0 Thinking about an AVERAGE WEEKDAY, how much time do you spend with each of the following media? Television None 50 3.6 Less than 15 minutes 59 4.2 16 to 30 minutes 98 7.0 31 to 59 minutes 176 12.6 1 to 2 hours 363 26.1 2 to 3 hours 316 22.7 3 to 5 hours 190 13.6 More than 5 hours 140 10.1 Radio None 88 6.3 Less than 15 minutes 163 11.7 16 to 30 minutes 230 16.5 31 to 59 minutes 265 19.0 1 to 2 hours 297 21.3 2 to 3 hours 135 9.7 3 to 5 hours 107 7.7 More than 5 hours 107 7.7 Newspapers None 159 11.4 Less than 15 minutes 272 19.5 16 to 30 minutes 384 27.6 31 to 59 minutes 261 18.8 1 to 2 hours 210 15.1 2 to 3 hours 54 3.9 3 to 5 hours 26 1.9 More than 5 hours 26 1.9 Magazines None 266 19.1 Less than 15 minutes 462 33.2 16 to 30 minutes 318 22.8 31 to 59 minutes 165 11.9 1 to 2 hours 121 8.7 2 to 3 hours 32 2.3 3 to 5 hours 12 .9 More than 5 hours 16 1.1 Internet None 5 .4 Less than 15 minutes 43 3.1 16 to 30 minutes 129 9.3 31 to 59 minutes 253 18.2 1 to 2 hours 382 27.4 2 to 3 hours 212 15.2 Final IPDI report 2/11/04 5:45 PM Page 46
  • 51. 47 3 to 5 hours 163 11.7 More than 5 hours 205 14.7 Other than for e-mail, how often are you online? Every day 1,020 73.3 5 to 6 times a week 228 16.4 3 to 4 times a week 101 7.3 1 to 2 times a week 31 2.2 Less than once a week 12 .9 Which of the following have you done in the past 2-3 months? Check all that apply. Donated money to a candidate or political organization 327 23.5 Donated money using the Internet to a candidate or political organization 127 9.1 Attended a presidential campaign event of any kind 103 7.4 Worked for a political campaign 103 7.4 Visited the Web site of a political party or candidate 660 47.4 Visited a news Web site, such as CNN.com or MSNBC.com, 791 56.8 for news about politics and campaigns Visited MSNBC.com 833 59.8 Received any e-mail from Meetup.com 119 8.5 Visited a political discussion group or chat room online 144 10.3 Visited or posted information to a Web log that 161 11.6 discusses politics or current events Received any kind of political e-mail 635 45.6 Sent or forwarded to someone else any kind of e-mail about politics 458 32.9 Signed up for or received e-mail from a politician or party representative 341 24.5 Gone online to get political information 686 49.3 Which of the following have you done PRIOR to the current presidential election? Check all that apply Donated money to a candidate or political organization 337 24.2 in any previous election Attended presidential campaign events in any previous election 252 18.1 Worked for a campaign in any previous election 205 14.7 Gone online to get political information in any previous election 658 47.3 What is your marital status? Single 420 30.2 Married 816 58.6 Final IPDI report 2/11/04 5:45 PM Page 47
  • 52. 48 Below is the list of Web sites from which Nielsen//NetRatings recruited respondents to the online survey. Our intention was to select a broad group of political Web sites that collected people access- ing information about mainstream American politics. We sampled from all the major parties and all the major presidential candidates. We also chose the political news pages of major news Web sites and a selection of political magazines. The number of respondents generally reflects the overall traffic to each Web site. Description of Web site Address Number of respondents George W. Bush re-election campaign home page georgewbush.com 27 Howard Dean campaign home page deanforamerica.com 100 John Edwards campaign home page johnedwards2004.com 7 Wesley Clark campaign home page clark04.com 21 Wesley Clark official campaign Web log campaign.forclark.com 1 Dick Gephardt campaign home page dickgephardt.com 2 John Kerry campaign home page johnkerry.com 12 Dennis Kucinich campaign home page kucinich.us 12 Joe Lieberman campaign home page joe2004.com 4 Carol Moseley Braun campaign home page carolforpresident.com 0 Al Sharpton campaign home page al2004.org 0 Democratic Party home page democrats.org 42 Republican Party home page rnc.org 0 Green Party home page greenparty.org 0 Libertarian Party home page lp.org 7 CNN home page www.cnn.com 62 CNN politics Web page cnn.com/allpolitics 48 Fox News home page foxnews.com 88 Fox News politics Web page foxnews.com/politics 14 New York Times home page nytimes.com 129 New York Times online edition politics page nytimes.com/pages/politics 0 Washington Post online home page washingtonpost.com 59 Washington Post online edition politics page washingtonpost.com/wp_dyn/politics 25 MSNBC home page msnbc.com 369 APPENDIX B: WEB SITES USED FOR THE ONLINE SURVEY Final IPDI report 2/11/04 5:45 PM Page 48
  • 53. 49 MSNBC Web site politics page msnbc.com/news/politics_front.asp 2 Slate magazine slate.msn.com 1 Lost Angeles Times online edition home page www.latimes.com 29 Los Angeles Times online edition politics page latimes.com/news/politics 5 ABC News home page abcnews.go.com 23 ABC News online politics page abcnews.go.com/sections/politics 2 CBS News home page cbsnews.com 43 CBS News online politics page cbsnews.com/sections/politics/main250.shtml 1 USA Today home page usatoday.com 73 USA Today online edition elections front page usatoday.com/news/politicselections/front.htm 1 Yahoo News news.yahoo.com 46 Meetup.com meetup.com 6 Moveon.org moveon.org 62 The Nation magazine thenation.com 5 National Review magazine nationalreview.com 18 The New Republic magazine tnr.com 2 The Weekly Standard magazine weeklystandard.com 7 C-Span home page c-span.org 5 Andrew Sullivan Web log andrewsullivan.com 1 Daily Kos Web log dailykos.com 4 Instapundit Web log instapundit.com 5 Talking Points memo Web log talkingpointsmemo.com 1 Vote Smart non-partisan political Web site vote-smart.org 3 Townhall.com conservative news townhall.com 18 Note: In most cases, the prefix “www” must be added to the site addresses above. In some cases the addresses are redirect pages, and visitors are automatically directed to the appropriate page. Final IPDI report 2/11/04 5:45 PM Page 49
  • 54. 50 In several instances we conducted simple linear regressions, emphasizing only those relationships significant at p < .001. We sometimes mentioned relationships where the statistical significance was greater than p < .05, but we describe those as less strong. Regression analysis summary for online political involvement variables predicting whether respon- dent is an Influential (Nielsen//NetRatings online survey data) Variable B SEB Beta Donated money online .05 .05 .03 Visited political Web site .16 .03 .16*** Visited news Web site .06 .03 .06* Visited political discussion .10 .04 .06* group/chat room Visited political Web log .01 .04 .01 Received political e-mail .14 .03 .14*** Sent or forwarded political e-mail .19 .03 .18*** R2 = .19 (n = 1,392) * p < .05 *** p < .001 Regression analysis summary for online political involvement variables predicting whether respon- dent is an Influential (Roper telephone survey data) Variable B SEB Beta Donated money online -.02 .06 -.01 Visited political Web site .07 .04 .07 Visited news Web site .04 .02 .06 Visited political discussion -.05 .05 -.04 group/chat room Visited political Web log .10 .04 .09* Received political e-mail .08 .03 .10** Sent or forwarded political e-mail .13 .04 .13*** R2 = .09 (n = 1,029) * p < .05 ** p < .01 *** p < .001 APPENDIX C: LINEAR REGRESSION INFORMATION Final IPDI report 2/11/04 5:45 PM Page 50
  • 55. 51 Regression analysis summary for media variables predicting whether respondent is an Online Political Citizen (Nielsen//NetRatings online survey data) Variable B SEB Beta Television -.04 .01 -.16*** Radio .01 .01 .05 Magazines .01 .01 .02 Newspapers .03 .01 .10** Internet .04 .01 .14*** R2 = .04 (n = 1,392) ** p < .01 *** p < .001 Regression analysis summary for media variables predicting whether respondent is an Influential (Nielsen//NetRatings online survey data) Variable B SEB Beta Television -.05 .01 -.16*** Radio .02 .01 .06* Magazines .01 .01 .04 Newspapers .04 .01 .12*** Internet 0 .01 .02 R2 = .04 (n = 1,392) * p < .05 *** p < .001 Final IPDI report 2/11/04 5:45 PM Page 51
  • 56. Institute for Politics, Democracy & the Internet The Graduate School of Political Management The George Washington University 805 21st St., NW, Suite 401, Washington, D.C. 20052 T (202) 994.6000 | toll free (800) 367-4776 F (202) 994-6006 www.ipdi.org Design/Edit: Reingold, Inc. Final IPDI report 2/11/04 5:45 PM Page 52

Related Documents