Much of the existing literature on the digital divide - the differences between the "haves" and "have nots" regarding access to the Internet - limits its scope to a binary classification of technology use by only considering whether someone does or does not use the Internet. To remedy this shortcoming, in this paper I look at the differences in people's online skills. In order to measure online ability, I assigned search tasks to a random sample of Internet users from a suburban county in the United States. My findings suggest that people search for content in a myriad of ways and there is considerable difference in whether individuals are able to find various types of content on the Web and a large variance in how long it takes to complete online tasks. Age is negatively associated with one's level of Internet skill, experience with the technology is positively related to online skill, and differences in gender do little to explain the variance in the ability of different people to find content online.
Introduction: Inequalities in Internet Use
Much of the literature documenting the Internet's spread has focused on the differences among those who have access to the Internet and those who do not, or the differences among those who use it and those who do not. Since the National Telecommunications & Information Administration published its first report "Falling Through the Net: A Survey of the Have Nots in Rural and Urban America" in 1995, many analyses have been written on the inequalities of access to and use of the medium. Existing studies of differential Internet access and use document inequalities among various segments of the population (Bucy, 2000) with particular attention to education (NTIA, 2000), race (Hoffman and Novak, 1999), gender (Bimber, 2000), age (Loges and Jung, 2001), income (Benton Foundation, 1998) and rural residence (Strover, 1999).
These studies have been essential in understanding inequalities in access to the Internet, or what has come to be known as the "digital divide". Here, I argue that as the medium spreads to a majority of the population (NTIA, 2002) it is increasingly important to look at not only who uses the Internet, but also to distinguish varying levels of online skills among individuals. Skill, in this context, is defined as the ability to efficiently and effectively find information on the Web. By exploring the differences in how people use the Web for information retrieval, we can discern if there is a "second-level digital divide" in the making as the Web spreads to the majority of the American population. To explore this question, I report findings from a project that explores people's ability to locate content online. Documenting differences in Web use skills allows us to distinguish how different kinds of people are able to take advantage of this medium in varying ways.
Refining the Current Approach to the Digital Divide
While most reports identify differences among various segments of the population, over time studies emphasize the increasing diffusion of the medium among the population at large (Howard, Rainie, and Jones, 2001; Katz and Rice, 2002; NTIA 2000, 2002; Pew Internet and American Life Project, 2000). As more people start using the Web for communication and information retrieval, it becomes less useful to merely look at binary classifications of who is online when discussing questions of inequality in relation to the Internet (DiMaggio and Hargittai, 2001). Rather, we need to start looking at differences in how those who are online use the medium, that is, differences in people's online skills. It is important to expand the research agenda to allow analyses of the differences among Internet users.
Some scholars have offered a refined understanding of the digital divide by suggesting that there are different levels at which divides exist. Kling (1998) identified differences in technical access (the physical availability of the technology) and in social access (the professional knowledge and technical skills necessary to benefit from information technologies). Norris (2001) pointed to divides at three levels: the global divide which encompasses differences among industrialized and lesser developed nations; the social divide which points to inequalities among the population within one nation; and, a democratic divide which refers to the differences among those who do and do not use digital technologies to engage and participate in public life. DiMaggio and Hargittai (2001) suggested five dimensions along which divides may exist:
- technical means (software, hardware, connectivity quality);
- autonomy of use (location of access, freedom to use the medium for one's preferred activities);
- use patterns (types of uses of the Internet);
- social support networks (availability of others one can turn to for assistance with use, size of networks to encourage use); and,
- skill (one's ability to use the medium effectively).
The goal of this study is to empirically investigate such refined understandings of a second-level digital divide by exploring differences in Internet users' online skills.
Information about people's online skills tells us to what extent they are able to use the medium in ways of most interest to them and in ways that are most useful to their particular needs. The ability to find different types of information online allows people to use the medium to their maximum benefit. If users often give up in frustration and confusion then merely having access does not mean that a digital divide has been solved because a divide remains in their capacity to effectively use the Internet (Wilson, 2000).
How can we talk about the Internet's effect on political participation if a user does not possess the skills to find political information? Similarly, how can the Internet prove to be a useful link between the government and citizens if people are unable to find official documents online? By measuring users' Internet skills, we can bridge the gap in the literature between mere structural measures of access and descriptions of what people do online in order to better describe the differences in online skills among individuals.
Methods and Data
The data presented in this paper are based on in-person observations and interviews with a random sample of 54 Internet users from the suburban towns and boroughs of a New Jersey county conducted during the summer and fall of 2001. Respondents were recruited through random sampling. Potential respondents were first sent a letter via postal mail explaining the project and requesting participation. They were also sent a brochure that presented more details about the study. People were also pointed to the study's Web site (www.webuse.org) for more information and were given the option of calling/writing the researcher to schedule an appointment. A few days after the letters had been sent, the households were contacted by telephone.
The eligible adult (i.e. Internet user adult over 18) with the next nearest birthday was selected in order to randomly sample from within households. If this randomly selected person from the household was not willing to participate then the household was coded as a refusal even if another member of the household would have been willing to take part in the study. Such strict measures of random selection assure that the participants of the study represent a truly random sample of the area's Internet user population. People who were identified as Web users were invited to participate in the study. Web users are defined as people who go online at least once every month for more than using e-mail. This is a low threshold for including people in the study; it is used to maximize variance in experience.
Respondents were offered $40 for their participation, which they received after the study session. Respondents were asked to come to the research site on the university campus and were offered assistance with transportation if they could not provide their own (one respondent took advantage of this option and was reimbursed for bus fare). The study's response rate was 64 percent, considerably high given the time cost involved in participation.
Studying how people find information online
Researchers in the library and information science community have conducted analyses on how people locate content online, but their projects limit their scope to people in particular academic communities (e.g. graduate students in information science programs (Wang, Hawk and Tenopir, 2000) or, at best, college students in general (Cothey, 2002)) making their findings impossible to generalize to the broader Internet user population (see Jansen and Pooch, 2001 for a review of the literature on Web searches). Moreover, these studies focus on the technicalities of searches - such as the number of search terms used in a search query - without exploring demographic variables that contribute to differences in how people find material online.
Information about subjects' usual Internet use and history as well as data on their demographic background were collected via surveys. People's ability to locate content on the Web was measured by giving respondents tasks to perform online. I analyzed the results of the following five tasks. Individuals were asked to find:
- Information about local cultural events in the area such as art shows, musical performances, theatre shows or movies;
- Music they could listen to online;
- A Web site comparing different presidential candidates' views on abortion;
- Tax forms; and,
- Art by kids.
These tasks were chosen to explore people's ability to find information on the Web in different topical domains. They explore whether users can find locally relevant content, can take advantage of the multimedia nature of the Web, can use the Web for political purposes, for government information, and for a random task such as finding children's art online. The last task proved to be useful in that it was new to all users and thus was able to control for people's previous experiences with any given task.
Respondents were given the choice of using a PC or a Mac so they could use the platform with which they are most familiar. The three most popular browsing software applications were all available on both machines; Internet Explorer, Netscape Communicator, and America Online to allow respondents to replicate their usual online experiences. The computers connected to the Internet on a high-speed university network line. Additionally, a program called Don't Panic (Panicware, 2001) was used to erase the browser and URL history on each browser program so that each respondent started out with a clean slate and was not influenced by previous users' actions. The search sessions were recorded with the Hypercam screen capture program (or SnapZPro on the Mac) that generated audio-visual files of the entire search sessions. The whole screen was captured with every action - e.g. click of the mouse, scrolling - and every verbal comment the respondent made during the search. These files were then analyzed to measure whether people successfully completed a task and how much time (in seconds) they spent on each task. This is the outcome of interest in this study: Skill is operationalized as success and time to completion of a task.
The researcher sat behind to the left of the respondent and refrained from influencing the respondents' strategies (e.g. never suggested any particular online actions, did not answer questions about spelling or whether a certain click would be useful). Respondents were encouraged to look for the information until they found it. No one was cut off from pursuing a search. In some cases when respondents looked frustrated or agitated they were given the option of moving on. However, when a subject simply stated that he or she was unable to perform a certain task, that person was encouraged to try several times nonetheless. When the subject suggested multiple times that they would not be able to complete a search, they were read the next task (Hargittai (in press) describes the study methodology in more detail).
Respondents range in age from 18-81 (see Table 1 for details). Half of the sample is male. Fifty-five percent of the respondents work full time and an additional 15 percent work part-time. Their occupations range from real-estate agents, environmental policy analysts, blue-collar workers to office assistants, teachers, service employees and medical professionals in addition to students, unemployed and retired persons. See the Appendix for a complete list of participants' demographic characteristics including their occupations.
Table 1: Descriptive Statistics for Independent Variables
a Education and Family income have no means as those variables were collected categorically.
Mean Standard Deviation Median Minimum Maximum Age 43.06 16.37 41.5 18 81 Educationa N/A N/A College Less than high school Ph.D. Family Incomea N/A N/A $80,000-$89,000 $17,500-$19,000 >$250,000 Number of years since first use of the Internet 6.24 3.54 6 0 16 Number of hours browsing the Web weekly 8.55 10.76 5 8 minutes 70
Eighty-nine percent of the respondents are White, there are four African Americans and one Asian American respondent, one person chose the "other" category for race; there is one Hispanic participant. Eleven percent live without another adult in the household, 45 percent live with a spouse, the rest live with roommates or others (parents in most cases). Eighty-two percent of the respondents have children. Of these 28 people, 53.5 percent (15 people) have children currently living with them.
On average, participants in this study are more educated than the general Internet user population as 26 percent have less than a college degree (among whom six are still enrolled in school), a third hold a college degree and the remaining 40 percent have a graduate degree. This suggests that findings from the study will be conservative with respect to effects of different educational levels on people's ability to use the Internet effectively.
The family income of respondents is greater than the national average although it is important to note that this county is one of the highest income counties in the country and thus despite the high median income, the sample is not out of the ordinary for the local population. The average median per capita income in this county in 2000 was almost $40,000 (based on U.S. Census data). Here, I look at family income, which is likely to be considerably higher on average. See Table 1 for descriptive statistics of the family income distribution. Respondents differ in their political leaning with 20 percent self-identified as conservative, 30 percent claiming to be middle of the road and half identifying as liberal.
Regarding Web use frequency and history, the group is diverse. Respondents were asked how much time they spend on the Web each week excluding e-mail use, that is, time spent specifically for browsing information online. Web use ranges from just a few minutes to over 30 hours weekly (see Table 1 for details). About half of the respondents use the Web more than they use e-mail, 18.5 percent use the two types of online services the same amount and the rest do more e-mailing than Web browsing in a typical week.
The group is similarly diverse in its overall experience with the medium. Users were asked to note when they first started using the Internet. One person went online the year of the study with an additional 18.5 percent only having used it for two years or less. However, many - 39 percent - of the subjects had been users for 5-7 years. There are also several long-term users among the respondents with 18.5 percent having used the Internet for more than ten years (see Table 1 for details).
Differences in Ability to Find Content Online
I measure people's online skills in two ways. First, the binary success/failure rate shows what portion of the respondents was able to complete a certain task. Second, the time to completion of each task is measured in seconds to show the gradual differences in how long people take to find information on the Web. The exact time spent on each task is recorded for every respondent so information is available both on when respondents successfully completed a task and when they decided to give up on a task.
There is some variance in the success rate for performing tasks and a large variance in the amount of time used to complete tasks. Half of respondents (27 individuals) were able to successfully complete all tasks and an additional 31.5 percent (17 people) succeeded in locating four of the five types of information sought (see Table 2 for details). However, the remaining ten people were only able to successfully complete 1-3 tasks. This is a considerable proportion given that people were encouraged to pursue tasks without any time constraint. Table 3 describes in detail what proportion of respondents was successful for each task with additional information on the amount of time people took to successfully complete each task.
Table 2: Number of Successfully Completed Tasks
Number of Tasks Completed Successfully Number of Respondents
(% in parentheses)
1 1 (2) 2 3 (5.5) 3 6 (11) 4 17 (31.5) 5 27 (50)
Table 3: Success Rate and Average Time to Successful Completion by Task
Task Success Rate (%) Mean (mins) Standard Deviation Minimum Time Maximum Time (min) N (success) Music 94.44 1.55 1.67 5s 7.83 51 Tax forms 94.44 2.48 2.04 27s 8.75 51 Local events 87.04 2.18 2.03 24s 9.78 47 Kids' art 85.19 2.09 1.67 11s 8.05 46 Political info 61.11 3.79 3.10 27s 13.53 33
Note the low rate of successful completion (61 percent) for the task that required respondents to find a Web site that compares different presidential candidates' views on abortion. To control for previous experience with such a task, respondents were also asked if they had looked at political campaign information online before. Among those who had not, 54.2 percent completed the task successfully in contrast to the 66.7 percent among those who had experience with looking for this type of content online. Although there is some difference among those who had previous experience with this task and those who did not, there are significant proportions of people in both categories who lacked the skill to successfully complete this task.
This finding has important implications when considering the potential effects of the Internet on political participation and its ability to inform citizens on political issues. A large percentage of users were unable to find a political comparison Web site even in a situation where they are not constrained by time and are not being distracted by other obligations and activities. This suggests that people have a very hard time finding political information that may be helpful to further their understandings of candidates' views in a political campaign. Although there are numerous resources on the Web that showcase this type of information, the mere presence of such content will be of little use to advancing political participation if people are not capable of finding their way to such sites.
Overall, people spent anywhere from two and a half minutes to 33 minutes on the five tasks regardless of whether they were able to successfully complete them or not. Figure 1 shows the cumulative distribution function of the total time spent on all tasks. The y axis denotes the proportion of people still working on a task and the x axis shows the amount of total time spent on searching. The circles on the graph signify the amount of time the respondent took to search for the five types of content on the Web.
Figure 1: Total Time Spent on All Five Tasks
There is a gradual increase in the amount of time people spent on all the tasks. Half of respondents were done in 12 minutes or less (so on average these people spent 2.4 or less minutes on each task), but the rest took longer, spending as much as 33 minutes on the tasks. Four respondents spent less than a total of five minutes on all tasks whereas four respondents spent more than 30 minutes on the five tasks.
The large variance is not simply due to people searching endlessly without successful completion resulting in lengthy search sessions. Figure 2 shows the distribution of time spent on the five tasks for those 27 respondents who successfully completed all five tasks. Their times-to-completion range from 2.5 minutes to 30.3 minutes suggesting that the quickest respondent successful with all tasks was as much as twelve times quicker than the slowest fully successful respondent.
Figure 2: Total Time Spent on All Tasks for Those Successful with All Five Tasks
How can we explain these differences in people's ability to find content online? Here, I look at the relationship between online skills and age, gender, education, and experience with the technology. I consider both the mean number of successfully completed tasks by subgroups and the average amount of time (in minutes) spent on all five tasks. I report time spent on all five tasks regardless of successful completion. The results are robust when only considering the total time of those who were successful with all five tasks.
There are clear generational differences in people's ability to use the Web. Table 4 shows the average number of successfully completed tasks by age broken down by decades. The 18 and 19 year olds were all successful with all five tasks whereas the people in their 70s and 80s averaged 3.33 successful tasks with people in their 60s also able to finish less than four tasks on average. The same table also presents information about the average time people spent on the five tasks broken down by age in decades. Again we see clear generational differences. People in their teens and people in their 20s are quicker than people in their 30s and 40s who are quicker than older respondents, although curiously, people in their 60s were relatively quick at finishing tasks.
Table 4: Average Number of Successfully Completed Tasks and Time Spent on Tasks by Age
Age by Decade Mean Number of Tasks Standard Deviation of Tasks Average Time on Tasks (min) Standard Deviation of Time Number of Observations 10s 5 0.00 6.7 3.96 3 20s 4.67 0.49 8.2 3.97 12 30s 4.17 1.33 15.7 6.11 6 40s 4.21 0.58 14.0 6.94 14 50s 4.13 1.13 19.1 8.44 8 60s 3.75 1.49 13.5 6.49 8 70s+ 3.33 1.15 24.4 7.50 3
On average, women completed 4.19 tasks compared to men's 4.26 average success rate. The average total time spent on the five tasks for women was 14.6 minutes whereas for men it was 12.9 minutes. Neither of these differences is statistically significant, suggesting that there is no influence of gender on whether people are able to efficiently navigate the content of the Web and how long they take to do so.
Education has consistently been a predictor of access to the Internet (NTIA 1995, 1998, 1999, 2000) and is likely to affect the level of Web use skill as well. Universities were the first ones to embrace the technology and thus those who attended college in the past decade would have had exposure to the medium during their schooling. Moreover, people with higher levels of education are likely to have had more exposure to computer technology in general, familiarity with which is an important first step in gaining access to the Internet. Moreover, higher educated individuals are more likely to possess the expertise and confidence with technology needed to download and install additional software, a necessary prerequisite for the browsing of many Web sites.
Here, respondents' level of education is broken down into three categories: less than a college degree, college degree and graduate degree. Table 5A shows the relationship of educational level to online skills. Those with the highest level of education do best in terms of the number of tasks completed while those with the lowest level of education are the quickest in completing tasks.
Table 5A: Average Number of Successfully Completed Tasks and Time Spent on Tasks by Education
Education Mean Number of Tasks Standard Deviation of Tasks Average Time on Tasks (min) Standard Deviation of Time Number of Observations No college degree 4.43 .94 10.4 5.13 14 College degree 3.78 1.21 18.4 8.44 18 Graduate degree 4.45 .67 12.0 6.42 22
However, because those currently in school are likely to have quite a bit of exposure to the technology, and because they would show up in an educational category that does not yet represent the degree they are currently seeking, in Table 5B, I present the relationship between education and online skill excluding those respondents who are currently in school (six college students and one graduate student were thus excluded from this table). Here, we see that those with a graduate degree do best both with successful completion of tasks and amount of time spent on tasks.
In the last column of this table, I have included the average age of those in the three educational subcategories because we have seen earlier that age is significantly related to level of Web skill. Consistent with that finding, we find that the relationship of education to skill noted in this table may be driven by the considerable differences in the average age of those who have no college degree (40) and those who do (48 for the college graduates and 47 for those with an advanced degree). The higher level of skill in the lowest educational category may thus be driven by the considerably lower average age of those respondents. Since the average age is similar for the latter two categories they are more comparable (as this controls for differences in age). We can see that those with a graduate degree are considerably better at finding information online than those who do not have an advanced degree when controlling for age.
Table 5B: Average Number of Successfully Completed Tasks by Education and Time Spent on Tasks Excluding Those Currently in School
Education Mean Number of Tasks Standard Deviation of Tasks Average Time on Tasks (min) Standard Deviation of Time Number of Observations Mean Age in Educational Subcategory No college degree 4.125 1.13 12.9 3.89 8 40 College degree 3.78 1.21 18.4 8.44 18 48 Graduate degree 4.42 .68 12.3 6.45 21 47
Prior experience with the technology
The amount of prior experience with the Internet is likely to affect online actions (Howard, Rainie, and Jones, 2001). People who spend more time online - whether at home or any other location - will likely acquire more knowledge about the Web and thus will have better online skills. Moreover, people who have been Internet users for longer are expected to be better at finding information online as they have more experiences to draw on. Additionally, early adopters tend to be more innovative, with a greater willingness to explore a new medium and familiarize themselves with it (Howard, Rainie, and Jones, 2001; Rogers, 1995).
The data suggest that the amount of time people spend online does affect their efficiency in finding information on the Web but is of most concern for those who use it minimally. According to Table 6, those who browse the Web less than an hour each week are able to find less information and take considerably longer on tasks than those who spend either 1-7 hours online per week or those who spend even more time surfing the Web weekly.
Table 6: Average Number of Successfully Completed Tasks and Time Spent on Tasks by Time on Web Weekly
Education Mean Number of Tasks Standard Deviation of Tasks Average Time on Tasks (min) Standard Deviation of Time Number of Observations Less than one hour 3.20 1.78 18.9 8.37 5 1-<7 hours 4.22 .93 13.7 7.90 27 7 or more hours 4.45 .67 12.7 6.80 22
Table 7 sheds light on the importance of being a veteran versus a newcomer online. Those who only recently started using the Internet - in the past three years - exhibit considerably lower online skills than those who have been online for longer. However, note again the age distribution among the subcategories. Late adopters are significantly older and so it may be a combination of age and time that is driving the differences in Web use skills.
Table 7: Average Number of Successfully Completed Tasks and Time Spent on Tasks by Number of Years Since First Internet Use
Number of Years Since First Use of the Internet Mean Number of Tasks Standard Deviation of Tasks Average Time on Tasks (min) Standard Deviation of Time Number of Observations Mean Age in Subcategory 0-2 years 3.18 1.25 19.6 6.01 11 52 3-6 years 4.47 .61 14.3 7.67 19 41 7 or more years 4.50 .78 10.6 6.56 24 40
In general, young people (late teens and twenties) have a much easier time getting around online than their older counterparts (whether people in their 30s or 70s). Some of this is clearly based on comfort with the technology they are using and not necessarily based on elaborate techniques they have mastered specifically with respect to the Web. As the amount of time people have been Internet users also matters in how well individuals are able to navigate the content of the Web, it is possible that those who are currently less skilled will learn over time and improve their ability to find content online. However, people may be discouraged by the difficulties of finding information on the Web and thus may end up spending less time with the medium. Given that time spent on the Web is also associated with level of Web skill, lower level skills may persist over time.
It is clear from the findings that there is great deal of variance in abilities to locate content online. Merely offering people a network-connected machine will not ensure that they can use the medium to meet their needs because they may not be able to maximally take advantage of all that the Web has to offer. Policy decisions that aim to reduce inequalities in access to and use of information technologies must take into consideration the necessary investment in training and support as well. Like education in general, it is not enough to give people a book, we also have to teach them how to read in order to make it useful. Similarly, it is not enough to wire all communities and declare that everyone now has equal access to the Internet. People may have technical access, but they may still continue to lack effective access in that they may not know how to extract information for their needs from the Web. Although providing Internet access may help alleviate some problems of the digital divide, information presented in this paper demonstrates that a second-level digital divide exists relative to specific abilities to effectively use the medium.
About the Author
Eszter Hargittai earned her BA in Sociology at Smith College and is currently a PhD candidate in Sociology at Princeton University. She has published on the comparative history of the radio's and Internet's early years, inequalities in the global diffusion of the Internet, how portal sites channel users toward some content and away from other online material, analysis of international telephone networks, and on conceptualizing the "digital divide". Her current project is an empirical look at how people find information online and how the organization and presentation of content on the Web influences people's online actions. She has recently launched an e-mail distribution list - "Eszter's List" - with pointers to interesting material on the Web. For more information, see http://www.eszter.com.
I would like to thank Paul DiMaggio for his insightful comments throughout this project, Stan Katz for his ongoing support, and Erica Field for helpful discussions. I am also grateful to Edward Freeland, James Chu, Carolyn Mordas, Jeremy Davis-Turak, and Inna Barmash for their assistance with various components of the project. Generous support from the Markle Foundation is kindly acknowledged. The project has also been supported in part by NSF grant #SES9819907, a grant from the Russell Sage Foundation, and through a grant from the Pew Charitable Trusts to the Center for Arts and Cultural Policy Studies, Princeton University. I am also grateful to the Fellowship of Woodrow Wilson Scholars at Princeton University.
Benton Foundation, 1998. "Losing Ground Bit by Bit: Low-Income Communities in the Information Age," at http://www.benton.org/Library/Low-Income/, accessed 25 March 2002.
B. Bimber, 2000. "The Gender Gap on the Internet," Social Science Quarterly, volume 81, number 3, pp. 868-876.
E.P. Bucy, 2000. "Social Access to the Internet," Harvard International Journal of Press/Politics, volume 5, number 1, pp. 50-61. http://dx.doi.org/10.1162/108118000568967
V. Cothey, 2002. "A Longitudinal Study of World Wide Web Users' Information-Searching Behavior," Journal of the American Society for Information Science and Technology, volume 53, number 2, pp. 67-78. http://dx.doi.org/10.1002/asi.10011
P. DiMaggio and E. Hargittai, 2001. "From the 'digital divide' to 'digital inequality': Studying Internet use as penetration increases," Princeton University Center for Arts and Cultural Policy Studies, Working Paper Series number 15.
E. Hargittai, in press. "Beyond logs and surveys: In-depth measures of people's Web use skills," Journal of the American Society for Information Science and Technology Perspectives.
D.L. Hoffman and T.P. Novak, 1999. "The Evolution of the Digital Divide: Examining the Relationship of Race to Internet Access and Usage Over Time," at http://ecommerce.vanderbilt.edu/research/papers/pdf/manuscripts/EvolutionDigitalDivide-pdf.pdf/, accessed 25 March 2002.
P.E.N. Howard, L. Rainie, and S. Jones, 2001. "Days and Nights on the Internet: The Impact of a Diffusing Technology," American Behavioral Scientist, volume 45, number 3 (November), pp. 383-404.
B.J. Jansen and U. Pooch, 2001. "A Review of Web Searching Studies and a Framework for Future Research," Journal of the American Society for Information Science and Technology, volume 52, number 3, pp. 235-246. http://dx.doi.org/10.1002/1097-4571(2000)9999:9999<::AID-ASI1607>3.0.CO;2-F
J.E. Katz and R.E. Rice, 2002. Social Consequences of Internet Use: Access, Involvement and Interaction. Cambridge, Mass.: MIT Press.
R. Kling, 1998. "Technological and Social Access on Computing, Information and Communication Technologies," White Paper for Presidential Advisory Committee on High-Performance Computing and Communications, Information Technology, and the Next Generation Internet, at http://www.slis.indiana.edu/kling/pubs/NGI.htm, accessed 26 March 2002.
W.E. Loges and J.-Y. Jung, 2001. "Exploring the Digital Divide: Internet Connectedness and Age," Communication Research, volume 28, number 4 (August), pp. 536-562. http://dx.doi.org/10.1177/009365001028004007
National Telecommunications and Information Administration, 1995. "Falling Through the Net: A Survey of the "Have Nots" in Rural and Urban America," at http://www.ntia.doc.gov/ntiahome/fallingthru.html, accessed 25 March 2002.
National Telecommunications and Information Administration, 1998. "Falling Through the Net II: New Data on the Digital Divide," at http://www.ntia.doc.gov/ntiahome/net2/falling.html, accessed 25 March 2002.
National Telecommunications and Information Administration, 1999. "Falling through the net: Defining the digital divide," at http://www.ntia.doc.gov/ntiahome/fttn99/contents.html, accessed 25 March 2002.National Telecommunications and Information Administration, 2000. "Falling Through the Net: Toward Digital Inclusion," at http://www.ntia.doc.gov/ntiahome/fttn00/contents00.html, accessed 25 March 2002. National Telecommunications and Information Administration, 2002. "A Nation Online: How Americans Are Expanding Their Use of the Internet," at http://www.ntia.doc.gov/ntiahome/dn/html/anationonline2.htm, accessed 25 March 2002.
P. Norris, 2001. Digital Divide: Civic Engagement, Information Poverty and the Internet in Democratic Societies. New York: Cambridge University Press.
Panicware, Inc., 2001. "Don't Panic! 4.0," at http://www.panicware.com/, accessed 25 March 2002.
Pew Internet and American Life Project, 2000. "Tracking Online Life: How Women Use the Internet to Cultivate Relationships with Family and Friends," at http://www.pewinternet.org/reports/pdfs/Report1.pdf, accessed 25 March 2002.
E. Rogers, 1995. Diffusion of Innovations. New York: Free Press.
S. Strover, 1999. "Rural Internet Connectivity," at http://www.rupri.org/pubs/archive/reports/1999/P99-13/, accessed 25 March 2002.
P. Wang, W.B. Hawk and C. Tenopir, 2000. "Users' Interactions with World Wide Web Resources: An Exploratory Study Using a Holistic Approach," Information Processing and Management, volume 36, number 2, pp. 229-251. http://dx.doi.org/10.1016/S0306-4573(99)00059-XE.J. Wilson, 2000. "Closing the Digital Divide: An Initial Review," at http://www.internetpolicy.org/briefing/ErnestWilson0700.html, accessed 25 March 2002.
Appendix: Demographics of Respondents
Education: <HS = less than high school; HS/GED = high school or equivalency exam; SC = some college; ASD = associates degree; BA = bachelor's degree (either arts or sciences); MA = Master's (either arts or sciences); PRO = professional degree; PhD = doctoral degree
a African American;bAsian American;c Hispanic
Occupation Age Sex Education Time on Web/week (in hours) Number of Years Since First Use of the Internet Student 18 F HS/GED 3 7 At homea 18 F HS/GED 10 5 Library office assistant/Student 19 F SC 10 4 Accounting manager assistant/Student 20 F SC 2 11 Sales data manager/Student 20 F SC 4.5 7 Bank customer service representative/Student 21 F SC 5 4 Computer lab tech support/Student 22 M SC 70 7 Financial markets data managerc 24 M BA 30 5 Financial services sales representative 24 M BA 21 6 Financial analyst 25 M BA 1 11 Golf course groundskeeper 26 M HS/GED 7.5 2 Hydrogeologist 26 F BA 5 7 Non-profit administrator 26 F MA 6 7 Graduate student (humanities)a 27 M PhD 10 11 Civil engineer 29 M BA 17.5 6 U.S. military pilot 33 M MA 7 6 Non-profit computer programmerb 34 M MA 25 7 Local government GIS administrator 36 F MA 15 6 Assembly worker (air conditioners)a 39 M <HS 0.13 1 Small company MIS director 39 M HS/GED 7 6 At home 39 F BA .17 1 Self-employed real-estate agent 40 F ASD 1.5 2 Medical equipment repair and installation 40 M BA 10 11 At home 40 M MA 4 9 Professor (mathematics) 40 M PhD 8 9 Circulating nurse 41 F ASD 9 8 Legal researcher for State 41 M PRO 5 8 Unemployed 42 F MA 21 8 Security officer 42 M PRO 4 3 Personal/professional concierge (self-employed) 43 F BA 8.5 2 Retail data administratora 44 F BA 0.25 2 Self-employed consultant 44 F MA 7 5 Environmental consultant 45 M MA 3 5 School teacher 47 F MA 3 10 Social work supervisor 48 F MA 1 6 Sales accountant 50 M BA 5 5 Professor (humanities) 50 F PhD 0.5 13 Utility company budget analystc 50 M BA 5 7 Retail customer service 52 F MA 4 4 Management consultant (comp info systems) 52 M MA 10 16 Technician manager 55 M ASD 3.5 12 Business development 56 M BA 3.5 4 Retired 57 F MA 5 11 Development officer 60 M PhD 7.5 8 Personal assistant to psychologist 62 F ASD 18 7 Retired 62 M MA 14 3 Retired 63 F BA 3 2 Crisis counselor 64 F BA 0.5 2 Real estate broker 66 M BA 15 2 Pharmaceutical products customer support rep. 68 F BA 1 2 Retired (was: scientific equipment marketer) 69 M PhD 10 6 - 72 F BA 10 13 Software programmer 72 M BA 3.5 5 Store sales staff (formerly company tech staff) 81 M BA 1 0
Paper received 21 March 2002; accepted 25 March 2002.
Copyright ©2002, First Monday
Second-Level Digital Divide: Differences in People's Online Skills by Eszter Hargittai
First Monday, volume 7, number 4 (April 2002),
A Great Cities Initiative of the University of Illinois at Chicago University Library.
© First Monday, 1995-2014.