Troubles with the Internet:
The Dynamics of Help at Home

Sara Kiesler, Vicki Lundmark, Bozena Zdaniuk, Robert Kraut,

William Scherlis, & Tridas Mukhopadhyay

Carnegie Mellon University

November, 1999

Abstract

Despite advances in technology, nearly everyone experiences technical difficulties and challenges using a home computer. In a field trial of household Internet usage, 89 percent of 93 families needed external technical support in the first year they used the Internet. However, usually only the most technically involved members of the family requested external technical support, and this behavior was associated with other computer-related behaviors in the household. We explore the process by which a family member with comparatively high technical skill or enthusiasm, often a teenager, becomes the most involved Internet user, the family guru to whom others in the family turn for technical help, and the person who makes external help requests. The family guru benefits from this role, influences the householdís adoption of technology, and represents an important link between households and computer support professionals. The role also is a fascinating social phenomenon in the evolution of inter-generation relationships.

Troubles with the Internet: The Dynamics of Help at Home

Although computer technology is practically ubiquitous in workplaces and growing increasingly popular at home, computers still pose substantial technical challenges to their users. The popularity of computer advice columns and training courses, the rise of the usability engineering profession, and the large budget allocated to computer support divisions attest to the complexities people encounter when using computers. Revenues of the problem resolution industry reached nearly $1.4 billion in 1997 and could increase more than a third by 2002 (Hoffman, 1999). Even technical workers with advanced computer skills encounter usability problems when they try to learn new systems and programs (e.g., Barley, 1988; Orlikowski, 1996).

In the workplace, when employees need technical help, they often can turn to in-house professional technical staff or to expert coworkers. At home, when people need technical help, experts and professionals may be unavailable. Customer support lines exist to answer peopleís questions, but these services impose attentional, monetary, or psychological costs that can discourage people from using them. With the modal U. S. household having only two individuals in 1998 (mean household size = 2.6, U.S. Department of the Census, 1998), if one family member is confused, there aren't lots of others to turn to.

In this article, we examine how individuals acquire technical support at home and the possible consequences of that support for themselves and the household as a whole. This article draws from a longitudinal study of familiesí first experiences with the Internet. Analyses of how family members used the Internet and its social impact after their first years online have been reported in Kraut, Scherlis, Mukhopadhyay, Manning, & Kiesler, 1996; Kraut, Patterson, Lundmark, Kiesler, Mukhopadhyay, & Scherlis, 1998, and Kraut, Mukhopadhyay, Szczypula, & Kiesler (in press). The current article uses data on technical problems and support that have not been presented in previous papers.

Setting

The data we analyze here are from 93 Pittsburgh families and 237 family members whom we provided in 1995 and 1996 with a computer and access to the Internet. The goal of the project, called HomeNet, was to track residential Internet use for at least a year and to assess its social impact. The sample was designed to be diverse, to help us understand income, education, and skill barriers to using the Internet. The first group in the sample consisted of 44 families with teenagers, who began using the Internet at home in spring, 1995. A second group of 49 families with teenagers or with an adult on the board of a community development organization began using the Internet about a year later. Children under age 10 and uninterested family members were not in the study. Details of sample selection are available at Kraut et al (1998).

The families were recruited through 4 high school journalism programs and 4 community development organizations in 8 Pittsburgh neighborhoods. The socioeconomic status of the neighborhoods varied, with an overall median household income of $42,500, somewhat higher than the U.S. median, but lower than that of Internet users in 1995 (Anderson, Bikson, Law, and Mitchell, 1995). About 25% of the households recruited were minority, mainly African-American. Within the households, about 60% of the family members in the study were 19 years old and older, and 51% were female. Twenty-five percent of the sample regularly used a computer at work, but the project excluded individuals if they had an active Internet connection from home or work. In sum, our sample is a diverse group of families and family members, most of them in their first experience with the Internet. Compared with more recent demographics of households online, our sample is similar except for our having about double the percentage of minority users (IDC, 1998). Our sample is not unique by virtue of its focus on novices in homes. As of 1998, a majority of people online at home had used the Internet there for a year or less (IDC, 1998).

The HomeNet project gave all families an Internet-ready package comprised of a Macintosh computer with pre-installed Netscape home page and ClarisWorks Office software, a modem and extra phone line, phone and Internet service, personal email accounts, training for family members, and regular hours of telephone and email access to a help desk (see Kraut et al., 1996, for more details on the package). Help desk staff included a professional technical director and Carnegie Mellon undergraduates majoring in computer science or information systems, with training in teaching basic computer skills. The help desk staff used beepers to receive calls and a program called Timbuktu (Netoptia, 1995), which allowed them to view and manipulate the participantsí computers remotely. When necessary, help desk staff visited homes to diagnose problems and make repairs.

The need for technical help

HomeNetís package was intended to reduce technical barriers to use of the Internet and to minimize participantsí requests for support services. Nonetheless, technical barriers remained. Over 70% of the households needed technical support just to set up their computer and connect it to the Internet for the first time. Eighty-nine percent of the households requested technical support from the help desk staff over the first year of the trial. The technical problems that family members encountered were numerous and diverse, ranging from not understanding that a modem and a telephone could not be used simultaneously on the same line, having to fix corrupted aliases, recovering from mistakenly reformatting a hard disk, being unable to debug a dial-in script that reached the local supermarket rather than the ISP, needing help with setting up peer-to-peer chat, and needing advice on how best to convert VHS video to digital format.

Table 1. Examples of participantsí computer and Internet problems reported to the help desk.

Many households made multiple requests to the help desk, consuming unexpectedly high project resources. Even though most families called the help desk at least once, more than half of the individual family members never called. Both this pattern and interviews with families suggested to us that family members adopted divergent practices for dealing with technical problems. At one extreme were those who simply gave up using their computer and the Internet without seeking external help. Others muddled through, or sought and got help from friends, family, or books. A third group relied heavily on help desk staff to solve their problems and to help them learn new skills. Two participants called the help desk over 30 times in the first year of using the Internet.

Giving and getting help

Who were the people who sought out technical support from the help desk. One might expect people who seek help to be those who need help mostófor example, computer novices, who need help to get started. Presumably, the benefits of help would accrue primarily to those seeking it, allowing them to take advantage of new technology and services. Yet this scenario According to this view, one would expect that people who are initially least skilled at using computers and who use them least would be the most frequent seekers of help. In addition, the help would benefit them, by allowing them to use the Internet more than they would have otherwise. Yet thisis inconsistent with research in many domains, including work on organizational learning (Cohen & Levinthal, 1990), politics (Neuman, 1986), consumer behavior (Punj & Staelin, 1983) and the diffusion of innovations (Rogers, 1996), all of which suggest that new knowledge accrues most to those who already have a substantial amount of it. People who are more interested and skilled in a domain are more likely to realize what they do not know, are more likely to have the confidence to challenge themselves and to stretch the limits of their expertise, and they are more likely to have the interest and background knowledge to seek out and benefit from additional instruction.

Although little is known about the dynamics of giving and receiving technical advice within the family, analogies with the world of work may be useful in understanding the phenomenon. The family, as a compact unit loosely tied to its supporting environment, is analogous to a work group. In the workplace, technical knowledge and advice typically flows into a work group through specialized information gate keepers (Allen, 1976; Tushman & Katz, 1982; Constant, Sproull, & Kiesler, 1996). Information gate keepers have distinctive personal attributes, such as having more seniority, competence, and organizational authority and centrality than those to whom they pass advice. They also have distinctive network positions¾ many social ties with people outside the group, from whom they import knowledge to the group, and many ties within the group, to whom they can pass on their knowledge. Acquiring new knowledge benefits both the gate keeper and the group. Once gate keepers import information into the group, it travels through the group 's pre-existing social networks, especially communication links among friends (Allen, 1976) and among those who occupy similar organizational niches or among those who have similar personal attributes (e.g., Rogers, 1995, p. 286)). , although specialized advice networks also exist (Krackhardt, 19xx?).

If processes underlying the spread of technical knowledge within work groups also apply to the spread of technical knowledge within families, we would expect the following: (1) Individuals who are initially most motivated and skilled in using computers and the Internet will be most likely to seek external help. (2) The seeking and giving of technical advice will be a specialized role within the family: If one person performs it, others need not. (3) Seeking help will benefit the entire household, not only the individual seeking it. (4) Trans-generational advice is likely to flow from parent to child (because parents have more seniority and more ties outside the family than children do), rather than the reverse. (5) Advice will flow through the family along already existing contact patterns, as between children similar in age and gender, who spend time together. (6) Gender segregation into more and less technical tasks will cause males to be more often the carriers of technical information than will females. As in many work organizations, where men predominate in technical jobs like programming and women predominate in nontechnical jobs like administrative assistance, many household tasks remain gender segregated (Berk, 1985). Taking care of household technology (car maintenance, household repairs) has been considered a masculine domain (White & Brinkerhoff, 1981; Bryant and Zick, 1996); hence one might expect that males would be the gate keepers when dealing with computer and Internet problems.

Overall, in this view, with seniority controlled, adults would provide advice to children and other adults, children would provide advice to those of the same gender, and males would provide advice to others in the family.

Differences between work groups and families in composition, values, and activities could undermine these predictions, however. For example, in contrast to work groups, families with children often devote considerable attention to the developing human capital of children. Efficiency of the unit is an explicit value in workgroups, but in modern families, the value of efficiency is downplayed. Instead, play, relaxation, companionship, and the socialization of children more important. Many families view computers as valuable because they enhance childrenís education and learning, and are comparatively safe entertainment. Hence, families may foster childrenís spending considerable time with computers and they may be especially encouraging of the development of technical computer skill by the children. Children in families also have more free time than adults do. Teenagers, in particular, have time for, and license to, take on technical challenges and develop skills to master them. Many teens, in turn, are fascinated with computer technology. Compared to adults, they are very heavy uses of computers and the Internet at home (Kraut et al, 1996). The likely consequence of these unique characteristics of families is that teenagers might be more important seekers of external help and sources of help within the family than one would expect based on the organizational literature. On the other hand, their youth and lower status within the household and their often stormy relationships with their parents could prevent teens from effectively sharing the knowledge that they develop through use and experimentation.

Traditional family dynamics condition who spends time with whom within the household, and as a result may channel the flow of technical advice as well. For example, parents tend to interact more with the child of their own gender (Bryant & Zick, 1996) and this trend tends to increase when children reach adolescence (Crouter, Manke, & McHale, 1995). Such interaction patterns may create more opportunity for exchanging computer help within same-sex parent-child pair than within the opposite-sex pair. We explore the gender and generational flow of technical help within the family in light of these earlier findings.

Method

Within the 93 families in the study, 237 members signed consent forms, were given email accounts on the Internet, and logged on at least once. The participants were followed over time, their Internet usage was monitored continuously, and they completed questionnaires before, during, and after one year. We used statistical analyses of help desk logs, usage data, and questionnaires to test the hypotheses, and qualitative data from home interviews to help us understand the motivations that led to the statistical patterns.

Requests for support from the help desk

Help desk staff completed a form for each telephone conversation they handled and they saved an archive of email messages received by the HomeNet help desk about any computer or Internet support. From these request logs, we tallied all telephone or email contacts each person and each family made with the help desk within the first 52 weeks of their householdís start date. We then counted the frequency of requests made by each person. We also tallied the number of requests within each household and computed for each person the number of other family members who requested help. Because help desk requests were skewed, we use log transformations of the data in our analyses.

Internet usage

Computer-generated usage data were collected by setting automatic probes on every personís Internet account, which identified each time the person logged on or off. To increase reliability, the computer-generated usage data were summed each week. Our usage variable, connect hours, is an average based upon the total hours each week that participants were connected to the Internet during the period when they were still active. In other words, this variable reflects usage for the entire first year or until the time a participant stopped using the Internet altogether. We also created two companion variables to distinguish average weekly connect hours in the first and last 26 weeks of the first year.

From the same automated logs of Internet connect hours, we also measured participants' survival online, the length of the period during their first year, when they were active Internet users. This is the period from a participant's first week of use to his or her last week of use (i.e., the week after which there was no further activity recorded). We credited participants who used the Internet after week 52 as surviving on the Internet for a full 52 weeks. We treated the data as right censored for an individual, if he or she moved out of the household or if the household computer were no longer available. Censoring primarily occurred when students moved off to college, terminating their participation in the data collection. If they took the household computer with them, this terminated data collection for the other household members as well.

Questionnaire measures

All participants completed a pretest questionnaire before they received their computers and Internet service, and a posttest questionnaire after a year had gone by. Short questionnaires were administered occasionally during the trial and after the first year. The first variable we discuss below is from one of the short questionnaires; the others are from the pretest and posttest questionnaires.

Computer skill. A 5-item scale using 5-point Likert ratings on the pretest and posttest asked participants how much they agreed with the following statements: I am very skilled at using computers; I use computers almost every day; I am afraid of using a computer; using computers is fun; I donít know much about using computers (Cronbachís alpha = .85).

Demographic attributes. From questionnaire items on the pretest, we created dummy variables for the analyses to represent generation (adults vs. teens under age 19), gender (males vs. females), and race (whites vs. minorities). These demographic attributes significantly predicted Internet usage in previous analyses (Kraut et al., 1996).

Helping others in the family. In May, 1997, we administered a short questionnaire to participants who were living at home. One of the questions asked the respondents to indicate the names of those in their family who help them in using the computer. The answers were transformed to create for each respondent a score indicating how many family members the person helped. To control for family size, this score was divided by the number of family members who answered the question. Thus, the variable, giving help, reflects the proportion of family members who named the participant as a helper. The answers to the above question were also used to examine gender and generation attributes of help givers and help receivers within the family.

Home interviews

The interviews from which we quote in this article were conducted in the homes of 25 participant families. We asked all family members to be present. Interviews were approximately three hours long and were conducted by three project staff members, one of whom videotaped the interview. The interview began with a discussion of family routines and interaction, and then a tour of the house. In the latter half of each interview, each member of the family who used the computer sat at the machine, connected to the Internet, and took an interviewer through the personís usual interactions (logging on, connecting to the Internet, reading email, using the Web, etc.). Interviewers encouraged participants to talk about their experiences using the Internet.

Results

First, we report simple descriptive statistics for the variables in the study (Table 2). Just under half of the individuals in the sample made contact with the help desk. (About half of those made 1-3 requests for external help; the remainder made 4 or more requests.) Table 2 shows that being a teenager, having more pretest computer skill, and having higher weekly Internet usage through the year were correlated positively with making help desk requests. Also, making help desk requests was correlated with longer survival online; that is, those who requested external support were more likely to continue using the Internet. Those who gave more technical help to other family members also tended to have more pretest computer skill, to use the Internet for more hours per week, and to survive longer online.

Table 2. Correlations among variables describing individual participants in their first year of using the Internet.

Predicting who requests external technical support

We used regression analysis to examine which variables uniquely predicted requests for technical help from the help desk, holding other variables constant. Table 3 shows tests of hypothesis 1, that pretest computer skill and Internet usage would predict the frequency of requests to the help desk. Since the frequency of help requests and the usage indicator, connect hours, were skewed variables, the analysis uses the log of the number of requests and the log of connect hours.

Table 3. Standardized regression coefficients predicting the number of help desk support requests made by participants (logged).

Table 3 shows that Internet usage was the most important predictor of a personís help desk requests. All other variables lacked statistical significance except for help desk requests by others in the family (a negative predictor that we discuss below). Given the significant bivariate correlations between making help desk requests and pretest skill and being a teenager, these findings suggest that the influence of prior technical skill and age on help seeking is mediated by ongoing involvement with the computer and the Internet.

Motivations. In home interviews we explored family membersí motivations for asking or not asking for external technical support. We did this by probing the events surrounding family membersí computer problems and how family members learned new computer skills. Calling the help desk was rarely the first response of anyone faced with a computer problem or a new task they wished to accomplish. Those who regularly used the Internet typically explored possibilities on their own or they asked advice of other family members. One 18 year old said,

"Iím basically a trial and error person. I learn a lot of things by myself. I donít like to sit down and listen to people telling me how to do stuff unless I know I have a problem in a certain area (H)."

When this teen had a problem she couldnít solve, she would then ask her older brother for help:

"At first I didnít know what I was doing. . .Iím saying, Ďwhatís going on?í and I asked my brother. ĎOh, itís downloading; itís in the computer now; itís not on the home net.í Itís like, oh OK ."

This teen called the help desk when her brother was unavailable. By contrast, those who did not call the help desk seemed to be comparatively inexperienced, timid users, an impression borne out by the statistical findings. These participants were uninterested in vigorously trying to solve their computer problems and were deterred by the costs of doing so. Participants said they would put off dealing with computer problems ("I would leave it for later"), hated waiting for help, and disliked thinking about computer problems, as in this exchange during a home interview:

Mom: . . . there were times when I would just throw up my hands and get up and walk away.

Daughter: Yeah, it would get me to the point where I would just [say] Ďforget ití (R).

Some participants could not diagnose their problems and did not have the vocabulary or background knowledge to discuss what went wrong with a technical person. An 18 year old said, "I swear to God, every time I go to use the computer it doesnít want to work for me. It doesnít like me (A)." Compounding their lack of knowledge, some inexperienced users were too embarrassed to call on technical help, as is revealed in this exchange during a home interview:

Q: Do you know what that is? [Interviewer points to hard drive icon on the desktop, which lacks a text label]

A: My granddaughter did that. I donít know why it went black.

Q. Did you call the help desk?

A. No, I thought we broke something. (M)

Specialization within families: family guru

In home interviews, those with an ongoing interest and hands-on involvement with the computer and the Internet described calls to the help desk when they needed technical help whereas others in the same family, typically less involved, did not. We hypothesized (Hypothesis 2) that help requests would be statistically concentrated within families, evidence of specialization. The negative correlation between participantsí help desk calls and calls by others in the family supports this hypothesis (Table 2. If family members can receive help from another family member, then they need not seek it externally. Table 2 also shows that those who requested external technical support from the help desk were also significantly more likely to give technical help to other family members. The regression analysis further supports the specialization hypothesis in that it shows, controlling for demographics and pretest skill, fewer help desk requests by others in the family help significantly predicted more help requests by a participant (Table 3).

In most of the families, one person made almost all contacts with the help desk during the year. In families with 2 participants (n = 34), the mean number of people who requested help from the help desk was 1.2; in families with 3 participants (n = 23), the mean number of people who requested help from the help desk was 1.4; in families with 4 participants (n = 10), the mean number of people who requested help from the help desk was 1.6; in families with 5 participants (n = 6), the mean number of people who requested help from the help desk was 1.5; and in the 2 families with 6 participants the mean number of people who requested help was 2.

To further pursue the possibility that calling the help desk was part of a specialized role as family computer guru, we evaluated characteristics of the top help desk caller in each family. Table 4 presents a household-level analysis in which we examine the attributes of the family member who called the help desk most often versus the rest of the family. Eighteen families with only a single Internet user were not included in this analysis. Table 4 shows that, within families, the person who made the most help desk requests did so significantly more frequently than others. Across all families, the participants who called the help desk most frequently accounted for 38% of all family members, but 80% of all calls. Those who made the most help desk calls also had significantly greater Internet usage than others in the family and more, but not significantly more, pretrial computer skill.

Table 4. Comparing participants who were the top help desk requester in their family with others in their family.

As shown in earlier research (Franzke & McClard, 1996) and suggested by the interviews with household members, getting help from those close byófrom other family members or friendsówas the initial and most common source of advice for technical problems. Who gave this help? Table 2 showed a significant correlation between a personís frequency of requesting technical support from the help desk and the proportion of the family members asking that same person for computer help. Both variables are also correlated with the helperís pretest computer skill and, more strongly, with the helperís Internet usage. Table 4 shows that the top external help requesters gave more computer help within the family than did other family members.

These relationships suggest that the heavy Internet users in a family not only acted to obtain external technical support for themselves and the rest of the household but also served as internal technical resources, i.e., family gurus. If so, we should see that a family memberís computer skill and Internet usage predicted helping others in the family (Table 5). The regression model in Table 5 shows that being a teenager was a significant predictor of giving help to others in the family. Hence, technical knowledge in these families flowed from less seniority to most seniority, the reverse of what has been the case in traditional work organizations. At home, teenagers gave more help to other family members than did adults. Internet usage also was a significant additional predictor of helping others in the family: Those who used the Internet more gave more computer help to other family members.

Table 5. Standardized regression coefficients predicting participantsí help to more family members (weighted by number of family members).

Family dynamics

To explore the interactions surrounding help giving in the household, we calculated the number of all possible pairs of participants within each family in which computer help could have been given. We grouped these pairs by generation and gender of the potential help giver and the potential help receiver. For each generation-by-gender group, we calculated the percentage of pairs in which help was actually given. This analysis is presented in Table 6.

Table 6. Percent of potential helper/recipient pairs in which computer help was reported, sorted by age and gender of pairs.

From older literature on information flow in organizations, we derived that advice in the family would be from adult to child and within age and gender categories, and would be male-dominated. We found that within-age helping characterized parents in our sample: they helped other adults in 30% of the adult-adult pairs, whereas they helped their teenage children in only 16% percent of all child-adult pairs (z=2.41, p<.10).5 However, cross-age helping from less to more seniority characterized the teen helpers. The teens helped other children in only 14% of all child-child pairs, whereas they helped adults in 38% of all child-adult pairs (z=-3.39, p<.05).5 Contrasts of the direction of support in cross-generation pairs showed that adults gave help in 16% of child-adult pairs, whereas teens gave help in 39% of these pairs (z=-3.72, p<.05).

We had hypothesized also that when help-giving occurred, it would tend to follow same-gender paths of communication. Since we found that teens tended to help adults in the household, we then tested whether boys give advice more to men than to women and whether girls give advice more to their mothers than to fathers. The results indicate that boys tend to give advice to both women (38% of all boy-woman pairs) and men (48% of all boy-man pairs) at similar rates (z=.65, ns). Girls tended to give more help to women (49% of all girl-woman pairs) than to men (17% of all girl-man pairs) but this difference was of marginal statistical significance (z=2.66, p<.10).5

Finally, we expected that males would be more likely to give computer advice than females because of the traditional gender role division in families. Our results indicate that men were not more likely than any other age/gender group to give computer help to other family members. The proportions presented in Table 6 do suggest that boys gave more help than did other family members (36% for boys as compared to 26% for women, 23% for girls, and 20% for men). However, none of the pair comparisons were statistically significant.

In sum, our strongest finding was that advice and help flowed from teenagers to adults. This finding is consistent with observations during home interviews, where we observed parents gaining computer and Internet skills while relying on their children for help.

"and I havenít done [insertion of audio clips] yet because I need Bobby [teenage son] to help me do that.. . .But ok, so now where were we, Bobby? Do I have to unconnect and reconnect? What do I do?" (41 yr old woman, B)

Impact of help on the helper

What were the consequences of being the technical resource for a family on that person. In hypothesis 3 we proposed that making more help desk requests and helping others in the family would lead to greater Internet usage and greater technical skill for the person who did so. We conducted regression analyses to evaluate the personal effects of making help desk requests and helping others in the family. Because, as we have seen earlier, using the help desk and giving help were themselves predicted by participantsí initial levels of skill and Internet usage, these factors had to be controlled when evaluating the consequences of making help desk requests and giving help to others in the family. We controlled initial levels of skill and usage by conducting analyses that, in addition to demographic variables, used pretest skill and usage during the first half of the year to predict posttest skill and usage during the second half of the year (Model 1). Then, in Model 2, we added variables to the equations in Model 1 representing help desk requests and helping other family members.

Results are reported in Table 7. They reveal that making help desk calls in the first six months of the trial was associated with marginally less usage six months later (p < .06), but making help desk calls in the second six months was strongly associated with increased usage in the second six months (p < .001). We believe the slight negative impact of earlier help desk calls on later usage reflects the mix of help desk callers in the first six months. One group, the "disappointed users," needed and called the help desk but gave up and stopped using the Internet early (so their help desk calls predict lower usage six months later). The other group, the "enthusiastic users" also called the help desk, kept calling, and kept using the Internet. Hence, the second six months of usage reflects participants who were more involved, more committed, and who often called the help desk to help them do new tasks on the computer not just to make repairs or fix bugs. Giving help to other household members did not predict either increases in Internet usage or technical skills.

Table 7. Impact of making help desk requests and of helping others in the family on participantsí subsequent Internet usage and skill, controlling for earlier usage and skill.

Several other significant trends in the data shown in Table 7 are worth noting. First, household income predicted increases in usage over the latter part of the year. Perhaps richer families could better afford upgrades to their machines and software that would support more use. Second, teens and minority family members show highly significant gains relative to adults and whites on the posttest computer skill measure. This result reflects an increasing skill gap favoring teens over adults in the household, and a narrowing difference between whites and minorities, who began the field trial with slightly lower skill scores.

The statistical analyses do not show the process by which changes in usage or skill happened, nor why calling for external support encouraged Internet use. The home interviews suggest that help desk staff gave useful information to involved participants that went well beyond scheduling repairs and fixing simple errors. Some participants developed an ongoing relationship with help desk staff they knew by name, and whom they relied on for tips and "know how". Here is an example from an interview with a 19 year old:

Q. "So you became competent in HTML?"

A. "Yeah, Iíve gone in and done that because . . . You know [first name of technical support staff]? He said it was real easy, you can go. . .I forget where itís at, find the source or whatever and copy from somebody elseís [code], or just repeat the steps. And thatís basically all I did (C)."

Impact of help on others in the family

In hypothesis 3 we predicted that the more a person made external help requests and gave technical help within the family, the higher would become the skills and Internet usage of other members of the family. However, because family membersí initial skills and early usage would have influenced how much others in the household offered help to them and how much they would seek help from external sources, these initial values had to be controlled statistically. Our analysis examines how one personís getting external help and giving help within the family was associated with other family members' later Internet usage and posttest computer skill. By controlling early usage and initial skill, the analysis is equivalent to an analysis of change in the outcomes, controlling for regression toward the mean and other artifacts (Cohen & Cohen, 1983). The participant in the household is the unit of analysis, independent variables are attributes of that participant (e.g., the participantís gender, age, or household income), and the dependent variables are the average weekly Internet usage in weeks 27-52 and the average skill from the posttest questionnaire of other household members excluding the participant. Table 8 shows no evidence that an individualís asking questions of the help desk or giving help to more members of his or her household increased the average usage or skill of the other household members.

Table 8.  Impact of a participantís making help desk requests and helping others in the family on other family membersí subsequent Internet usage and skill.

The statistical analysis, which focuses on how one individual's help behavior influences the changes in overall computer skill and use by other family members, doesn't reveal the whole picture. Interviews suggest that the expertise developed by family gurus had contradictory consequences for other family members. According to the stories told by respondents, the expertise developed by the household guru seemed to have simultaneously encouraged and discouraged Internet and computer use and skill development by others in the family. The gurus' expertise kept the hardware and software running and helped other users solve particular problems. At the same time, though, in the process of developing and exercising their expertise, the gurus monopolized the machines and personalized them, making them less attractive to other family members. The following excerpts from interviews illustrate these points.

Taking charge. One way family gurus were helpful to others in the family was by taking responsibility for the introduction of the computer into the household, setting up the machine, organizing family files, and establishing routines for gaining access and using the Internet.

Q. "As I understand it, you have only one bookmark file for the family?"

A. "No, we have separate ones. My brother is, like, director of the house. . .Iím second in command." (H)

Gurus also mediated external technical support for others and encouraged and gave advice to others in the family. They championed use of the Internet and the acquisition of new technical skills.

"My niece is always trying to teach me how to use it. . .You know so thereís a lot of things I donít understand, or she shows me one time and I forget by the next time I have to use it. We tend to spend the weekends when sheís home and stuff talking about what sheís found, what Iíve found. [I ask] how you do this, and she shows me. Of course when she shows me, her fingers fly so fast I canít see it, I have to make her slow down, but I think itís brought [us] closer together." (52 yr old female, W)

Taking charge, however, could have some negative effects on others. Many gurus, by virtue of their being the most involved users of the Internet in their households, lobbied to have the computer installed in their room, and monopolized the machine, making it inaccessible to others at convenient hours.

". . [my husband] is really quite obsessive about it and . . he checks, he logs on at 6:00 AM and itís the last thing he does at night. I get a crack at it periodically. . " (50 yr old female, F)

For many of the household experts, especially the teenagers, taking charge of the computer meant customizing the machine in ways that made it less attractive to others. Teenagers installed flashy screen savers and downloaded multiple programs from the Internet, and they created unusual user interaction preferences that claimed implicit ownership of the machine and confused other family members. Some downloaded programs that emitted unpleasant noises when anyone used the computer. One teen, for example, changed the error sound, so that the machine spoke an expletive whenever someone made a mistake. Some even disassembled machines.

Q. "Then you were using [the computer] too?"

A. "Yeah. Limited. Because it was in my sonís bedroom and so he got to do everything, including taking it apart" (47 yr old male, H)

Good and bad teaching. Teen gurus were both good and bad teachers. They were often close by when others needed their help, a great convenience.

". . .I can never get it back and thatís one of the things that Iíd have to ask Carla how to do it and sheíd say ĎIíve told you this how many times ?í and sheíd have to tell me again." (48 yr old female, referring to her daughter, age 18)

But teens could be thoughtless.

"It seems that every time I have my [files] on here, I donít know what happens to them. I donít know if you can erase them and thatís what my brother does to me, but like I had all my college ones on here, and I think he just erased most of them." (18 yr old female, A)

And unsympathetic.

"and my Dad just doesnít know. Itís tough to explain it to him because heís not used to it at all. Totally different generation" (18 yr old female, H)

Authority structure. The interdependencies between teenage gurus and others in the family had unexpected consequences in some families. Other family members found themselves living with a teen who had become passionate about the computer and Internet and whom others considered technically far ahead. They felt fortunate that the teen was there to help family members when they had problems, to fix "frozen" computers, to make suggestions, and to interpret what could seem an alien world of computing (Sproull, Kiesler, & Zubrow, 1984). However, many parents had initially joined the trial because they believed it was "good for the kids." Now, the teen was becoming part of an alien world, and was adopting jargon, expertise, and interests family members and friends did not always share. The social differences this change generated are reflected in the following interview with a teenage girl (H).

Q. ". . .So you think that using the Internet makes you stand out-- makes you different from your friends?

A. "A little bit. It makes me more knowledgeable when it comes to certain things."

Q. "But they donít mind and you donít mind?"

A. "I mind teaching them everything. Like they donít know how to put in their names or whatever, plug in those little blocks. You have to click on them first. They donít know that. They just start typing and theyíre like, Ďwell, what happened, my name didnít come up?í You have to type it, you have to click on it. Itís like Iím on a certain page in my book and theyíre on page one and I canít deal with that."

Within families, teensí technical expertise shifted intellectual authority in the family, as the same teen discusses:

Q. "Is it like that with your father? If he were trying to use something and he couldnít figure it out and he asked you. . .?"

A. "Sometimes if Iím not doing anything, Iím just like washing dishes or something, oh, he canít access something, I can help him. Sometimes [he says] ĎI know what Iím doingí (she lifts her eyebrows, indicating skepticism). I donít know, maybe he gets upset that I know more about this than he does.

Q. "Does he ask you very often?"

A. "Not very often. . . Heíll stumble across a few things now and then that he never knew they were there, and Iíll show him."

Q. "Is that fun for you?"

A. "A little bit. Gives me the upper hand."

The potential shift in authority was one explanation that some fathers gave for resisting getting help from their children, especially their daughters. This teenís experience with the Internet caused her to gain technical proficiency that her parents and siblings valued and admired, but her skill also created differences in outlook and interests to which everyone had to adjust. In many cases, however, parentsí belief in the value of computers and computer skills overcame their reservations about these gaps, and if they took steps, it was often in the direction of trying to obtain more experience themselves, but often by themselves or with friends their own age.

Discussion

Unlike most consumer products and services, computers require technical competence and know-how. Even todayís user-friendly home computers are far more difficult to use than the average home appliance. When the washing machine stops working, it is easy to diagnose that the machine is at fault rather than the homeís wiring, the soap, the clothing, or the userís behavior. When the computer stops working, diagnosis is much more difficult. We began this analysis to find out why people did or did not obtain available external technical support when they faced difficulties using their computer or the Internet. We wondered why those calling seemed to be the very people who knew the most, and why some people dropped out rather than call for help. We discovered that calling for external technical support is rooted in commitment to using the computer and the Internet, and that requesting support is part of a behavioral pattern characteristic of involved Internet users. These same involved users became a major source of technical support for other members of their families.

On the pretest we gave before anyone began the trial, those with the most computer skill, confidence or enthusiasm ultimately became the most involved users of the Internet. This involvement led to their making requests for external support to solve problems and to accomplish new tasks, which in turn led to more Internet usage. Highly involved Internet users often became family gurus that others in the family relied on to mediate external help and to provide direct help and encouragement. Their help, however, had mixed results, leading to no net gains that could be detected statistically. Many of the family gurus were teens; in 31 out of 58 (54%) of families with teens, the top help desk requester was a teenager. Teens in the role of technical gatekeeper could be especially uncomfortable to adults in the family who were unused to this experience. Typically, this same teenager was not the family member who called the electrician, fixed leaks, or repaired the dishwasher.

The flow of technical information within households we observed in this field trial shared some similarity with information flow in organizations. As in organizations, the gate keeping role tended to be specialized, with much of the technical information coming into the group through a small number of points of contact, who then redistributed this knowledge. As in organizations, the gate keepers tended to be the most technically competent in the group. However, the gate keeping structure in families did not conform the structure of information distribution that has been attributed to large work organizations with a flow of technical information from more to less senior employees. Instead, computer expertise within families flowed informally from teens to adults, much as from graduate students to faculty, or from newly trained computer professionals to older programmers.

When we considered both gender and generation in our analyses, we found some interesting complications. Men helped their wives, but hardly anyone else in the family. The help giving between mothers and daughters tended to be reciprocal, but that between sons and their fathers was not. There was some tendency for boys to serve as gurus more than girls, as would be expected based on their greater Internet usage, but the differences were not significant. Many of the teen girls in this field trial considered themselves every bit as "nerdy" and computer-skilled as boys.

Taken as a group, the teen gurus in the families were a fascinating group. In the families with teenagers, teens were usually the family members who lobbied for a PC and Internet access, and it was mainly for their teensí benefit that parents welcomed the computer (see Kraut et al., 1996). Teens, especially boys, began with the most technical confidence and interests, and dominated usage by a clear margin (Kraut et al., 1996). Whereas many of the adults had little time for recreational computing, teens had the time to explore the Internet and develop new computer skills. They introduced new services to the household, downloaded programs, and served as the familyís technical resource. Our statistical analyses show how being a teen was associated with these technical attributes and with the role of family guru. Our home interviews show how teen gurus introduced a new, sometimes abrupt, dynamic into families unused to the teenís role at the interface of the family and the new world of the Internet. Teen gurus were admired for their abilities, and sometimes they were held in awe. Most of the parents were proud of their teensí computer accomplishments, and were far more willing to tolerate hours spent before a computer screen than before a television screen.

Applications to support services

Our analysis has some implications for policymakers and designers, and for those who provide Internet and computer support services. Our data suggest, first, that current estimates of the demand for technical support assembled from professional support services underestimate the amount of technical support that people actually need. These estimates do not include informal help provided by friends and family, and the help that people never seek out or get. Furthermore, if our data can be generalized, the calls to professional services probably represent a more sophisticated set of problems than most home users have. The people who called the HomeNet helpdesk were the most technically involved in the household and many of their questions reflected more technical sophistication than the average family member had.

Simply providing people with access to the Internet does not assure that everyone can or will use this access. A modicum of computer expertise and confidence still seem critical to using the Internet, and to taking advantage of it. The problems our participants experienced suggests that the industry needs to put more resources into assessing and improving usability, based on the experiences of the range of actual users. Given the centrality of teens in using the Internet and providing informal technical support, one strategy might be to train high school students to give technical help more effectively within their families and communities. In this vein, an interesting project group in Austin, Texas gives low income teens instruction and parts to build a computer. After the teen builds a computer for a nonprofit agency, he or she gets to build one for the teenís own household. Similar projects with incentives and instruction to help others should benefit teens as well as others. And such programs would benefit of the computer industry, because there of the strong links between skill, involvement, and helping others.

Our analysis has several limitations that could be addressed in future research. First, although in the home interviews we asked family members to "walk through" their use of the computer and the Internet, we were rarely able to observe the process of help seeking and help giving. Better measurement of family interactions would allow us to better understand the kinds of complaints, dissatisfactions, and assistance needs that led to different problem solving strategies, including requests for family or external support. Second, we did not document the results of all help desk calls to know whether these calls led to solutions or further problems. For instance, we do not know if, or how often, repeat calls occurred because participantsí problems were not resolved by the first request. More frequent requests for external support (exerted by those with more technical involvement) might be a consequence of higher standards for service, the pursuit of an unsolved problem, or both.

Implications for research on technological change

The last decade has seen a huge increase in the number of U.S. households using computers and the Internet. Estimates of households online as of 1999 range from one/fourth to one/third of total households. Print, radio, telephone services, and television were previous technologies that brought the outside world into the household. The Internet continues and extends this phenomenon, giving individual family members easier ways to communicate and pursue their interests beyond the confines of home, neighborhood, and organization. Our study and industry estimates show that teenagers are leaders in taking advantage of these technological opportunities. Teens are among the first to adopt new, computer-based ways to communicate, to learn, and to be entertained, and the pace of technological change suggests that they will be in the forefront of change for some time to come.

Our study suggests that teens are helping families adjust to technological change and, at the same time, they are carriers of social change. Joshua Meyrowitz (1984), in his analysis of the social impact of television, argues that television created a new household information system. Television, he says, helped to blur the boundaries between childhood and adulthood by exposing children to otherwise difficult to obtain information about the world, adult behavior, and social possibilities. The Internet is our own timeís most important new household information system, and it may be contributing to a further blurring of childhood and adulthood. It is not just that teens on the Internet have a vast choice of Internet content and interactions. With the advent of the teenage guru, the child in the family plays a new role of child-as-technical-advisor, a role (given the high status of technical expertise in the U.S.) that confers on the teen authority and probably independence as well. With their teen guru, we observed that parents were likely to give in when there was contention over the use of the computer. Many were happy with the skills that their teens gained with nights and weekends spent at the screen, as long as the content was appropriate.

In future research, the social consequences of increasing technical expertise among teens should be studied further. For example, how does the acquisition of technical expertise differentiate a person from her or his peers and from other family members? If there is a competency multiplier effect, the more technical person might grow further apart intellectually from her or his peers and family even as she or he helps them (Patterson, 1999). In an earlier report from the HomeNet study (Kraut et al., 1998), we showed that in the first year of use, those who used the Internet more, also engaged in slightly less family communication overall. Whether this was due to their time alone on the Internet, to a divergence of interests in the family, to the cognitive effects of expertise, or to something about technical skill per se, is an important question for understanding the social impact of our increasingly technological society

 

Endnotes

References

Allen, T. (1977). Managing the flow of technology. Cambridge, MA: MIT Press.

Barley, S. R. (1988). The social construction of a machine: Ritual, superstition, magical thinking and other pragmatic responses to running a CT scanner (pp. 497-539). In M. Lock & D. R. Gordon (eds.), Biomedicine examined, NY: Kluwer Academic Publishers.

Berk, S.F. (1985). The gender factory: The appointment of work in American households. New York: Plenum.

Berselli, B., (1997) When ĎHelp gives way to ĎHELP!í Top-level techies pick up the line. Washington Post, A1, A7.

Bryant, W.K., & Zick, C.D. (1996). An examination of parent-child shared time. Journal of Marriage and the Family, 58, 227-237.

Cohen, W. & Levinthal, D. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 128-152.

Crouter, A. C., Manke, B.A., & McHale, S.M. (1995). The family context of gender intensification in early adolescence. Child Development, 66, 317-329.

Franzke, M. & McClard, A. (1996). Winona gets wired: technical difficulties in the home. Communications of the ACM, 39(12), 64-66..

Hoffman, C. (1999). Technical support knowledge systems: An overview of the problem resolution industry. (Webcast presentation). International Data Corporation, 5 Spleen St., Framingham, MA 01701.

Hoyman, M. (1987). Female participation in the informal economy: A neglected issue. Annals of the American Academy of Political and Social Science, Sept., 64-82.

Kraut, R, Mukhopadhyay, T., Szczypula, J., Kiesler, S. (in press). Information and

Communication: Alternative Uses of the Internet in Households, Information System

Research.

Kraut, R., Patterson, M., Lundmark, V., Kiesler, S., Mukhopadhyay, T., & Scherlis, W. (1998). Internet paradox: A social technology that reduces social involvement and psychological well-being. American Psychologist, 53(9):1017-1031.

Kraut, R., Scherlis, W., Mukhopadhyay, T., Manning, J., & Kiesler, S. (1996). The HomeNet field trial of residential Internet services. Communications of the ACM, 39, 55-65.

Lyons, W. E., & Lowery, D. (1989). Citizen responses to dissatisfaction in urban communities: A partial test of a general model. Journal of Politics, 51, 841-868.

Marascuilo, L.A. & Serlin, R.C. (1988). Statistical methods for the social and behavioral sciences. NY: W.H. Freeman and Company.

Meyrowitz, J. (1985). No sense of place: The impact of electronic media on social behavior. NY: Oxford University Press.

Netopia, Inc. (1995). Timbuktu User's Manual. Alameda, California: Author.

Neuman, Russell (1986). The paradox of mass politics: knowledge and opinion in the American electorate. Cambridge, MA: Harvard Press.

Orlikowski, W. J. (1996). Improvising organizational transformation over time: A situated change perspective. Information Systems Research, 7, 63-92.

Patterson, M. (1999). Social distancing effects of interest in computers and other expert avocations. Unpublished ms. Carnegie Mellon University.

Punj, Girish & Staelin, Richard (1983). A model of consumer information search behavior for new automobiles. Journal of Consumer Research, 9, 366-380.

Sproull, L., Kiesler, S., & Zubrow, D. (1984). Encountering an alien culture. Journal of Social Issues, 40, 31-48.

Constant, D., Sproull, L., & Kiesler, S. (1996). The kindness of strangers: On the usefulness of weak ties for technical advice. Organization Science, 7, 119-135.

Tushman, M. L. (1977) Special boundary roles in the innovation process. Administrative Science Quarterly, 22(4), 587-605

US Census Bureau (1998). Household and Family Characteristics: March 1998 (Update) (P20-515)Summary Tables. http://www.census.gov/population/www/socdemo/hh-fam-sum98tab.html

White, L,K., & Brinkerhoff, D.B. (1981). The sexual division of labor: Evidence from childhood. Social Forces, 60, 170-181.