Of the 137 subjects, 106 were undergraduate students, two were faculty members with doctorates and 27 others were graduate students or staff, while two did not report their educational level. The age of the subjects ranged from 18 to 48, with a mean of 21.7. Eight of them were over age 30. Of the 116 subjects who identified their race, 16 (13.8%) were black, three were Hispanic, three were Asian American, six were Asian foreign students, and one man chose the eighth option I offered: “mixed race.” Of the 87 whites in the study, eight identified themselves as not being American citizens, which could alienate them from our hegemonic society at times, too.
It turned out that exactly two-thirds of the subjects were women, which reflects the predominance of women at the University of Maryland, College Park; Howard University; and The American University--especially in the journalism programs. When distinguished by sociological gender, however, a much slimmer majority of 52.6% of the study subjects scored higher on the feminine scale than on the male scale, indicating that a significant number of these female incipient journalists were already exhibiting the competitive, masculine-associated tendencies that society ascribes to journalists (Walsh-Childers et al., 1996).
Appendix K shows the tally of reasons the study subjects listed for accessing each Web page or e-mail message they opened. It shows they chose to spend an average of one-quarter of their 20-minute online session reading or writing e-mail, while they spent three-quarters of their time on the Web. The single-most popular online activity was visiting sports Web sites such as ESPN, CNNSI, and CBS Sportsline. Many of these page views seemed to have been spurred by preseason hype for the Maryland Terrapins basketball team. Another factor that may have biased this study slightly is the local popularity of washingtonpost.com, which drew the vast majority of the 156 “hits” on newspaper sites. Later analysis showed that these sports and newspaper Web pages were almost invariably hegemonic. The crux of my later analysis was deciding the latent hegemony in the many Web pages that fell into the catch-all category, “Web use for personal information.” Most of these pages ultimately were coded as counterhegemonic or pluralistic.
I built this research study deductively. My theory of hegemony online is based on a literature review of related theories and the history of how other communication media evolved. I developed hypotheses to test my theory, and then I developed the digital diary and questionnaire to operationalize those hypotheses. Following is a review of each hypothesis and a report on whether or not the data support it (nine of 14 were supported), at an alpha level of at least .05.
H1: People will spend most of their online time reading hegemonic Web pages. Supported
Of the 1,142 online destinations identified by the subjects, 424 or 37% were coded as hegemonic, 313 were classified as dominant culture Web pages, 344 (mostly e-mail messages) were coded 0, there were 17 closed counterculture pages, and 44 were counterhegemonic.
H2: The longer people use an online message, the higher they'll rate its quality and
I tested this hypothesis with a linear regression between the Z score for the quality rating that subjects gave each Web page or e-mail message and the minutes they spent on each online destination (Table 1).
Summary of Linear Regression Analysis for Using Time Spent on a Web Page or E-Mail Message to Predict Its Perceived Quality (N = 1,145)
Variable B SE B Beta
(Constant) 0.315 0.033
Time spent on message 0.065 0.010 0.193***
Note. R2 = .037; adjusted R2 = .037.
***p < .0005
H3: The less time people spend on e‑mail each week, the less time they spend in chat
groups, and the less time they’ve had a home page, the more use they will make of hegemony online. Not supported
H4: The more time people spend on the Web each week, the higher their ratio of Web
usage to e‑mail usage, and the more purchases they make online, the more use they will make of hegemony online. Not supported
H5: The more experience people have on the Web, the more use they will make of
hegemony online. Not supported
H6: People who have a home page with ads or commercial credits--but no chats--will
make more use of hegemony online. Not supported
H7: People who do not list a counterhegemonic Web page as their favorite will make more use of hegemony online. Supported
This inverse relationship was supported by an independent t-test of the hegemony rating for the favorite Web page listed by subjects and the hegemony ratings of the pages they chose during their 20-minute online session, t (1, 1063) = -2.861, p = .004. Among those who listed a counterhegemonic Web page as their favorite, the average hegemony coding for the pages they chose during their 20-minute online session was .15 (from a possible range of -1.0 to 1.0), compared to a mean of .48 for those who listed a hegemonic or pluralistic page as their favorite.
H8: People of color will make less use of hegemony online. Supported
Although there was little difference in use of hegemonic Web pages among the races, cross tabs show that whites chose counterhegemonic Web pages 2.9% of the time, while nonwhites chose them 7.8% of the time, F (1, 4) = 11.445, p = .022. And a further breakdown shows that African Americans, specifically, chose counterhegemonic Web pages 13.9% of the time, F (1, 28) = 74.102, p < .0005.
H9: The higher people’s household income is, the more use they will make of hegemony
online. Not supported
H10: The more education people have, the more use they will make of hegemony online.
Cross tabs show that undergraduates chose hegemonic Web pages 33.9% of the time, compared to 46.8% for graduate students, faculty, and staff. Conversely, the undergrads chose pluralistic destinations 31.4% of the time, compared to 25% for grad students, faculty, and staff, F (1, 4) = 16.903, p = .002.
H11: The older people are, the more use they will make of hegemony online. Supported
Since age is a continuous variable that yields ratio data I was able to test this hypothesis by regressing age on the study subjects’ hegemony scores (Table 2). I also plotted age against each page’s hegemony ranking, using Spearman’s rho, which yielded a correlation coefficient of .09 at .003 significance.
Summary of Linear Regression Analysis for Using Age to Predict the Usage of Hegemony Online (N = 134)
Variable B SE B Beta
(Constant) 2.409 2.519
Age 0.229 0.113 0.173*
Note. R2 = .030; adjusted R2 = .023.
*p < .05
Although age is a continuous variable, there was relatively little age variation among my study subjects, so I also analyzed age as an ordinal variable, breaking down the age data into high, medium and low. The “low” third comprised study subjects 19 and younger, the second third comprised ages 20 and 21, and the “high” third comprised those 22 or older. Even these narrow divisions yielded a highly significant Chi-Square (p < .001) for cross tabs with the chosen page’s hegemony ranking, further supporting the age hypothesis.
H12: Men will make more use of hegemony online than women will. Supported
This relationship was made evident by an independent t-test that showed the mean hegemony score for men was 10.01 and only 6.08 for women, t (1, 134) = -3.59, p < .0005. Cross tabs also showed support for this hypothesis, F (1, 4) = 67.215, p < .0005 (Appendix L).
H13: The higher people rate toward the masculine end of the gender scale, the more use
they will make of hegemony online. Supported
Cross tabs between the hegemony ratings for each of the online messages chosen by study subjects and the Z scores of their responses on the gender scale yielded results that were in the predicted direction and highly significant, F(1, 4) = 18.451, p < .001 (Appendix M).
H14: The lower people rate on the alienation scale, the more use they will make of
hegemony online. Supported
Along with the 30-item gender scale, the seven-item alienation scale proved to be quite robust. Linear regression yielded a Pearson Correlation of -.22 between the subjects’ alienation and hegemony scores, at .005 significance. The relationship’s adjusted R2 was .041, F (1, 135) = 6.87, p = .01 (Table 3).
Summary of Linear Regression Analysis for Using Alienation to Predict the Usage of Hegemony Online (N = 137)
Variable B SE B Beta
(Constant) 5.966 0.753
Alienation score -2.976 1.135 -0.220**
Note. R2 = .048; adjusted R2 = .041.
**p < .01
Quantitative Research Question Answered
Along with these 14 predictive hypotheses, I tested the hypothesis that some combination of variables would significantly aid in predicting the use of hegemonic online communication. A forward selection multiple regression, entering as independent variables alienation, months of Web experience, the log 10 of weekly hours on the Web, and the log 10 of weekly hours on e-mail yielded a model with an adjusted R2 of .065, F (4, 119.568) = 3.250, p = .014. A backward elimination method of multiple regression on these four independent variables enhanced the perspicuity of the model by eliminating months on the Web as an independent variable while improving the significance and only slightly diminishing the adjusted R2 to .063, F (3, 127) = 144.619, p = .010.
Deleting the five extreme outliers from the Web variable and the two extreme outliers from the e-mail variable, improved the adjusted R2 to .066, F (3, 122) = 145.903, p = .010 (Table 4). Also, this model presented no collinearity problem (the highest condition index statistic for the independent variables was 5.319, well below the danger point of 15) and partial regression plots showed the residuals randomly scattered above and below 0 on both axes.
Summary of Simultaneous Regression Analysis for Variables Predicting the Usage of Hegemony Online (N = 126)
Variable B SE B Beta
(Constant) 5.368 1.369
Alienation score -2.764 1.184 -0.205*
E-mail hours per week -3.365 1.615 -0.197*
Web hours per week 2.909 1.524 0.180#
Note. R2 = .089; adjusted R2 = .066.
*p < .05
#p < .06
In essence, this model tells us that people who make the most use of hegemonic Web pages are those who are least alienated from society, spend relatively little time on e-mail, and spend a lot of time on the Web. The model explains only 6.6 percent of the variance in the hegemony scores among the subjects in this sample, but that explains more than any other combination of variables in this study--or any other study published to date. Moreover, it provides a good foundation for further research on hegemony online.