WJMCR 11 (January 2008)
This national survey found that a notable percentage of people depend on the Internet as a valued source of in-depth information about health, science, and business. Between 31% and 50% of the respondents said they use the Internet weekly for in-depth information in one of the three areas. These respondents valued the Internet more than magazines, books, or friends and families as a source of in-depth information. In-depth information is useful to people who depend on media for understanding and orientation about issues and topics. Individual background variables were better predictors of whether people use the Internet for such information than they were of people�s evaluation of that information�s quality.
For more than three decades, media-systems dependency theory (MSDT) has provided a theoretical basis for explaining the relationship among individuals, institutions, and media at both macro and micro levels. During that period, digital media have grown to become an integral part of daily life for many people. In 1998 Ball-Rokeach1 discussed the implications of this growth, anticipating the impact of the Internet on media dependency, when she wrote: �The Internet thus intrudes on traditional relations by being integrated into an expanded media system that may expand the reach of understanding, orientation, and play goals that individuals, groups, and organizations may attain through media dependency relations.�2
Since Ball-Rokeach wrote those words, the Internet has indeed become integrated as an important source of information and entertainment in the United States. In 1996, 28.8 million adults in the United States had access to the Internet, 16.4 million used the Internet, and 11.5 million used the Web.3 By early 2006, 147 million adults in the United States used the Internet.4
Although a great deal of effort has gone into describing the growth of the Internet and the types of information accessed, little research has used the media-systems dependency theory as a framework for exploring how the Internet has become integrated into individuals� lives. This study uses data from a national survey of people�s regular use and evaluation of the Internet for detailed, in-depth information about specialized topics to explore micro-level dependency on the Internet. We assume that �regular use� of the Internet, or any other source for that matter, for specialized information indexes individuals� dependency on that source more than does general and non-specific exposure. Moreover, dependency on a source or medium does not require exclusive use of that medium, nor even daily use, but regular use indicates whether the medium constitutes an important part of the individual�s information mix. In addition, the study explores the characteristics of individuals who go online for three types of specialized information (health, science and business) to look for common predictors of dependency and perceived usefulness.
This research is important because it analyzes the integration of the Internet into people�s lives using a well-established theory, and it is a first effort to evaluate the degree to which Americans depend on the Internet for in-depth specialized information, a class of information that is useful for the orientation and understanding dependencies as specified by MSDT.
Media Systems Dependency Theory and Related Approaches
Media-systems dependency theory emerged from a 1976 article by Ball-Rokeach and DeFleur5 that attempted to explain why media could have varying cognitive, affective and behavioral effects on different people. Over time, it evolved into a more complex theory that deals with the relationship between media and individuals at the micro level and media and social institutions at a macro level.6
At the micro level (individuals), MSDT assumes individuals are goal-oriented and active in the selection and use of media content. It suggests three dependency areas, with two sub-areas each, in which different individuals depend on media to varying degrees7: 1. for solitary play and social play; 2. for self-understanding and social understanding; and 3. for action orientation and interaction orientation. In particular, the action and interaction orientation dependencies assume that people act purposefully in deciding how they will behave to obtain goals.
As Ball-Rokech et al.8 pointed out, orientation dependency is closely related to understanding dependency. Both these dependencies require the acquisition of information for specific goals that often involve complex topics and issues; because of that specificity and complexity, the information needs to be detailed and in-depth.
Although MSDT has been used as the theoretical basis for some studies of the Internet9, no published research was found that also addresses the particular relationship (e.g., displacement vs. complementarity) between the Internet and more traditional media with MSDT as a framework.
However, in 1993 Lacy et al.10 used MSDT and uses and gratifications concepts to create a typology of media uses from an economic perspective, and in 2000 Lacy11 extended this typology to a wider range of information. This approach hypothesized a media mix, represented by a matrix of uses and media types, for each person. The particular mix of media for meeting an individual�s goals varies from person to person and across time. The media mix is a subdivision of a larger information mix, which includes interpersonal communication as an information source, for fulfilling an individual�s goals. This mix does not assume that a single medium is the sole source of all the information a person needs to reach a goal or fulfill a need. However, the typology suggests that the mix is relatively stable, with some variation, during the short run. The mix goes through major shifts when disruptions occur in the availability of existing media and when new media products are made available.
Both MSDT and the media-mix model indicate that the development of the Internet would introduce new sources of dependency for information. Neither suggests that the use of the Internet will displace traditional media for everyone, although it might for some. Rather, the two approaches suggest that some people will replace the time spent with existing media for time spent online, and that this displacement may vary with the type of dependency motivating the Internet use.
Uses of the Internet
Although little research has been published addressing individual dependency on the Internet, scholars have begun to examine how and why people use the Internet. Some have applied the �uses and gratifications� approach to the study of the Web.12 For example, in 2000 Papacharissi et al.13 studied people�s motivations for �computer-mediated communication� in general, compared to interpersonal communication, and identified five dimensions via factor analysis: interpersonal utility, information seeking, passing the time, convenience, and entertainment. More narrowly, a 2001 survey of 98 U.S. adults14 identified five primary factors for Web site use: search, cognitive, new and unique, social, and entertainment. An even more narrowly focused study in 2002,15 analyzed an online survey of 308 U.S. adults about using the Web for political information and found four primary motivations: guidance information-seeking, surveillance, entertainment, and social utility.
Despite their focus on �new media,� researchers have found that some of these uses and motives are common to traditional media. Researchers point out that the Internet is used in a manner similar to other, more traditional media,16 a conclusion supported by research that found people seeking the same broad content areas (politics, sports, business, health, science) online as they seek in traditional media. This research indicates at least some people might be substituting information acquired online for information gained from traditional media.17 The uses and gratifications approach aims to identify motivation for the individual use of media content. However, this approach has often been viewed as limited for a number of reasons. For example, the specific uses and gratifications that have been identified often vary from study to study, which highlights the atheoretical nature of much uses and gratifications research. In addition, surveys are limited in their ability to identify internal gratifications.18
In many ways, media-dependency systems theory is more attractive than uses and gratifications for explaining the micro-relationship between individuals and media for a variety of reasons.19 It is a theoretical approach to studying how people interact with media, and the theory covers a much wider range of behaviors toward media. It becomes even more attractive given that many of the uses (if not gratifications) identified through research fit within the six individual areas of media dependency in the MSDT typology.
Displacement of Traditional Media by the Internet
Even though MSDT has not been used as the basis for studying media substitution, some studies have examined whether or not and to what degree online information access is displacing the use of traditional media. Most of these studies have addressed the substitution of the Internet for traditional news media such as newspapers, radio, and television news.
In 2005 De Waal et al.20 surveyed about 1,000 Dutch adults about their media use and concluded that use of online newspapers varies along several demographic variables�age, gender, education, and income. Overall, however, they concluded: �Once we compare the usefulness of online newspapers for specific areas of information with print newspapers and the other information channels, online papers cannot compete with printed newspapers or television. Both are considered more suitable for all kinds of information.�21 In effect, online newspapers were more likely to be complements than substitutes for other news media.
This finding was supported by a survey conducted in Austin, Texas. In 2002 Chyi et al.22 concluded that a print version was preferred to reading the same newspaper online. The online newspapers in the study had gained only a small number of non-print newspaper readers, and Chyi et al.23 concluded that online and print newspapers were more complements than substitutes. Unlike the 2005 study by De Waal et al,24 which examined the correlation of time spent with various news media, Chyi et al.25 explored whether study respondents would replace the print versions with the online version.
In a 2000 study of Spanish newspapers that had gone online, Dans26 found that online and print newspapers varied in how they were read. �The use of the newspaper is different too: reading of Internet newspapers is usually more functional and goal oriented, as indicated by the small number of pages read per visit,� he wrote.27 However, he said there was a risk that online newspapers could be substitutes for print newspapers. Indeed, Gentzkow28 studied Washington daily newspapers and concluded that online and daily newspapers were substitutes, but the impact was not as great as some have predicted.
Waldfogel29 used existing data to explore correlations among the use of various media and concluded �certain media appear to compete with each other for consumers� attention.�30 Waldfogel lists a number of possible substitution patterns but only explores substitution in one topic area�political news. The study concluded that if substitution of other media for newspapers were complete, civic activity would not be affected. However, it also concluded that substitution is not yet complete.
Kaye et al in 2003 surveyed Internet users interested in politics in 2000 and compared the results with a 1996 survey. The comparison indicated that the time spent seeking political information stayed about the same, but those who were changing their media patterns were abandoning traditional media in a way that suggested the Internet was serving as a substitute rather than a complement to traditional media.
Overall, these studies suggest that some people use the Internet in lieu of traditional media, but the impact has not become as extensive as futurists expected. It appears that some people use the Internet as a substitute for traditional media but others do not. However, these studies were conducted as the use of the Internet was expanding, and the nature of Internet use has possibly changed during this period. In addition these studies do not address particular uses or, with the exception of political information, the type of information people are seeking online.
In-depth, Specialized Information and the Internet
The growing use of the Internet reflects many factors, but at least four characteristics of the Internet stand out when compared to traditional media: 1. The Internet features almost instantaneous delivery of information, which makes it faster than traditional print media. 2. The Internet is interactive, which allows more immediate feedback than any other medium. 3. The Internet allows for the distribution of multimedia content, which many media do not. 4. The Internet allows for the distribution of large amounts of low-cost, in-depth information. These four relative advantages of the Internet help explain its spread and it its potential importance in the micro-level dependency between individuals and media producers.
The studies of Internet information use have tended to concentrate on patterns of substitution and complementary use of the Internet, primarily with regard to news, thus focusing on instantaneous delivery and multimedia content. Use of the Internet for in-depth information remains largely unstudied as a class of behavior. This study aims to fill that gap by examining individual dependency on the Internet for business, science, and health information. Each of these complex topics affects people�s everyday life and requires a depth of understanding. Further, business and health topics often relate to important subsequent behaviors and actions by individuals. While researchers have dealt tangentially with Internet use for business and economic news, often in terms of its competing with or complementing other media,32 little if any attention has been paid to people�s use the Internet for in-depth business information. However, one study�which also examined individual background variables�concluded that people who looked for business and finance news online were more likely to be male, older, and more educated than non-users.33
Research focusing on the use of the Internet for science information is also somewhat limited. In 2006 Horrigan34 reported that 20% of adults said they use the Internet as their main source of science news, and 33% of broadband users say they get most of their science information from the Internet. However, Horrigan did not explore the concept of dependency.
Unlike science and business information use, Internet health information has seen a boom in research. Studies show that health and medicine have become the fourth most popular online search subjects, behind news, travel, and weather.35 Further, some 52 million Americans sought health-related information over the Internet at least once a month in 2004, and 40% of them indicated that online health information affected their health care decisions.36
A national survey of more than 4,700 Americans found that 40% of respondents with Internet access reported using the Internet to look for advice or information about health or health care in 2001.37 Significant use of the Internet by patients also was reported in another study, which surveyed 512 randomly selected primary care patients. The study found that 53.5% had used the Internet for health information.38 The average age of those using the Internet for health information was 45.8 years, and a large proportion was college-educated with annual household incomes greater than $50,000. One-fifth (19%) of those using the Internet for health information said they went online at least once a week and 35% said they used the Internet at least monthly. In terms of the quality of health information found on the Internet, 62% said it was �excellent� or �very good,� 32% said it was �good,� and 6% agreed the quality was �fair.�39
A 2006 survey by the Pew Internet & American Life Project40 estimated that 113 million U.S. adults, which is 80% of the Internet users, have searched for health information on one or more of seventeen health-related topics, and some types of Internet users were more likely to go online for health information than other types. The probability of searching for online health information was higher for women, people under the age of 65, college graduates, experienced online users, and, not surprisingly, those with home broadband access. Some of these studies of specialized Internet information seeking have explored the role of demographic or background factors, an approach also taken with the present research. While basic Internet access and potential information access vary along social structural lines, as argued by those exploring broader questions about �digital divides� and knowledge gaps,41 the question remains: What background variables come into play in determining who seeks in-depth specialized information online?
Unlike much previous research, this study does not concentrate on patterns of general media use (television, newspapers, radio) and general (almost �passive�) information acquisition (e.g., �Where do you get most of your news about what�s going on in the world?”). It explores in-depth specialized types of information attended to by large segments of the public (business, science, and health) and media sources typically used in active seeking of information. It is assumed that understanding and orientation dependency requires in-depth and specialized information, which can be found in some but not all newspapers. For example, Loges et al.42 found that social- and self-understanding was assciated with time spent reading newspapers but action-orientation dependency was not.
Although research has examined use of some forms of online specialized information, such as health and science, no published research could be found that explored MSDT as a foundation for comparing the usefulness of different sources across more than one form of detailed, in-depth information. Because little research exists in this area, this study will address research questions instead of hypotheses. These questions aim to discover people�s dependence on the Internet for in-depth specialized information; how people evaluate the usefulness of the Internet compared to other sources of in-depth information; and what individual background factors are related to the dependency and the evaluation of the Internet for in-depth specialized information.
RQ1. What percentage of people in the United States depends on the Internet for in-depth specialized information about business, science, and health?
RQ2. How do the people who depend on the Internet for in-depth specialized information in business, science, and health evaluate it relative to other sources of in-depth specialized information?
RQ3. Do these evaluations vary with the type of in-depth specialized information?
RQ4. Do people who use the Internet for in-depth specialized information about business, science, and health differ from those who do not?
RQ5. What sorts of background variables are associated with the evaluative ratings of the Internet for in-depth specialized information?
Trained interviewers employed by a professional survey center and using computer assisted telephone interviewing equipment conducted a national survey of 1,031 adults at randomly selected phone numbers between Oct. 7 and Oct. 24, 2006. Numbers were generated by randomly selecting zip codes and working telephone prefixes serving the zip codes, and then generating the final four digits of the phone number. Non-working numbers and refusals were replaced by adding �1� to the number and dialing again. Of 2,038 viable dialings, 1,031 or 51% yielded completed interviews. A conservative estimate of sampling error associated with a probability sample of 1,031 is +/- 3% at the 95% confidence level. Of course, sampling error is only one source of error, and for many sub-analyses the sample size is smaller than 1,031.
As a measure of dependency, interviewees were asked whether they regularly (�at least once a week�) �seek out detailed, in-depth information� about a particular information type (business, science and health). All three of these areas are complex and require in-depth and specific information in order for individuals to understand (understanding dependency) and act (orientation dependency) on decisions about these types of issues and problems. As noted earlier, dependency on a medium does not require exclusive use of that medium, or even daily use, but regular use indicates whether the medium is an important part of the individual�s media mix.
In order to evaluate the level of dependency on the Internet relative to other media, a battery of questions designed to measure the usefulness of four different information sources (the Internet, magazines, books, and �friends and family�) were asked for three types of specialized information (science, health, and business). The questions asked respondents to evaluate (using a 1-10 scale, with 1 being �least useful� and 10 being �most useful�) each of the four sources: the Internet, magazines, books, and �friends and family.� The comparison sources of information (magazines, books, and family and friends) were selected because they are frequently sources of in-depth information that does not appear on a regular basis in newspapers or on television news. In addition, comparison of Internet use to such sources has not been researched to the degree that comparisons with news media have been studied.
In addition to providing demographic information, respondents also characterized their political partisanship. As is often the case, the unweighted sample data overrepresented women (59%,) and 87% identified their race as white. Both sample percentages were higher than the national population estimate (51% female and 75% white) from the U.S. Census.43 A fourth (24%) of the sample claimed to be four-year college graduates, slightly less than the national population level of 27%,44 and 7% reported having less than a high school degree.
Among income brackets for those responding, the largest proportions (at 19% each) were in the $40,000-$60,000 range and the $25,000-$40,000 range. Fifty-seven percent were married and 75% reported having children. More than a third of respondents (35%) described their residence as a �large city,� while 15% reported living in a rural area. Thirty-eight percent identified their political party as Democrat, while 30% reported being Republican. The sample was slightly more likely to be Democrat and slightly less likely to be Republican than the percentages reported by the Gallup Polls.45 Despite these variations, the national sample provides a level of information about Internet use that is often missing in academic studies.
The research questions were addressed with a variety of statistical procedures. The cut-off point for statistical significance was p < .05. RQ2 (asking about differences in evaluation of in-depth information among sources) and RQ3 (asking about differences in evaluation of sources by information type) were addressed with t-tests that compared respondent means for the evaluation scales.
RQ4 (asking whether those who depend on the Internet for in-depth specialized information differ in background from those who do not) was addressed using logistic regression analysis, which predicts a dependent variable in terms of �group membership� (e.g., seekers versus non-seekers of depth science information), on the basis of a set of independent variables. The logistic regression model calculates regression coefficients (B), which illustrate the relative contribution of each independent variable in predicting membership in the two groups.46 Also, the model calculates an Exp(B) for each independent variable, which indicates the increase or decrease in the probability that a case will be classified in a group for each unit increase in the independent variable. For example, an Exp(B) of 3 indicates a one-unit increase in the independent variable for a case means the case is three times as likely to fall into the group with the characteristic under study. An Exp(B) equal to 1 indicates the independent and dependent variables are independent, which means there is 50% chance that the case will be in the group defined by the dependent variable.47 In such a case, prediciting group membership becomes a coin flip.
After the regression equation is calculated, how well it predicts can be determined by using it to classify cases. Then the percentage of correctly classifed cases is calculated to see how well the overall equation predicts. The higher the percentage, the better the equation predicts and the stronger the relationship between group membership and the independent variables�in this case, background or demographic variables.
In order to examine the relationship between the evaluation of specialized in-depth information and demographic variables (RQ5), ordinary least squares multiple regression analysis was used. The data were first examined for violations of regression assumptions. The data did not contain outliers (defined as cases three standard deviations above or below the mean), skewness figures were less than +/- 1.0, and there was no evidence of multicollinearity.
The independent variables used in the regression equations included age, a ratio level variable, and socio-economic status, calculated by multiplying the six-interval education level variable by the eight-interval income variable. The two were combined because the combined variable had higher correlations with the dependent variables than did the education and income variables separately; both have been shown frequently to relate individually to general Internet access.48 The remaining independent variables were dummy variables, including whether the respondent was employed or not, whether the respondent lived in an urban area or not, whether the respondent was politically conservative or not, and whether the respondent was a woman or man.
RQ1 asked what percentage of people in the United States depends on the Internet for in-depth specialized information. Data from Table 1 indicate that the percentage of respondents who went online varied significantly for the three types of specialized information. Health information was sought by the largest proportion of the respondents, with 50% reporting that they sought in-depth health information at last once a week. In-depth science information ranked second with 40% seeking such information at least once a week. In-depth business information was the least sought of the specialized information types, with 31% of respondents looking for business information at least once a week.
Using the confidence interval of plus or minus 3%, the sample data indicate that between 47% and 53% of people in the United States went online at least once a week for health information, between 37% and 43% went online for science information, and between 28% and 34% went online for business information.
|Percentage of Respondents who Access In-depth Information at Least Once a week||Science||Health||Business|
|Do Not Access||56%||45%||65%|
|Margin of error = + 3%.|
RQ2 asked how people who depend on the Internet for in-depth specialized information evaluate it relative to other sources of in-depth specialized information. The respondents who had indicated that they used the Internet for each type of in-depth information were then asked to rate the Internet, books, magazines, and friend and families for usefulness (1=least useful, to 10=most useful).
Data from Table 2 show that the respondents who depend on the Internet for in-depth information about these topics consider the Internet more useful than magazines, books, and family and friend for most forms of specialized information. The differences exceeded one on a 10-point scale in all comparisons except between the Internet (7.6) and books (7.36) for science information. The lowest rating among all media for all three types of information was 5.79 for family and friends as a source of science information. The highest average ratings for any medium and information type other than the Internet for all three types of information was 7.36 for books and science information and 6.93 for books as a source of health information.
|Means for Usefulness of Various Media for In-depth Information (Standard Deviation and N in Parenthesis)||Science*||Health**||Business***|
|Family & Friends||5.79|
|* The differences between the Internet and the other three types of media is statistically significant at the p < .001 level.|
|** The differences between the Internet and the other three types of media is statistically significant at the p < .001 level.|
|*** The differences between the Internet and the other three types of media is statistically significant at the p < .001 level.|
RQ3 asks if source evaluations vary with the type of in-depth, specialized information. Table 2 data show that the variation among the three types of specialized information was small. For all three types of information, the Internet averaged an overall rating higher than 7.50, with health information averaging 7.59, business 7.68, and science 7.76. The Internet was rated highest in all three areas.
RQ4 asked if people who depend on the Internet for in-depth specialized information differ from those who do not. Table 3 reports the logistic regression results for the three specialty areas. The results represent the relationship between background variables and group membership.
Four background variables were statistically significant in predicting who did and did not depend on the Internet for in-depth science information. A one-unit increase in socio-economic status resulted in a 16.2% increase in the likelihood of a person accessing in-depth science information online. Three background variables were significantly but negatively related to science information seeking. Conservatives, women, and employed people were less likely to access in-depth science information online. However, the overall model correctly predicted only 59.8% of the cases, which indicates that the model is missing some important variables, that the measurement needs to be improved, or both.
|Logistic Regression Predicting Who Will Access In-depth Specialized Information Online||Science Information||Health Information||Business Information|
|% cases correctly classified||59.8%||58.3%||66.1%|
|d.f.||6, n=694||6, n=693||6, n=693|
|* p < .05|
|**p < .01|
|***p < .001|
The second logistic regression model in Table 3 showed that only two background variables reached statistically significant levels for predicting whether a person depended on the Internet for in-depth health information. Somewhat surprisingly, given the concerns of digital divide critics about gaps in important knowledge�like health�that result from differential Internet access, socio-economic status was not a significant predictor. Employed people were less likely to look online, but the strongest predictor was gender. Women were 59.1% more likely to look online for health information. However, the overall regression was not very good at differentiating between those who looked online for health information from those who did not. The equation correctly predicted only 58.3% of the cases.
The third logistic regression equation in Table 3 shows the relationship between background variables and depending on the Internet for in-depth business information. Only three variables reached statistical significance. Being a woman was negatively related to depending on the Internet for business information, while having high socio-economic status and living in an urban setting were positively related. The percentage of cases correctly predicted was higher than the other two logistic regression equations with the model correctly predicted 66.1% of the cases.
Ordinary least squares regression was used to test RQ5 about what sorts of background variables are associated with the evaluative ratings of the Internet for specialized information. Five demographic variables were used. As noted above, socioeconomic status (SES) was calculated by multiplying the education variable by the income variable. In addition, age was included, along with three dummy variables–conservative political leaning, being currently employed, and living in an urban area. The regression equations are found in Table 4.
All three regression equations were statistically significant but explained only small amounts of variance (e.g., only 3.1% of the variation in the rating for Internet science information, and 4.4% of the variation in the usefulness rating for Internet business information). The equation with evaluation of Internet health information as the dependent variable had the strongest association, explaining 8.2% of the variance.
|Regression of Demographic Variables on Usefulness of the Internet for In-Depth Information (Beta weights)||Science Information||Health Information||Business Information|
|Live in urban area||.054||.031||.021|
|(d.f. = 5, 274)||(d.f. = 5, 330)||(d.f. = 2, 217)|
|* p < .10|
|** p < .05|
|*** p < .01|
Age was the only variable that was statistically significant in all three equations, and it was negatively related to the usefulness rating for all three specialized information areas. Age had a -.127 beta weight for predicting the science information rating. The beta weights for age were -.224 for health information and -.223 for business information. As the age of the respondents increased, the ratings for all types of information online decline, although the decline is much greater for health and business information.
A key to understanding Media Systems Dependency Theory and its application to digital media is to study variations in how people depend on the Internet for in-depth, specialized information and how they evaluate the Internet relative to traditional media for acquiring this class of information. This study used a national probability sample to examine the dependency of the Internet for in-depth specialized information about science, health, and business.
Results show that a significant percentage of people, from a third to a half, depend on the Internet weekly for these types of information. In-depth, specialized information is important in helping people who depend on media for understanding and for orientation. Because the cost of storing and distributing this type of information online is inexpensive compared to traditional media, the Internet holds great promise for allowing people to access in-depth specialized information. This potential is enhanced even more by the search capabilities of the Internet
Such a survey cannot explain why dependency varies among the types of information. However, about half of the people said they depended on the Internet for health information. This is consistent with existing research and is likely explained by the fact that everybody faces health problems at some point in their lives. However, an interest in science or business likely reflects an acquired interest. This suggestion is supported by findings that those who sought health information online and those who did not varied on only one demographic variable�gender. However, those who sought science information differed on five demographic variables when compared to those who did not, and those who sought business information differed from those who did not on four demographic variables. With both science and business information, those with higher socio-economic status sought in-depth information online.
The only common significant predictor of online information seeking was gender. A higher percentage of women depended on the Internet for health information, but a smaller percentage of women sought in-depth online information about science and business. Gender was related to self-understanding, social-understanding, and action orientation in a study of dependency and time spent reading newspapers.49
Not surprisingly, people who depend on the Internet for in-depth specialized information rate it higher than other types of media and even family and friends, although there were some variations in responses. For example, people who sought in-depth science information online evaluated books highly as a source of science information. Those who sought health information online also rated books high, but those who sought in-depth business information online rated books very low.
These results for business information are particularly important, because they suggest a higher dependency on Internet than other sources for these respondents. However, the dependency was not exclusively on the Internet. The mean evaluations for usefulness exceeded the neutral point, five, for all sources of information and all topics. The results also are consistent with the idea of people having a media or information mix.50 As hypothesized, the development of digital media has resulted in a reformulation of individual information mixes.
Demographic or background characteristics were clearly related to the use of some forms of online information, but this study also investigated a relationship of demographics to the evaluation of the Internet as a source for three types of information. However, the three regression equations did not explain much variance in the evaluation�only 3.1% to 8.2%. This may reflect the limited range in the dependent variables, which were measured on a 1-10 scale. But even with a better measure of the usefulness of in-depth specialized Internet information, it is likely that cognitive and affective variables are related to these evaluations. These variables could take a variety of forms, such as salience, the success of previous decisions based on the information, or the ease of finding and understanding the information. The small variance explained suggests further investigation of the relationship among user and content characteristics.
The one consistent predictor of the evaluation of the usefulness of in-depth Internet information was age. The older the respondent, the less likely he or she was to rate the Internet information high in usefulness. This result is consistent with most studies about the impact of age on the use and evaluation of the Internet. Age was related to self-understanding, social-understanding, and action-orientation dependency in a study of newspapers and MSDT.51 The relationship between age and the evaluation of the Internet as a source is consistent with the idea that habit (stability of media mix) plays a role in the evaluation of media. However, the magnitude of the relationship found here was small. This probably reflects the fact that all of the respondents asked to evaluate the Internet did, in fact, use the Internet for in-depth specialized information, or the limited range of the measurements.
Interestingly, socio-economic status was related only to the evaluation of the usefulness of health information. One might assume a connection between SES and evaluation of business information would be found, but it was not. However, socio-economic status was related to whether respondents sought business information online.
Overall, these national data indicate that the Internet is a highly valued source of in-depth specialized information in multiple areas and that a notable portion of people depends on the Internet for health, science, and business in-depth information. These results are consistent with the study of dependency and newspapers.52 As Ball-Rokeach53 predicted, individual media dependency is shifting from traditional media to the Internet for some forms of information.
Of course, this study has its limitations. Although it used a national probability sample, comparison of the sample demographics with Census data indicates that the sample was slightly older with more women than the U.S. population. The background variables, with the exception of age, were measured with nominal scales, which introduced an unknown level of measurement error. In addition, this study assumed a relationship between in-depth, specialized information and two categories of dependency (understanding and orientation) based on the theoretical definitions of these dependencies. This assumption should be investigated.
Future research should aim to correct these limitations. In addition, the research should investigate media dependency on the Internet for in-depth specialized information in additional topics. The three studied here are common topics. Perhaps an examination of more narrow information topics such as hobbies, narrowly focused topics, or even entertainment would yield useful findings as well. In addition, different measures of dependency could be used to examine the relationship between in-depth, specialized information and various forms of dependency.
Riffe is a professor and Presidential Research Scholar in the E.W. Scripps School of Journalism at Ohio University and Lacy is professor in the Department of Communication and School of Journalism at Michigan State University, where Varouhakis is a student in the Media and Information Studies Ph.D. Program