Journal of Interactive Advertising, Volume 3, Number 2, Spring 2003
This study introduces Internet Dependency Relations (IDR) as a predictor of online consumer activities. IDR is based on the theoretical perspective of Media System Dependency theory, which postulates dependency relations between individuals and media based on the perceived helpfulness of media in meeting understanding (social/self), orientation (action/interaction) and play (social/solitary) goals. Using a cross-sectional email survey of 166 respondents randomly drawn from the faculty, staff, and student population at a large mid-western university in the United States, the predictive influence of IDR on online shopping, chatting, and news reading was empirically tested. On average, consumers in the survey had bought eight products online in the last six months, spent twenty-one minutes daily reading news online, and chatted ten minutes daily on the Internet. They also displayed moderate, though positive dependency relations with the Internet. IDR significantly explained online shopping activities and online news reading, but did not predict online chatting. In terms of specific IDR goal dimensions, the predictive influence of action orientation on online shopping, solitary play on online chatting, and social understanding on online news reading was confirmed.
Whether dealing with the consumption of goods, news, or other types of online content, it has been suggested that consumer activities in online environments indicate a more instrumental than ritualistic use of media. Even more so than any other medium, the Internet anticipates an active rather than passive audience, implying that, at the present time, its use is more purposive and goal-directed. Therefore, it is possible that the personal and social goals that people seek to meet through the Internet may be important motivating factors in the activities that they pursue online. In this study we attempt to tie goal-directed motivations of Internet users with online shopping, chatting, and news reading.
Shopping, chatting, and news reading are fast proliferating activities among U.S. users in today’s online environment. In March 2001 alone, more than 100 million U.S. consumers shopped online, collectively spending over $3.5 billion (Nielsen/NetRatings and Harris Interactive 2001). Similarly, thousands of chat rooms of every nature report hosting over a million chatters daily (Palm Coast/Flagler Internet 2000), testifying to the growing popularity of instant messaging and related chat forms (Pastore 2001). And recent research from Pew Internet & American Life (2000) rated online news reading as the third most popular daily Internet activity in the United States, after sending email and surfing the Web for fun.
Two theoretical approaches available to study how individual goals are met through media (including Internet) resources are Uses and Gratifications and Media System Dependency theory. Unlike Uses and Gratifications research, which is premised on consumer control over accessing media content according to their goals/needs, we focus on consumer dependency on Internet resources to satisfy goals. We believe that such a dependency on the Internet leads, over time, to the development of a consumer-Internet dependency relationship, which, in turn, may likely affect the nature and extent of consumers’ online activities.
In this study, we propose — and test — the multidimensional construct of Internet Dependency Relations (IDR) as a possible predictor of online activities. Conceptualized as the extent to which people depend on the Internet to meet their social and personal goals, IDR is derived from Media System Dependency theory (Ball-Rokeach 1985, 1998; Ball-Rokeach and DeFleur 1976; DeFleur and Ball-Rokeach 1982, 1989), which defines individual-media relations in terms of both overall intensity of the dependency relationship, as well as the extent to which individuals relate to a medium to meet specific goals. MSD goal dimensions include understanding (self and social), orientation (action and interaction), and play (solitary and social) goals that individuals seek to meet through media resources (Figure 1).
Figure 1Goal Dimensions of Media System Dependency Relations
The central issue this paper addresses is the extent to which Internet behaviors can be explained by IDR both as a summed intensity, and as the intensity of six specific goal dimensions. We argue that overall IDR intensity will significantly influence Internet users’ online shopping, news reading, and chatting experiences. These activities were selected both due to their growing popularity among Internet users, and their intuitive corresponding match with MSD goal dimensions: shopping with orientation, news reading with understanding, and chatting with play. We further hypothesize connections between the intensity of specific goal dimensions and specific online activities, and examine predictive linkages between action orientation and online shopping, social understanding and online news reading, and solitary play and online chatting.
According to MSD theory, a media dependency relationship is one "in which the satisfaction of needs or the attainment of goals by individuals is contingent upon the resources of the other party” (Ball-Rokeach and DeFleur 1976, p. 6). MSD suggests that in today’s society individuals have to rely on media information resources in order to attain their various goals. Information resources include all media products (Loges and Ball-Rokeach 1993), including commercial and advertising information. The intensity of media dependency relations depends on the perceived helpfulness of the media in meeting goals. The goal scope (dimensions) of these relations (Figure 1) covers a wide range of individual goals — understanding (social and self), orientation (interaction and action) and play (social and solitary) — that may be met through media resources (Loges 1994). Understanding goals deal with people’s needs to understand the world and themselves; orientation goals focus on the need to behave effectively in interactions with others as well as in personal behavioral decisions; and play goals deal with the need for entertainment and escapism (Morton and Duck 2000). While these goal dimensions are exhaustive, they are not mutually exclusive — and more than one kind of goal can be activated (and satisfied) by the same medium (DeFleur and Ball-Rokeach 1989). Both intensity and goal scope may be determined by how exclusive media resources are perceived to be in attaining these goals, and vary for different individuals as well as for the same individual over time (Ball-Rokeach 1985, 1998; Ball-Rokeach and DeFleur 1976; DeFleur and Ball-Rokeach 1982, 1989).
The incorporation of the Internet into daily lives is reflected in the kinds of activities many Americans pursue online. On a typical day in March 2000, 58 million Americans logged on to the Internet (Pew Internet & American Life 2000) to send email, surf for fun, get news, buy a product, or chat in a chat room or a discussion forum, among other things. Internet users surveyed in a recent study said the Internet had improved their connection to family and friends, the way they pursue hobbies, and their ability to learn new things. Many found the Internet helpful in doing jobs, getting information on health care, shopping and managing personal finances (Howard, Rainie, and Jones 2001). The diversity and intensity of online activities point to the need to investigate what factors might intervene in the activities. Recent models of media selection and use have suggested that, in addition to demographics and media attributes, factors such as assessment of needs fulfillment, appropriateness, social norms, and peer evaluations are important in determining the nature of media use (Flanagin and Metzger 2001). Therefore, from an MSD perspective one might argue that individual goals –and the Internet’s ability to meet them — may exert some influence on consumer activities in the online environment.
A few years ago, shopping or “purchase” was rated among the least prolific uses of the Web (Katz and Aspden 1997; Poindexter 1999). However, most marketers believe it is only a matter of time before the majority of consumers shop in their virtual storefronts. A March 2001 survey of U.S. users found that e-commerce has hit mainstream, with 48.2% of all Americans over 18 years old –100.2 million people — having bought products online. Despite downturns in the dotcom boom, consumers’ online spending has steadily increased. In March 2001 alone, more than $3.5 billion was spent online, a jump of 35.6% from $2.6 billion in April 2000 (Nielsen Netratings/Harris Interactive 2001). A Yahoo!/ACNielsen Internet Confidence Index report found that US consumers planned to spend at least $10 billion online between July-September, 2001 (Cyberatlas 2001).
Even so, generally speaking, buying online still does not appear to be one of the primary reasons why people visit web sites, despite the overall increase in commercial activities on the Internet. Poindexter’s 1999 study found this to be true of both Baby Boomers and Generation Xers, even though youngsters spent more than 10% of their disposable income on purchasing diverse products through the Web (Forrester Research Report 2000). While GVU’s 10th WWW User Survey (1998) found that quality information, easy ordering, and reliability were more important to respondents than security, Korgaonkar and Wolin (1999) found that, among other things, security concerns and transaction anxiety appeared to be the most prevalent causes for not buying on the Web.
Studies using demographic variables to explain online shopping behavior have often reported conflicting or confusing results. While Li, Kuo, and Russell (1999) found age and education level played an important role in online shopping, as did consumers’ shopping orientation, Bellman, Lohse, and Johnson (1999) considered demographics an imperfect surrogate to explain online purchasing. They found that while demographics explained why people were online in the first place when compared to the national U.S. population, they did not significantly predict online purchase behavior. Donthu (1999) observed that distinction was often not made between online users and online shoppers. His study found online shoppers to be older, more affluent, with a positive attitude towards advertising and direct marketing, less price and brand conscious and largely convenience seekers. A Forrester Research Report (1999) suggested otherwise: younger consumers (40%) bought more frequently on the Internet as compared to more mature adults (30%); and fully 62% of all young U.S. consumers were likely to shop online by 2003. A variety of studies have also pointed out the increasing online shopping sophistication of today’s 16-22 year olds, as evidenced by their use of various aids such as price comparison web sites and online coupons to buy a wide variety of products on the Internet.
In terms of motivational variables found to influence online shopping, the Wharton Virtual Test Market results reported Wired Lifestyle–characterized by years’ experience with the Internet, reception of large amounts of emails and work on the Internet in the office every week, and Time Starvation–a result of the increasing number of hours worked by members of a household especially in dual-income households, as predictive of online shopping (Bellman, Lohse, and Johnson 1999).
While Media Dependency Relations has been previously used in purchase contexts (Grant, Guthrie, and Ball-Rokeach 1991; Skumanich and Kintsfather 1998), it has been studied only in television shopping environments. Grant, Guthrie, and Ball-Rokeach (1991) modeled relationships between viewers/buyers, the television shopping program, and the television medium by extending MSD theory to dependency on the television shopping genre. Skumanich and Kintsfather’s (1998) study found viewer relationship with the medium, the genre and the genre personae (i.e. the tele-shopping host) highly predictive of purchase behavior.
Our research examines the relationship between Internet Dependency Relations and online shopping. Since it is evident that the online shopping experience involves a range of diverse activities like conducting product information searches, price and brand comparisons, searching for discounts, as well as actual online product purchase, this study conceptualizes online shopping as both a range of activities, as well as the actual number of products bought online.
In view of the mixed findings related to the use of demographics as online shopping predictors, we include demographic variables. Specifically, we ask the following research questions:
RQ1: To what extent do intensity of Internet Dependency Relations, age, gender, and income influence consumers’ online shopping activities?
RQ2: To what extent do intensity of Internet Dependency Relations, age, gender, and income influence the number of products bought online?
Based on definitions of goals dimensions provided in MSD theory, we further hypothesize a connection between individuals’ action oriented goals and online shopping:
H1: Stronger action orientation goal dimension will positively predict consumers’ overall online shopping activities.
H2: Stronger action orientation goal dimension will also positively predict actual online product purchase.
As an easy way to get instant answers to messages and to carry on conversations with friends, colleagues, and strangers around the globe, online chatting is one of the fastest growing activities on the Internet (Pastore 1999). Almost every portal or online community on the Internet today hosts some type of chat activity. The chat statistics released by America Online indicated that there were more than 40 million registered users of its Buddy List and Instant Messenger services, and more than 750 million daily messages were sent through Buddy List and ICQ services. A NetValue survey in 2001 found that online chatters as a group were among the heaviest users of the Internet; they generated twice as many online sessions as non-chatters (Pastore 2001). The study found that women were more likely to chat online than men, and spent two more days per month on online chatting than did males (Pastore 2001).
The paucity of research on this fast proliferating online consumer activity, and industry-reported evidence of its growing popularity among Internet users, leads us to investigate the connection between Internet Dependency Relations and time spent chatting on the Internet. Considering that demographics provided some understanding of online chatting, we ask:
RQ3: To what extent do intensity of Internet Dependency Relations, age, gender, and income influence time spent chatting online?
Again, drawing upon the goal scope of MSD theory, we propose the following predictive relationship:
H3: Stronger solitary play goal dimension will positively predict time spent
Practically every major mainstream newspaper or magazine in the United States is available in an online edition; the same holds true for broadcast news networks. In addition, most Internet portals themselves incorporate online news services. Millions of other official and unofficial news-based sites are online, providing collective information resources that appear to be virtually limitless.
Market Facts/MSNBC reported in 1998 that 20.1 million U.S. residents used the Internet as a source for news (Levins 1998). A Pew Research Center biennial news consumption survey revealed that there was a jump in online news activities between 1996-1998, from 6% of Americans to 20% searching for news at least once a week. For these users, science, health, finance and technology were big news draws (Pew Research Center for People and Press 1998). More recently, Scarborough Research’s first National Internet Study, surveying more than 2000 U.S. adult Internet users, found that more than two out of five Internet users (45%) had read an online newspaper in the last 30 days. Half (55%) had logged on to a national newspaper web site like the New York Times, Wall Street Journal, and USA TODAY (Scarborough Research 2001). Scarborough Research (2001) also indicated that, generally, online news readers tend to be younger (41% were between the ages of 18-34) as compared to traditional newspaper readers (only 23% in the same age category). Hence, it appears that online editions of newspapers intentionally or unintentionally target a new younger audience. Even though it is too early to claim that online news has entered the mainstream, Noack (1999) argued that the obvious advantages of Internet-based news (accessibility, convenience, in-depth research, and information) would be key in attracting readers, encouraging them to spend more time reading news online.
Along with increased usage of online news, research has also found an increasingly positive attitude toward news as well. For instance, among 550 Internet users polled by ScreamingMedia, more than half believed that the Internet had the most interesting information and provided in-depth, accurate, up-to-date information (Astor 2000). Examining electronic newspaper usage, Weir (1999) concluded that media consumption was purposeful and adopters of electronic newspapers used them to get information important to them.
Prior research exploring the connections between media dependency relations and news in a print newspaper context indicated that intensity of dependency relations added a significant amount of explanation in newspaper reading variance when demographic variables were controlled; social understanding, self understanding and action orientation were important dimensions of newspaper dependency relations (Loges and Ball-Rokeach 1993). In another study on television media dependency relations, a linkage between dependency and news was also investigated and confirmed (Ball-Rokeach, Rokeach, and Grube 1984).
No prior studies have examined the connections between online news behavior and media dependency relations. Following Loges and Ball-Rokeach’s (1993) suggestion to consider both media dependency relations as well as demographic factors in analyzing media use, we ask the following question:
RQ4: To what extent do intensity of Internet Dependency Relations, age, gender, and income influence time spend reading news online?
Based on MSD’s identification of social understanding goals as leading to individuals’ information seeking behavior, we also hypothesize its predictive relationship with online news reading.
H4: Stronger social understanding goal dimension will positively predict
time spent reading news online.
To measure intensity of Internet Dependency Relations (IDR), the 18 item MSD scale developed and refined by Ball-Rokeach, Rokeach, and Grube (1984), Grant, Guthrie, and Ball-Rokeach (1991), and Ball-Rokeach, Grant, and Horvath (1995) is used. IDR is thus operationalized as respondents’ composite mean score on the 18 item MSD scale (See Appendix). To measure action orientation, solitary play and social understanding, composite mean scores for three items each on the MSD scale were used. Hence, mean scores for each of the three goal dimensions used in this study are subsets of the overall dependency mean score.
The dependent variable online shopping was conceptualized both as a range of shopping-related activities that consumers engaged in online, as well as the number of actual products they bought online. Operationally, a five-item interval scale developed by Patwardhan (2001) was used to measure respondents’ self-reported online shopping activities, while actual purchase was measured at the ratio level as the number of products bought online in the last six months.
Online news reading was conceptualized as the extent to which respondents accessed and read news on any news-based web site, including web sites of newspapers, broadcast media, or other news-based organizations. It is operationalized as a ratio level measure of the amount of time spent reading news online every day.
Online chatting was conceptualized as respondents’ use of chat and instant messaging services available on the Internet, and measured at the ratio level as the average amount of time spent daily chatting with others via the Internet.
Demographic variables were measured as follows: data on age were collected by asking for the year of birth; income was measured categorically (Please select the appropriate range to reflect your annual income from all sources including salary, and/or parental support if students. Unemployed students were advised to select their parents’ income range); Gender was the only dichotomous (nominal) variable.
The study used a cross-sectional email survey. The population of interest was students, faculty, and administrative staff at a large mid-western university in the United States. The sampling frame was the university email directory. This sample was particularly desirable for this theory testing study because university communities are known to have a high proportion of Internet users. Future research will build on the foundations of this study, surveying national/international Internet user populations.
Respondents were selected using multi-stage stratified random sampling. Stratified random sampling is a superior method to simple random sampling and ensures representativeness in terms of variables important to the study. The university population was first stratified into three groups — students, faculty, and staff — to ensure age and income variability. Each population group was further stratified on the basis of gender, to ensure gender representation. Five percent of students, (since students were a much larger group than the others), and 10% of faculty and 10% of staff were sampled. Thus the total sample size selected for the study was 1,462 respondents (1,200 students 127 faculty and 82 staff). While exploring differences by group was not the primary purpose of this study, this sampling method ensured variability on demographic factors (age, income, and gender) important to this study.
Reliability analysis was conducted for IDR and online shopping scales. Multiple Regression was used to answer RQ1 through RQ3. Bivariate regression was used to test Hypotheses 1 through 4.
After creating an email address list for respondents in the selected sample, the survey questionnaire was delivered via email. Conforming to research “netiquette,” the survey was accompanied by a cover letter explaining purpose and nature, time required to complete the survey, and the researchers’ academic affiliation. The option to opt out was offered, as were confidentiality assurances. The letter and questionnaire were sent out as inline text. Respondents were requested to hit the reply button, respond to questions, and “send” the survey back to the researchers. Two mailings were done, with some textual adjustments made for the second mailing to overcome problems in administration. For example, a respondent reported that the message was truncated when the reply button was hit, so the revised cover letter offered suggestions for other return routes like cutting and pasting the survey into the reply, or sending via campus mail.
Reliability and Validity
Post-test reliabilities for the scales were tested using Cronbach’s alpha. While scale reliability for the overall 18-item IDR scale was fairly high (.88), it was somewhat lower for each of the six individual goal dimensions – self understanding (.74), social understanding (.67), action orientation (.64), interaction orientation (.59), self play (.85), and social play (.65). This was probably because only three items were used to measure each dimension (larger number of items generally increases reliability). For the dependent variable online shopping, reliability for the five-item scale was .90, similar to pre- and post-test reliabilities reported for this scale (.91) in previous research (Patwardhan 2001).
Despite an inability to generalize the results of this study beyond the population of interest (university faculty, students, and staff), it may be argued that university populations are likely to reflect many of the characteristics of Internet users in the United States. The study does have high external validity in terms of generalizing from the sample to the university population, since probability sampling was used.
A total of 1,462 questionnaires were emailed over a 10-day period. Four hundred and eight emails were returned as failed deliveries; and there were twelve refusals to participate. [The high number of failed deliveries were mostly from the student group in the sample, suggesting that student email addresses with the university are not necessarily current]. Hence for a total of 1,001 emails successfully delivered, 176 responses were received, a response rate of 17.6%. Subsequently, ten incomplete (truncated) replies were discarded, leaving 166 usable sample questionnaires.
We acknowledge that low response rate is a major limitation, and offer two possible defenses. First, many online and email surveys (including the GVU Surveys) use non-random sampling methods since it is difficult to obtain a sampling frame of all Internet users in a particular population. A review of the literature also suggests that response rate for email and online surveys is generally much lower than mail or telephone surveys. Our response rate falls well within the range reported by researchers using probability sampling in email surveys. However, like other researchers conducting surveys using these methods, we would caution against generalized interpretations of our results.
As a precaution against sampling error, we conducted a check for non-response bias, and found a good match. We also compared the sample demographic profile with the population profile obtained from university sources. There was a good match on age for all three groups (faculty, students, and staff). In terms of gender, a moderate skew toward female respondents was observed among student and staff respondents but no skew was detected in the faculty group. For the faculty group, higher income was over- represented and lower income was under-represented. However, the overall match between the sample and the population increases confidence in the generalizability of our results to the university population.
In terms of demographics, a little more than half the respondents (55%, n = 92) were students, 24% (n = 40) were faculty, and 21% (n= 34) were administrative staff. Representation of females was slightly higher (56%) than males (44%). Over half the respondents (52%, n = 78) were in the age group of 18-34, 40% (n = 59) were between 35-54 years old, and only eight percent (n = 12) were more than 55 years old. In terms of income distribution, over 61% of respondents (n = 89) had an annual income below $50,000: of these, half (n = 45) earned below $25,000 and half (n = 44) above. Twenty eight percent of the respondents (n = 41) earned between $50,000 to $100,000 annually, and a smaller number (10%, n = 15) earned more than $100,000.
Mean IDR and Online Behavioral Activities
The mean intensity of overall Internet Dependency Relations among respondents suggested a positive but somewhat restrained dependence on the Internet’s resources to satisfy individual goals (mean = 3.1 on a scale of 1 to 5, with higher score indicating greater dependency). Similarly positive but moderately intense dependency for understanding (mean = 3.0), orientation (mean = 3.2) and play (mean = 3.1) were also observed (Table 1). Interestingly, the highest means were found among the three specific goal sub-dimensions used in this study: action orientation (mean = 3.6), solitary play (mean = 3.3) and social understanding (mean = 3.6), when compared to other social and self dimensions in the MSD goal scope.
Table 1Mean Internet Dependency Relations and Online Behavioral Activities
In terms of Internet-based activities, most respondents in the survey engaged in online shopping activities fairly frequently on the Internet (mean = 2.2 on a scale of 1 to 5 running from “very frequently” to “never”), and had bought an average of eight products online in the last six months. While there were no significant differences by group in the use of the Internet for online shopping activities in general, differences were observed in the number of products bought online by staff (15 products in six months) and students (6 products in six months) (F = 4.3, p = .02) (Table 1).
Respondents also spent about 10 minutes daily chatting online. Students and faculty differed significantly in the time they spent chatting online (F = 3.6, p = .03). On average, students spent the most time chatting online daily (mean = 15 minutes), followed by staff (mean = 6 minutes), and faculty (mean = 3 minutes) (Table 1).
On average respondents spent about 21 minutes daily reading news online, and no statistically significant differences by group were observed in time spent reading news online, though faculty spent the most time on this daily activity (mean = 26 minutes), followed by students (mean = 20 minutes) and staff (mean = 18 minutes).
RQ1: IDR and Online Shopping
RQ1 investigated the extent to which intensity of Internet Dependency Relations, and demographic factors, affected consumers’ online shopping activities. Linear multiple regression was used to check the relationships. The overall model found that eight percent of variance in online shopping was explained by IDR and demographic variables (R square = 8.3, F = 2.75, p = .03) (Table 2). However, none of the demographic factors (age, gender, income) were significant predictors. IDR was the only factor that significantly explained almost all the variance in online shopping (8%, t = 3.22, p = .00).
Table 2 Multiple Regression Analysis of Age, Gender, Income and Internet Dependency Relations to Predict Variance in Online Shopping
RQ2: IDR and Number of Products Bought Online
RQ2 examined the extent to which the intensity of Internet Dependency Relations, and demographic factors, affected number of products bought online. Linear multiple regression was used to test the relationships. It was found that the regression model did not significantly predict the number of products bought online.
H1: Action Orientation Goal Dimension and Online Shopping
H1 expressed a relationship between action orientation goal dimension and the online shopping experience. Bivariate linear regression was used to test the hypothesized relationship. Action orientation was found to be a strong, significant predictor (R square = .26, F = 54.05, p = .00) (Table 3). H1 was, therefore, supported.
Table 3 Linear Regression Using Specific IDR Dimensions to Predict Variance in Online News Reading, Online Chatting, Online Shopping and Number of Products Bought Online
H2: Action Orientation Goal Dimension and Number of Products Purchased Online
H2 asked whether action orientation goal dimension also predicted the number of products bought online. Surprisingly, despite moderate significant correlation between online shopping activities as measured on the scale, and the number of products bought online (r = .493, p = .02, one tailed), the predictive relationship between action orientation and the actual number of products bought online was not significant (R square = .02, F = 3.38, p = .07) (Table 3). H2 was, therefore, not supported.
RQ3: IDR and Time Spent Chatting Online
RQ3 investigated the extent to which IDR, and demographic factors, influenced time spent chatting online. The overall multiple regression model explained 18% of variance in online chatting (R Square = .18; F = 6.57, p = .00) (Table 4). However, the variance in time spent chatting online was not explained by IDR, but by demographic variables (age and income). Of the total amount of variance in online chatting, age had a unique contribution of six percent, income explained 11%, and the rest was shared.
Table 4 Multiple Regression Analysis of Age, Gender, Income and Internet Dependency Relations to Predict Variance in Time Spent Chatting Online
H3: Solitary Play Goal Dimension and Time Spent Chatting Online
H3 investigated the predictive relationship between solitary play and time spent chatting online. Bivariate regression analysis found that solitary play significantly predicted the time spent chatting online (R square = .05, F = 7.6, p = .00) (Table 3). Hypothesis 3 was, therefore, supported. However, only five percent of variance in the dependent variable was explained by solitary play, suggesting that the predictive relationship between the two was not very strong.
RQ4: Demographics, IDR and Time Spent Reading News Online
RQ4 examined the extent to which IDR intensity, and demographic variables, affected the time spent reading news online. The linear multiple regression model accounted for 19% of variance in the dependent variable (R square = .19, F = 6.69, p = .00) (Table 5). Interestingly, the two significant predictors were gender (T = -4.35, p = .00) which explained 11% of variance in the dependent variable, and Internet Dependency Relations (T = 2.85, p = .00) which explained five percent of variance. The negative relationship in the case of gender suggested that male Internet users were more likely to read news online than female Internet users.
Table 5 Multiple Regression Analysis of Age, Gender, Income and Internet Dependency Relations to Predict Variance in Time Spent on Online News Reading
H4: Social Understanding Goal Dimension and Time Spent Reading News Online
H4 hypothesized a predictive relationship between social understanding and time spent reading news online. Bivariate regression analysis indicated that social understanding was significant in predicting time spent reading news online (R square = .08, F = 13.39, p = .00) (Table 3). Hence, H4 was supported.
As the fastest growing communication medium of all times, the Internet is not only changing people’s personal lifestyles but also reshaping the interdependence between individuals, media, and society. Dependency is the flip side of control. As we argue for greater consumer empowerment and control over what media content we consume in Internet environments, we are also more likely to grow increasingly dependent on its resources to meet our goals. In terms of individual-media relationships that develop over time, our study suggests tenable connections between individual goals and dependency on Internet resources. On average, Internet users did display moderately intense Internet Dependency Relations, indicating that the medium has become an integral part of individuals’ media environments. IDR intensity appears to be strongest among younger people. In the case of different IDR goal dimensions, students, more than faculty or staff, appear to be more strongly motivated to seek out Internet resources to meet their overall play – and solitary play – goals, emphasizing the entertainment value of media to the younger generation.
This study also finds support for previous research attesting to the growing popularity of online shopping, chatting, and news reading activities among Internet users. Consumers in our study had bought an average of eight products online in the last six months, spent at least thirty minutes per day reading news online, and chatted ten minutes daily on the Internet. Differences by group were, however, evident in the fact that faculty spent the most time reading news online, students spent the most time chatting online, and staff did the most shopping online.
Our study also focused on the extent to which Internet Dependency Relations influenced online shopping, chatting, and news reading. At this stage of the Internet’s development, IDR appears to be a moderate determinant of behavioral responses. In the case of online shopping, the study is consistent with previous research findings that suggest demographic variables are not significant in explaining online shopping variance. However, statistical significance alone is not sufficient to draw conclusions about the predictive strength of IDR, considering the low R square, and future replications are necessary to investigate the impact of IDR intensity on online behaviors, considering the criticality of the Internet-user interface in the commercial world.
We also hypothesized a predictive link between specific goals and online activities. Examining the connection between action orientation and online shopping, we found that individuals who depended on the Internet to meet their action orientation goals were also more likely to engage in shopping-related activities online. This suggests that greater consumer dependence on Internet resources to help make personal behavioral decisions (action orientation) does indeed influence the online shopping experience. However, action orientation did not predict the number of products actually bought online. A possible explanation might be provided by the differences between groups in the number of products bought online. Since staff bought the most products online as compared to students and faculty, it suggests to us that many of the purchases were work-related. If the above conjecture is correct, it is possible that Internet use in work-related shopping contexts may differ from use in personal shopping contexts, and we may argue that media dependency relations, based on the satisfaction of individual personal goals, may not influence purchase behavior in the workplace. In future research, clearer distinction should be made between work-related and personal online shopping. Questions related to the kinds of products bought online may also be included.
While demographics did not affect online shopping, supporting previous research findings, this study indicated that they still have potential to predict other types of Internet use. Age and income are important predictors in online chatting at present; and it appears that online chatting is an activity that younger people with associated lower incomes engage in for longer periods of time than others. The significant correlation of age and play goals also makes sense in the light of the greater intensity of the overall play — as well as the more specific solitary play — dimension among younger people. Hence younger people, who are more dependent on the Internet to meet their play goals, were also the ones more likely to chat online for longer periods of time.
IDR was a significant predictor of the amount of time spent reading news online. A significant gender difference was also observed, with males spending more time than females on this activity. The strong predictive correlation between social understanding and online news reading indicated that people do depend on the Internet’s information resources to understand the world around them. Previous analysis of newspaper readership and dependency relations theorized that social understanding was linked to newspaper reading because a reader’s goals of increasing integration in the community were addressed by newspaper content (Loges and Ball-Rokeach 1993). The same appears to hold true in the Internet-based news environment as well.
By introducing IDR and its goal dimensions as a possible source of variance in online consumer behavior, we hope discussion of its importance and relevance will be further stimulated. Because of its relational aspect, IDR is potentially a better measure than a simple quantification of the extent of Internet use. The Internet itself is inherently more consumer-involving, increasing the likelihood of developing a relationship with it, which in turn is likely to influence the nature and extent of online activities. At present, this relationship appears to be of moderate intensity, but we believe it will strengthen over time. However, the findings in this study are exploratory, and need to be further validated through future research with more general populations.
Our research has some implications for industry as well. E-commerce companies, for example, are strongly motivated to discover reasons that drive shoppers online. Internet portal companies are anxious to increase web site traffic by uncovering motivations that lead people to use chat and instant messaging features. And online newspapers and news web sites are keen to understand how news readers/viewers can be attracted to content on their web sites. Though variance in online shopping and news reading explained by IDR was small, and IDR did not explain variance in online chatting, significant linkages between specific IDR goal dimensions and online activities were observed. Therefore, it is suggested that online purchase action could be made easier, convenient, and action-oriented to serve online shoppers better; facilitating understanding goals could be the strategic focus to serve online news readers; and chat sites can increase traffic by focusing on meeting play goals by making sites fun and entertaining to use (for example, the use of emoticons, view cams, and other devices to make the online chatting experience multi-dimensional).
This study has some limitations. Email surveys generally result in lower response rate than those of telephone or mail surveys. Exploring Shaeffer and Dillman’s (1998) suggestion of using a multi-method approach (combining email with other surveying methods like mail surveys, for example) and initiation of multiple contacts (this study used just two mailings) to improve response rate, may provide some solutions in the future.
Second, the speed of technological advances constantly alters the nature and scope of Internet activities; this may in turn alter the nature and scope of dependency relations as well. Hence, tracking relations through longitudinal analysis may provide a more consistent understanding of the development of individual-Internet relations over time than the cross-sectional approach adopted in this study.
The use of the MSD theoretical perspective in this research may also invite some criticism, due to its limited use in media effects research. In our considered opinion, despite its complex conceptualization, MSD provides a comprehensive and organized conceptual framework to explore individual-media relations. In terms of operationalization, MSD measurement allows cross-media as well as cross-genre comparisons, making it a strong and stable measuring instrument in media analysis.
Future research could examine not just overall Internet dependency, but also dependency on specific types of Internet content, for example online advertising, political information, commercial information, or health/medical information. Replication with national and international Internet user populations could provide another perspective on the development of Internet Dependency Relations and its effects on online consumer behavior. Comparative studies of dependency on Internet and other media, and the extent to which the Internet is/is not affecting dependency on other media or information sources also offer exciting possibilities for future investigation.
Astor, David (2000, May 15), “Survey Finds More Net Use and Trust, Editor & Publisher, 153, 35.
Ball-Rokeach, Sandra J. (1985), “The Origins of Individual Media System Dependency: A Sociological Framework, Communication Research, 12 (4), 485-510.
——— (1998), “A Theory of Media Power and a Theory of Media Use: Different Stories, Questions, and Ways of Thinking, Mass Communication & Society, 1 (1/2), 5-40.
——— and Melvin A. DeFleur(1976), “A Dependency Model of Mass Media Effects,” Communication Research, 3, 3-21.
Ball-Rokeach, Sandra J., August Grant, and A. Horvath (1995), “A Scale for Measuring Media Dependency,” Typescript, Los Angeles, CA: Annenberg School of Communication.
Ball-Rokeach, Sandra J., Milton Rokeach, and Joel W. Grube (1984), The Great American Values Test: Influencing Behavior and Belief Through Television, NY: Free Press.
Bellman, Steven, Gerald L. Lohse, and Eric J. Johnson (1999), “Predictors of Online Buying Behavior, Communications of the ACM, 42 (12), 32-38.
Cyberatlas (2001, June 27), “Yahoo, AC Nielsen Declare E-Commerce Strong and Healthy <http://cyberatlas.internet.com/markets/retailing/article/0,,6061_792621,00.html>
DeFleur, Melvin L. and Sandra J. Ball-Rokeach (1982), Theories of Mass Communication. (4th ed.), NY: Longman.
——— and ——— (1989), Theories of Mass Communication. (5th ed.), NY: Longman.
Donthu, Navin (1999), “The Internet Shopper,” Journal of Advertising Research, 39 (3), 52-58.
Flanagin, Andrew J. and Miriam J. Metzger (2001), “Internet Use in the Contemporary Media Environment, Human Communication Research, 21 (1), 153-181.
Forrester Research Report (2000, June), “The Manufacturer Growth Spiral,” Press Brief issued 6/26/2000. Available on request from [email protected]
——— (1999, November), Brand Doesn’t Drive Young Consumers
Grant, August E., K. Kendall Guthrie, and Sandra J. Ball-Rokeach (1991), “Television Shopping: Media System Dependency Perspective, “Communication Research, 18 (6), 773-798.
GVU 10th WWW User Survey (1998), Report <http://www.cc.gatech.edu/gvu/user_surveys/survey-1998-10/tenthreport.html#ex>
Howard, Philip E. N., Lee Rainie, and Steve Jones (2001), “Days and Nights on the Internet, American Behavioral Scientist, 45, 1-31.
Katz, James and Philip Aspden (1997), “Motives, Hurdles and Dropouts,” Communications of the ACM, 40 (4), 97-103.
Korgaonkar, Pradeep K. and Lori D. Wolin (1999), “A Multivariate Analysis of Web Usage,” Journal of Advertising Research, (March-April), 53-68.
Levins, Hoag (1998), “Growing U.S. Audience Reads News on Net,” Editor & Publisher, Feb. 21, 1998, 14.
Li, Hairong, Cheng Kuo, and Martha G. Russell (1999), “The Impact of Perceived Channel Utilities, Shopping Orientations, and Demographics on the Consumer’s Online Buying Behavior, Journal of Computer Mediated Communication, 5 (2) <http://www.ascusc.org/jcmc/vol5/issue2/hairong.html>.
Loges, William E. (1994), “Canaries in the Coal Mine: Perceptions of Threat and Media System Dependency Relations, Communication Research, 21 (1), 5-23.
——— and Sandra J. Ball-Rokeach (1993), “Dependency Relations and Newspaper Readership,” Journalism Quarterly, 70 (3), 602-614.
Morton, Thomas A. and Julie M. Duck (2000), “Social Identity and Media Dependency in the Gay Community: The Prediction of Safe Sex Attitudes,” Communication Research, 27 (4), 438-460.
Nielsen Netratings/Harris Interactive (2001, April 24), “Nearly Half of all Americans Buy Online, according to Nielsen/NetRatings and Harris Interactive,” <http://www.harrisinteractive.com/news/allnewsbydate.asp?NewsID=273>.
Noack, David (1999, January 16), “Poll says Web News Use in Mainstream,” Editor & Publisher, 132 (3), 26.
Palm Coast/Flagler Internet 2000, “Chatting on the Internet,” <http://www.pcfl.net/channels/chat/default.htm>.
Pastore, Michael (1999, July 15), “Internet Users Taking to Chat,” CyberAtlas <http://cyberatlas.internet.com/big_picture/traffic_patterns/print/0,,5931_1625>
——— (2001, February 7), “Online Chatters among Heaviest Web Users,” CyberAtlas <http://cyberatlas.internet.com/big_picture/traffic_patterns/print/0,,5931_582491,00.html>
Patwardhan, Padmini (2001), “Do Purchasing Involvement, Technology Relationship, and Visual/verbal Orientation Predict Consumer Opinion about and actual use of the Internet for Product Information Searches and Online Shopping?”, Presented to the American Academy of Advertising Annual Conference, March 29-April 1, Salt Lake City, Utah.
Pew Internet & American Life. (2000), “Daily Internet Activities,” <http://www.pewinternet.org/reports/chart.asp?img=6_daily_activities.jpg>.
Pew Research Center for People and Press (1998), “Pew Research Center 1998 Biennial News Consumption Survey: Event Driven News Audience,” <http://www.people-press.org/med98rpt.htm>.
Poindexter, Paula M. (1999), “Xers and Boomers: Are they that different in their Relationship to Web Advertising? In Proceedings of the 1999 Conference of the American Academy of Advertising, M.S. Roberts, ed., Gainsville, Florida: University of Florida.
Scarborough Research (2001, May 9), “First Scarborough National Internet Study Reveals Changes in how Online Consumers use Traditional and Internet Media,” <http://www.scarborough.com/scarb2000/press/pr_internetstudy1.htm>.
Schaeffer, David R. and Don A. Dillman (1998), “Development of a Standard Email Methodology,” Public Opinion Quarterly, 62, 378-397.
Skumanich, Stephanie A. and Kintsfather, David P. (1998), “Individual Media Dependency Relations within Television Shopping Programming: A Causal Model Revisited and Revised,” Communication Research, 25 (2), 200-219.
Weir, Tom (1999), “Innovators or News Hounds? A Study of Early Adopters of the Electronic Newspapers,” Newspaper Research Journal, 20 (4), 62-81.
Individual Media Dependency Relations Scale
(Five point scale from not at all helpful to very helpful)
In your daily life, how useful is the Internet to¡¦
Gain insight into why you do some of the things you do
Imagine what you’ll be like when you grow older
Observe how others cope with problems or situations like yours
Stay on top of what is happening in the community
Find out how the country is doing
Keep up with world events
Decide where to go for services such as health, financial, or household
Figure out what to buy
Plan where to go for evening and weekend activities
Discover better ways to communicate with others
Think about how to act with friends, relatives, or people you work with
Get ideas about how to approach others in important or difficult situations
Unwind after a hard day or week
Relax when you are by yourself
Have something to do when nobody else is around
Give you something to do with your friends
Have fun with family or friends
Be a part of events you enjoy without having to be there
Online Shopping Scale
(Five point scale from Very Frequently to Never)
I shop on the Internet
I buy many different products on the Internet
I make use of online discounts on goods and services
I follow up on good deals on the Internet
I buy a product online even if other buying options are available
Padmini Patwardhan is an Assistant Professor in School of Mass Communications at Texas Tech University.
Jin Yang is a Doctoral Student in College of Mass Communication & Media Arts at Southern Illinois University Carbondale.
*This is an invited article
Copyright © 2003 Journal of Interactive Advertising