Several authors have noted the social inequalities in access to telecommunication and information resources and the resulting unevenness in the ability of social and cultural groups to participate in this revolution. Yet, even if social inequalities were erased, integration of a technology into daily family life is not guaranteed; there is still uncertainty over the extent that communication technologies will become integrated into everyday life in the home. Despite ownership data and optimistic forecasts of continued adoption (i.e., increasing ownership), descriptive data are sparse concerning what actually happens after a family acquires the technology. Does it entertain in the living room, facilitate work in the study, or gather dust in the closest? Current research provides little insight into the usage intensity and usage context of communication technologies. Clearly, a greater understanding is needed of the factors that contribute to the intensity and variety of usage of communication technologies.

While research on usage profiles and technology clusters is important to managers, researchers and regulators, a problem exists in the current innovation literature in that no research is available which can be generalized across a product or technology segment, with any degree of confidence. This paper explores the usage of household communication technologies and the theoretical relationship of usage intensity, usage context, and occupational prestige. As multiple technologies are available to satisfy our individual needs, our choice of the 'best' technology(ies) is based on our experience with the medium itself, the social circumstances surrounding its use, and the social standing of the consumer. This study provides greater understanding of the usage profiles for communication technologies and lays the foundation for further theoretical and empirical work.

The merger of the computer and the telephone is also prompting the emergence of a communications revolution. Traditionally, telephones have served as the medium that links individuals in remote locations. Today, information technologies are changing the way we work, play, and think. Cellular phones and pagers allow us to stay in touch without being tethered to a fixed location Ðoffice or home. Electronic mail is replacing the pen and paper and studies have found that heavy e-mail users are better informed than their colleagues are (Baig 1994). Computers and faxes make it possible for telecommuters to avoid the daily rigors of traveling to the office. New communications and computing technologies allow consumers throughout the world to manipulate and distribute information as never before through the ownership and usage of the newer communication technologies. Furthermore, the explosion in communication capacity and networking technologies (e.g., fiber optics, wireless transmission, satellites, and Infobahns) will put amazing amounts of low-cost, high-speed communications at our disposal (Verity 1994).

As individuals we are confronted with new communication devices almost every day. Arnold (1997) observes that we are in the 'midst of the rapid proliferation of goods and services throughout certain segments of society that facilitate the flow of information and improve interpersonal communication.'

At the same time, some are not participating in the information revolution. To some individuals, the idea of interfacing with a digital, virtual world is daunting. Others have noted that social inequalities may limit those who will participate in the information revolution. Several authors (e.g., Schuler 1994, Nowotny 1982, Calhoun 1986, Robins and Hepworth 1988, Murdock and Golding 1989, Graham and Marvin 1996) have noted the social inequalities in access to telecommunication and information resources, and the resulting unevenness in the ability of social and cultural groups to participate in this revolution.

The world of information technology is a world made for the very fortunate few . . . who sit with their little computers and their telephones and deal with information all over the world. . . . And then they'll be the rest, who don't have access to this technology, who don't know how to use it, who don't know how to make products out of it. (Graham and Marvin 1996, 189)

Ultimately, it will be our collective acceptance of these new technologies and products that will determine the breadth, depth, and tempo of the information revolution and technology diffusion. Yet, even as these products diffuse throughout society, many will be excluded from participation in the information and communication revolutions.

Despite the massive research effort represented by the existing body of diffusion research, many observers are critical of the methodology, assumptions and concepts associated with such studies (c.f., Hirschman 1979; Mahajan, Muller and Bass 1990). One specific criticism is that some authors fail to make a distinction between trial (i.e., single use of the product) and adoption (i.e., multiple uses of the product) in discussing innovative behavior. Those studies that did consider usage levels were generally restricted to low technology products. Thus, additional research that explicitly considers the usage level and usage purpose (e.g., business, social, transaction, and entertainment) for higher technology products is warranted.

The questionnaire was mailed to respondents by a commercial market research firm, utilizing its a consumer mail panel. Each subject was asked to complete the questionnaire and indicate ownership, utilization (i.e., how often he or she had used each of the technologies), and the purpose of their usage for 22 communication technologies. Each adult was asked to also provide demographic characteristics.

After the research firm had processed and tabulated all closed-ended questions, the questionnaires and data were returned to the researchers. Open-ended questions were coded by the researchers. Non-response, while generally a significant problem with mail surveys, did not pose significant threats to validity as the response rates for this consumer mail panels was very high (i.e., over 70%).

Figure 1 illustrates the technologies selected for use in this study. A mixture of both products and services were selected for inclusion in the list of communication goods. However, the products/services were selected and presented in a manner so as to minimize ambiguity.

For the purposes of this study, product ownership is evident if the respondent indicates that he/she either owns or leases a communication good, or subscribes to a communication service. However, in order to isolate the variance that could be caused if the communication good is provided by an employer or some other arrangement (i.e., products not evaluated and selected by the respondent), these goods are excluded from analysis. The exclusion of employer-provided products acts to prevent any possible inflation of the expected relationship between occupational prestige and ownership.

In spite of its superiority as a measure, Ram and Jung (1990) noted that, historically, product usage has been a relatively neglected construct in marketing, as both its conceptualization and measurement have generally been inadequate. The authors defined product usage as consisting of usage frequency and usage variety. Usage frequency, according to Ram and Jung (1990, 68), refers to "how often the product is used (usage time) regardless of the different applications for which the product is used."

Usage variety, on the other hand, refers to the "different applications for which a product is used or the different situations in which a product is used, regardless of how frequently it is used" (Ram and Jung 1990, p. 68). Variations of the usage variety conceptualization have appeared in Gatignon and Robertson's (1985) notion of product width and depth, and Zaichowsky's (1985) conceptualization of depth and breadth of product usage.

Ram and Jung's (1990) usage frequency scale has proved to be highly reliable and of demonstrable convergent validity. This study incorporates Ram and Jung's (1990) scale for measuring usage frequency, and modifies their concept of usage variety to better fit the nature of communication technology. Figure 2 below illustrates the scale used.

The key to our understanding of consumer ownership and usage of communication technologies is identifying the motives or needs that elicit behavior (Williams and Dordick 1985). This study is built on the belief that interactive communication technologies facilitate interpersonal communication for business, social and transactional purposes and that this communication is integral to a consumer's social participation and the maintenance of his/her social network and lifestyle. Furthermore, the degree to which the business and social domains coalesce is positively related to a consumer's social standing (Fisher and Boughton 1991, Gilbert and Kahl 1982). Therefore, the purpose of the interpersonal communication and the consumer's social standing will affect the acceptance and usage of interactive communication products.

Otto (1975) proposed that occupational prestige is the single, best indicator of a family's social position in modern industrialized societies because of the inadequacy of composite measures of class or status, and because of the extraordinary reliability of the occupational prestige construct. Given the limitations of the composite approach to measurement, many scholars have reached a consensus that occupational prestige is the preferred status indicator and the best single indicator of a family's social status.

Occupation prestige has been shown to be relatively robust construct. The evaluation of occupations has proved to be largely invariant under widely varying methods of measurements (Otto 1975) and occupational prestige has been shown to be a stable phenomenon (Dunkerly l975). Dunkerly (1975) noted that while there had been large changes in both the education and income associated with different occupations between 1925 and 1963, the prestige attached to those occupations has remained virtually unchanged.

Occupational prestige is most simply defined as societal judgment as to the worth of individuals or groups (Montagna 1977, 30) and, as such, refers to differential societal evaluations of occupations according to their social rank or standing. The most common method of measuring occupational prestige in the United States is the reputational approach, where occupations are rank ordered on the basis of their 'general standing' (or social standing, social status, desirability, et cetera) in society. For example, Nakao and Treas (1994) instructed respondents to arrange cards identified with occupational titles on a "ladder" according to what they thought about the "social standing" of the particular occupation. We employed a measurement approach based on Nakao and Treas' occupation category scheme; this approach is discussed in greater detail later.

Coleman (1983), in agreement with Schaninger (1981) and others (e.g., Reynolds 1965), suggested that researchers should consider both social class and income when trying to understand the consumer. The thesis developed in this paper relies on the argument that consumption of these goods, while certainly affected by income, is not driven by income in the way it is by the family's social standing.

Consistent with Otto's (1975) definition of family income as one of the "conceptually unambiguous" constructs, total family income is gathered via an item with a categorical response format.

Age is an important explanatory variable in any study concerning the investigation of consumption patterns (Arnold 1997) as studies have found an important relationship between age and usage of high technology goods (c.f., Rosen and Weil 1995; Zeithaml and Gilly 1987).

Age has also proven itself a valuable indicator of one's location or position within the family life cycle. A common approach to measuring a family's membership in a life cycle category is based on Wells and Gubar's (1966) suggestion that the family's life cycle position be ascertained from the age of the male head of the household (Wilkes 1995). Gilly and Enis (1982) and other researchers have opted to use the female's age in defining the life cycle categories because they believed that the presence or absence of children was a critical determinant of life cycle stage and that age 35 was an important milestone for women regarding childbearing.

Since we have evidence that both the husband and wife's ages are important and since the unit of analysis for this study is the family, age is gathered for both spouses and averaged before inclusion in the research model.

When measuring occupational prestige in family research (i.e., considering the family as the unit of analysis), an analytical problem arises regarding how to incorporate the respondent's working spouse's prestige score. Traditional sociological analysis has relied on using the status of the male head of the household as a proxy for the family's social status. This approach has assumed that the husband was the primary breadwinner, that the wife works to provide luxuries for the family, and the family is unaffected by the status normally associated with the wife's occupational position (Philliber and Hiller 1978). However, many contemporary families are chiefly supported by the wife's occupation. Given the changing makeup of the workplace, there definitely appears to be a need to incorporate the prestige of the working wife. Coleman (1983) urged researchers to gather information regarding the wife's occupation and educational attainment, as these variables are indirect measures of the family's social horizons. Philliber and Hiller (1978) suggested that the wife's occupational attainment has significant implications for the social status of the family as a whole and have challenged the traditional assumption that both husband and wife share the same social status. Others (Baxter 1994, Davis and Robinson 1988) have reached similar conclusions.

Otto (1975) addressed the issue of dual income households by noting that the Census "demonstrates sensitivity" to this contemporary structural reality by assigning socioeconomic status to a family on the basis of the family's "chief" earner rather than on the basis of the male earner. Consistent with the Census approach of assigning a family's socioeconomic status, the approach adopted in this study will be to gather information regarding the "chief" earner and define the family's social status based on the occupational prestige score of the 'chief' earner.

After the data are collected, the responses to each of the two occupation items was inspected and an occupational prestige score will be assigned based on Nakao and Treas' (1994) occupational prestige inventory for 889 individual occupational titles. Responses to the open-ended occupation item will be matched with an appropriate occupational title in Nakao and Treas' (1994) inventory and the associated occupational prestige score will be assigned to that case. This approach and additional formalized coding rules for converting occupational titles to occupational prestige scores are outlined by Treiman (1977, 214-222). These rules provide a rigorous framework for coding and suggest remedies for cases that are difficult to code.

For Hypothesis 1, a count will be taken of the number of interactive communication devices used for business purposes by a respondent. The count will be weighted according to frequency of usage whereby 'seldom' use, 'occasionally' use, and 'frequently' use are given weights of 2, 3, and 4, respectively. This weighted value of usage frequency will be the dependent variable. The independent variables will be occupational prestige, income, education and age.

In order to eliminate the problem of dealing with different units of measurement and to allow us to later assess the relative impact on the criterion (dependent) variable, the predictor (independent) variables will be transformed or standardized into new measurement variables with a mean of 0 and a standard deviation of 1. According to our thesis, we should see (1) positive significance for occupational prestige after removal of income, education and age effects and (2) the beta coefficient for occupational prestige should be great than that for income, education, or age.

Hypothesis 2 suggests that occupational prestige will have a greater influence on the number of interactive communication devices used for social purposes by a respondent than income, education, or age. The approach here will be similar to that of Hypothesis 2 Ða weighted value of usage frequency will be calculated for interactive communication devices used for social purposes by a respondent and will serve as the dependent variable. The independent variables will be occupational prestige, income, education and age. Support of the hypothesis is provided if there is (1) a positive significance for occupational prestige, after removal of income, education and age effects and (2) the beta coefficient for occupational prestige greater than that for income, education, or age.

Likewise, Hypothesis 3 suggests that occupational prestige will have a greater influence on the number of interactive communication devices used for service transactions by a respondent than income, education, or age. Again, the approach here will be similar to that of Hypothesis 1 Ða weighted value of usage frequency will be calculated for the interactive communication devices used for service transactions by a respondent. This weighted usage value will be the dependent variable. The independent variables will be occupational prestige, income, education and age. Support of the hypothesis is provided if there is (1) a positive significance for occupational prestige, after removal of income, education and age effects and (2) the beta coefficient for occupational prestige is greater than that for income, education, or age.

H1: A household's social standing will be a superior predictor for usage of technologies used for business purposes than will be age, education or income.

For Hypothesis 1, the aggregate usage for business purposes was employed as the continuous dependent variable. The standardized values for occupation prestige, age, education, and income were the continuous independent variables. In order to eliminate the problem of dealing with different units of measurement and thus reflect the relative impact on the criterion (dependent) variable, the predictor (independent) variables were transformed into standardized variables with a mean of 0 and a standard deviation of 1.

Table 1 shows that occupational prestige is significantly related to the aggregate business usage of interactive technologies after controlling for age, education and income effects. The standardized beta value of 0.2005 for prestige confirms the positive relationship between prestige and overall usage. The t-statistic for prestige (t=6.199) indicates that the relationship is significant at the p<.0001 level. The adjusted R ² (e.g., adjusted for the number of predictor variables) is 0.2784, indicating that the four predictor variables explain almost 28 percent of the variation in overall usage. We also note that the F-statistic is significant (F=82.110, p<.0001), indicating that the amount of variation explained by the predictor variables was not 'by chance.'

However, Table 1 also shows that the t-statistics for age and income (t= -5.119 and 9.915, respectively) are significant at the p<.0001 level and the t-statistic for education (t= 3.110) is significant at the P<.002 level. Furthermore the standardized beta value of 0.3313 for income is greater than beta value for prestige. As the beta coefficient for income is greater than that for occupational prestige, Hypothesis 1 is not supported.

However, it should be noted that prestige is a superior predictor than either age or education. While previous research has shown that age and education are important factors in product consumption Ðand are significantly related to the aggregate business usage of interactive technologies in this study Ðboth age and education are inferior to occupational prestige as predictors of aggregate business usage. We further observe that, among adults, age is inversely related to the frequency of business usage of telecommunication products.

H2: A household's social standing will be a superior predictor for usage of technologies used for social purposes than will be age, education or income.

With references to Table 2, the t-statistic for occupational prestige (t=6.539) indicates that the relationship to aggregate social usage of interactive technologies after controlling for age, education and income effects is significant at the P<.0001 level. The standardized beta value of 0.2144 for prestige confirms the positive relationship between prestige and overall usage. The adjusted R ² is 0.2638, indicating that the four predictor variables explain over 26 percent of the variation in overall usage. Furthermore the F-statistic (F=76.331) is significant, indicating that the amount of variation explained by the predictor variables was not very likely to occur 'by chance.'

However, Table 2 again shows that the t-statistics for age and income (t= -7.316 and 9.706, respectively) are also significant at the p<.0001 level. Finally, the standardized beta value for income (0.3283) is greater than the beta value for prestige. As the beta coefficient for income is greater than that for occupational prestige, Hypotheses 2 is not supported.

We further note that the standardized beta coefficient values for prestige and age (beta= 0.210 and -0.2182, respectively) are almost identical thereby indicating that the magnitude of their respective effects on the dependent variable are similar, although age is inversely related to the social usage frequency of telecommunication products. Furthermore, in a comparison to business usage (see Table 1), we find that the contribution of age as a predictor of social usage has increased (beta of -0.2182 versus -0.1511) while the predictive power of income and prestige are relatively similar for business and social usage. However, in contrast to its significance as a predictor of business usage, education is not a significant predictor of social usage as shown by its t-statistic (t= 0.947).

Hypothesis 3

Hypothesis 3 proposes the existence of a positive and significant relationship between prestige and the service usage level for interactive technology devices after removal of income, education and age effects. Hypothesis 3 also proposes that prestige is a relatively superior predictor of service usage that is income, age, or education.

H3: A household's social standing will be a superior predictor for usage of technologies used for service purposes than will be age, education or income.

As shown in Table 3, the occupational prestige construct is significantly related (t= 5.174, p<.0001) to the aggregate service usage of interactive technologies after controlling for age, education and income effects. The standardized beta value of 0.1760 for prestige confirms the positive relationship between prestige and overall usage. The adjusted R ² (0.2074) indicates that the four predictor variables explain almost 21 percent of the variation in overall usage. Finally, the significant F-statistic (F=56.022), suggests that these values were not derived 'by chance.'

However, Table 3 also reveals that age and income are also significant at the p<.001 level

(t= -3.837 and 9.418, respectively). Again, the standardized beta value for income (0.3310)is greater than beta value for prestige.

Given that the beta coefficient for occupational prestige is less than that for income, hypothesis 3 is not supported with respect to a household's social standing being a superior predictor for service usage than is income. However, support is provided with respect to a household's social standing being a superior predictor for service usage than is age or education.

In a comparison of service usage to business and social usage (see Tables 1 and 2), we find that the contribution of prestige and age as predictors of usage has decreased while the predictive power of income is relatively similar for all three usage contexts. In contrast to its significance as a predictor of business usage, education is not a significant predictor of service usage as shown by its t-statistic (t= 1.006).

H4: A household's social standing will be negatively related to the aggregate usage of interactive communication technologies used for entertainment purposes.

The aggregate usage for entertainment purposes by a respondent is the continuous dependent variable and the continuous independent variables are the standardized values for occupation prestige, age, education, and income. Again, multiple regression analysis is employed to analyze the relationship between a single metric dependent variable and several metric independent variables.

According to our thesis, the regression coefficient for occupational prestige should be significant as indicated by the t-value and the regression coefficient should be negative, after removal of education and age effects.

Table 4 shows that occupational prestige is significantly related to aggregate usage of interactive technologies for entertainment after controlling for age, education and income effects (t= 4.735, p<.001). However, the standardized beta coefficient of 0.1633 for prestige confirms a positive relationship between prestige and overall usage. The tests for the regression model revealed no serious violations of the underlying assumptions of the regression model. As the regression coefficient for occupational prestige is positive, hypothesis 4 is not supported.

However, the standardized beta coefficients for prestige and income (beta = 0.1633 and 0.2670, respectively) are smaller values than for any other usage context (i.e., business, social, or service). Thus, the contribution of prestige as a predictor of service or entertainment usage is less than that for business or social usage.

In the late 1960s, the usefulness of social class as a significant stratification variable was questioned. Critics challenged the value of social class because demographic trends Ðchanges in income, education, leisure time, population shift, and migration to suburbia Ðcut across traditional class boundaries leading some to conclude that social class differences in shopping behavior were diminishing (cf. Rich and Jain, 1968). Other critics contended that the absence of more 'objective status indicators around which classes may crystallize makes consensus regarding class identifications ambiguous if not impossible (Otto, 1975). Finally, because some indicators of social class have failed to demonstrate the existence of groupings or because the standards of living were becoming similar for the upper working and lower middle classes, many critics have argued that the U.S. was becoming 'classless' (cf. Faris 1960, Wilensky 1966).

More recently, proponents have contended that social class is a valuable segmentation base because it captures lifestyle differences that income ignores. However, they admitted that the relative superiority of a given segmentation variable will vary among product classes.

This research builds on the foundation provided by Hisrich and Peters (1974), Schaninger (1981), Arnold (1997) and others. Consistent with Arnold's (1997) findings, we have demonstrated that occupational prestige, as a measure of social standing, is very significant and relevant in explaining the depth of consumption of interactive technologies and we provided support of a linear relationship between usage and socioeconomic variables. However, we found that income also is a significant factor in explaining depth of usage of interactive technologies.

Schaninger (1981) generalized that social class is a better segmentation basis for low-cost goods that reflect an underlying lifestyle, value, or homemaker role differences not captured by incomes. Schaninger's research, while providing a valuable contribution to the social class versus income controversy, was generally atheoretical in nature. The research methodology was one of selecting an assortment of goods (for example, appliances, soft drinks, mixers, and alcoholic beverages), finding a relationship between these goods and the socioeconomic variables, and then explaining the relationships ex post. Arnold (1997) improved on Schaninger's contribution by preceding the data analysis with a carefully articulated theoretical basis and focusing his investigation on a single product class. However, Arnold implied that the superiority of social class as a determinant of usage could be generalized to all interactive technologies for business and social purposes, although prestige was significant at p<.10 for only six of the fourteen technologies analyzed.

Contrary to Arnold's (1997) and Schaninger's (1981) conclusions, we found that occupational prestige, as a measure of a family's social position, was inferior to income for a class of products that are analyzed as a single cluster of products. The superiority of income was supported for all usage purposes, namely business usage, social usage, transactional or service usage, and entertainment usage.

An analysis of the relationship between social class, income and consumption on a disaggregated basis using ordinary least square regression was completed. It is interesting to observe that occupation prestige is a relatively superior predictor of usage for some technologies. We also observe that occupational prestige is a significant predictor (p<.10) of business usage for five of the technologies and of social usage for 2 technologies, but is significant for only 1 technology for service usage and is not significant for any technologies with regards to entertainment usage. This relationship is summarized in Table 5. From this observation one inference consistent with the data is that occupational prestige is a stronger explanatory factor for communication technology usage only when used for purposes with a social dimension (e.g., business or social purposes) as compared to purposes that are noninteractive in nature (e.g., entertainment). These results based on a disaggregated view of the product class may be interpreted as somewhat consistent with Schaninger's and Arnold' generalization regarding social class effects. This interpretation between an aggregated and disaggregated viewpoint will be discussed in further detail later.

b. p<.10

Consistent with the previous research of Arnold (1997), the product cluster examined was defined based on interactivity. However, contrary to Arnold's conclusion based on an analysis of each product individually, our aggregated analysis found social prestige to be inferior to income in its influence of overall usage and usage for four specific purposes of interactive technologies. Based on this discrepancy, we can only assume that viewing all interactive technologies as a single cluster is an oversimplification. Rather, we believe that the cluster must be divided into smaller baskets based on technology type (e.g., voice/data technologies) and product scarcity in order to understand better the product's social meaning and the social prestige of its respective users.

The analysis results provide initial support for our belief that the cluster boundaries must be drawn based on multiple technology attributes (e.g., interactive technology, simulates face-to-face interaction, voice technology, etc.). Furthermore, analysis suggests that the product cluster boundaries are dependent on product attributes and the social context in which the product was used. This later point further testifies to the notion advanced by Fisher and Boughton (1991) and Arnold (1997) that products are often rich with social meaning.

Clearly, the traditional perspective on diffusion theory must expand to incorporate not only the adoption of an innovation, but the breadth and depth of its usage. Furthermore, the research presented here proposes that diffusion theory must also focus on the social system in which the product is used. Finally, this research provides a basis for defining the usage of interactive communication technologies and for assembling these products into unique packages.

Furthermore, while occupational prestige is positively and significantly related to the usage of an aggregated cluster of interactive technologies for all four purposes, income had a greater influence on usage for all four purposes studied. Thus, while we agree with those who have suggested that interactive data technologies are increasingly utilized for emotional and interpersonal communication, this research suggests that income is required to reduce the economic constraints on access to interactive technologies.

Hypotheses 1 through 4 analyzed the aggregated usage of interactive technologies. A contrasting of these results against those of Arnold (1997), suggest that a single cluster of 14 interactive technologies may be too broad of product grouping. Division of the 14 technologies into smaller baskets of technologies, for example, voice or data technologies, may uncover some baskets which are influenced predominantly by social class and other baskets which are influenced predominately by income.

Upon examination of a disaggregated analysis of the 14 technologies and the relationship between social class, income and consumption, we note a significant relationship between occupational prestige and business or social usage for individual voice technologies such as the cordless phone cellular phone, voice mail service, and answering machine. Such findings are somewhat consistent with Arnold's proposition that voice technologies used for business and social purposes reinforce the social position of the higher-prestige users.

We also see from the results that prestige is significantly related to business usage for a small set of data devices, namely the PC modem, on-line service, and personal computer. In a business context, these devices provide emotional, immediate, and interpersonal non-oral communication and can dramatically improve access to and control over information while allowing the user to control the amount of time allocated to the communication. It is of further interest that the usage of these data devices for each purpose is best explained by different social-economic variables, for example personal computer usage for social purposes is best explained by age and education, income is the best descriptor of usage of personal computers and on-line services for service transaction, and only age is a significant descriptor of entertainment usage. This finding appears to support Fisher and Boughton's (1991) position that those of a high social position may use communication technologies to enhance their social position, but goes beyond that position and advances the theme that descriptors for usage of a single device must consider the usage purpose.

One of the fundamental issues facing sociologists and regulators, which is likely to have major impacts on both economic and political systems, is the problem of unequal access to information and information systems. There has been considerable concern that the monetary or non-monetary costs facing potential users may wrongly influence the distribution of power within society.

Some (cf. Blake 1978, Schiller 1985) have argued that the current class-based social system is likely to replicate itself as access to information is controlled by the ability to pay. While there exists the possibility that access to information and its corresponding benefit, education, can promote both aspiration and achievement, some regulators are concerned with the possibility that the gaps between the information rich and information poor is apt to increase.

Given the concern of an increasing information gap and the view that access to information is a basic necessity, some regulators are considering the possibility of state subsidies to promote access to information. Such subsidies are already available for basic telephone access under 'life-line' and similar programs.

The present research, consistent with previous research, recommends that regulators should look beyond technology ownership to technology usage. In doing this, regulators will observe that multiple variables, not just family income, determine communication technology usage. Furthermore, regulators should be concerned with the usage context they wish to promote. For example, we note that 26% to 27% of the variance in usage of communication technology devices for business and social purposes is explained by occupational prestige, age and income, while only 18% of the usage variance is explained by these three variables for entertainment usage.

The range of communication technological products and services studied here is narrow with only 14 products and services tested. Future research incorporating a broader array of products and services may provide a more complete understanding of the boundaries around any given cluster and its associated technology attributes, social-economic consumer variables, and usage purpose which determine its usage intensity.

Issues surrounding the measurement of social class remain unresolved. The plethora of social class indicators leaves the individual researcher to choose the best measure of social class. While occupation and occupational prestige are deemed among the most important indicators of social class, 'The distilling of a few relatively homogeneous occupational groups from the many thousands of distinct jobs held in the labor force is a difficult task that probably has no fully satisfactory solution' (Hodge and Siegel, 1968). Thus, any research involving social position and the construct of social class suffers from validity and measurement problems. Nonetheless, occupation prestige has been proven to be a highly reliable construct and achieves appreciable face validity within the particular context for this research.

Finally, the unit of analysis may be considered a limitation. For this research, participants were limited to the traditional married-couple households. This was done to effectively delimit household consumption of technologies, assign family prestige scores, and therein reduce variance associated with household type. However, we acknowledge that a large percentage of today's households are comprised of single adult households. Future research should account for any differences that may exist within other household structures.

First, future research should continue to emphasize the frequency of use of many products and services. It is also important that future research explore post-adoption usage patterns. Preferably, studies that investigate patterns of usage should be longitudinal in nature, such that a history of product usage over time may be collected.

Second, research should document the usage context or purpose. This research demonstrated that the usage purpose affected both the usage intensity across products and the significance of certain segmentation variables as correlates of buyer behavior. It is important that future research compose other usage contexts and more potential predictor variables so as to assess the best correlate or correlates for a particular product and context. More predictor variables would allow a more precise understanding of the relationship between the predictor variables and the consumption of these technologies.

Third, replication of this study in another society or country would provide insight into the influence of culture and economic development as moderating variables on the usage of telecommunication products.

Finally, several methodological opportunities exist. This study, employing regression analysis, produced alternate conclusions to those of Arnold (1997) and Schaninger (1981) who used analysis of variance techniques. Such findings raise the question of multiple dependence relationships and suggest the use of a multivariate technique, such as structural equation modeling, that combines aspects of regression analysis and factor analysis to estimate a series of interrelated dependence relationships simultaneously. Structural equation modeling provides for an estimation of multiple and interrelated dependence relationships and the ability to represent unobserved concepts. Leveraging the results of this and other research efforts, a theoretically based model that acknowledges the interrelated dependence relationships could be developed and its causal relationships assessed using structural equation modeling. These steps would be a logical progression toward a fuller understanding of the usage intensity and diversity of communication technologies.

[fnA]

References

Anderson, Robert L. and David J. Ortinau. 1988. Exploring Consumers' Postadoption Attitudes and Use Behaviors in Monitoring the Diffusion of a Technology-Based Discontinuous Innovation. Journal of Business Research 17:283-298.

Arber, Sara and Jay Ginn. 1992. Research Note Class and Caring: A forgotten Dimension. Sociology 26 (November):619-634.

Arndt, Johan. 1967. Role of Product-Related Conversation in the Diffusion of a New Product. Journal of Marketing Research 4 (August):291-295.

Arnold, Mark J. 1997. Consumer Acceptance of Interactive Communication Technology: An Empirical Investigation into the Significance of Social Stratification. (Ph.D. dissertation, Saint Louis University)

Baig, Edward C. 1994. Ready, Set--Go On-Line, Business Week, May 18:124-133.

Bass, Frank M. 1969. A New Product Growth Model for Consumer Durables. Management Science 15 (January):215-227.

Blake, F. M. 1978. Public Access to Information in the Post-Industrial Society. In The Information Society: Issues and Answers, ed. E. J. Josey. Phoenix, AZ: Oryx Press.

Calhoun, C. 1986. Computer Technology, Large-Scale Social Integration and the Local Community. Urban Affairs Quarterly 22 (2):329-349.

Caron, Andre H., Luc Giroux, and Sylvie Douzou. 1989. Uses and Impacts of Home Computers in Canada: A Process of Reappropriation. In Media Use in the Information Age, eds. Jerry L. Salvaggio and Jennings Bryant, 147-162. Hillsdale, NJ: Lawrence Erlbaum Associates.

Chafee, S., J. McLeod, and C. Atkin. 1971. Parental Influences on Adolescent Media Use. American Behavioral Scientist 14:323-340.

Churchill, Gilbert A. 1991. Marketing Research: Methodological Foundations. 5th ed. Orlando: Harcourt, Bruce, Jovanovich Publishers.

Clarke, Alfred. 1956. The Use of Leisure and Its Relation to Levels of Occupational Prestige. American Sociological Review 36:301-07

Coleman, James S., Elihu Katz, and Herbert Menzel. 1966. Medical Innovation: A Diffusion Study. Indianapolis: The Bobbs-Merrill Company, Inc.

Coleman, R. P. 1960. The Significance of Social Stratification in Selling. In Marketing: A Maturing Discipline, ed. Martin L. Bell, 171-184. Chicago: American Marketing Association.

David, Fred R. 1995. Strategic Management. Englewood Cliffs, NJ: Prentice Hall.

Dickerson, Mary Dee and James W. Gentry. 1983. Characteristics of Adopters and Non-Adopters of Home Computers. Journal of Consumer Research 10 (September):225-235.

Dizard, Wilson. 1982. The Coming Information Age. New York: Annenberg.

Dominguez, Louis V. and Albert L. Page. 1981. Stratification in Consumer Behavior Research: A Re-Examination. Journal of the Academy of the Marketing Science 9 (Summer):250-271.

Douglas, Mary T. and Baron Isherwood. 1970. The World of Goods. New York: Basic Books.

Dozier, D., and R. Rice. 1984. Rival Theories of Electronic Newsreading. In The New Media, ed. R. Rice, 103-128. Beverly Hills, CA: Sage.

Dunkerly, David. 1975. Occupations and Society. Boston, MA: Routledge and Kegan Paul.

Easingwood, Christopher J. and Simon O. Lunn. 1992. Diffusion Paths in a High-Tech Environment. R&D Management 22:69-81.

Ellen, Pam Scholder, William O. Bearden, and Subhash Sharma. 1991. Resistance to Technological Innovations: An Examination of the Role of Self-Efficacy and Performance Satisfaction. Journal of Academy of Marketing Science 19 (Fall):297-307.

Elliott, Robert K. 1994. Confronting the Future: Choices for the Attest Function. Accounting Horizons 8 (3):106-124.

Featherman, David L. and Robert M. Hauser. 1976. Prestige or Socioeconomic Scales in the Study of Occupational Achievement? Sociological Methods and Research 4 (May):403-422.

Fisher, James E. 1987. Social Class and Consumer Behavior: The Relevance of Class and Status. In Advances in Consumer Research, eds. Melanie Wallendorf and Paul Anderson, 492-496. Provo, UT: Association for Consumer Research.

________, and Paul D. Boughton. 1991. Information, Technology and Social Class. In Proceedings of the Annual Meeting of the Southern Marketing Association, ed. Robert L. King, 11-15. Richmond, VA: University of Richmond

Gatignon, Hubert. 1989. Technology Diffusion: An Empirical Test of Competitive Effects. Journal of Marketing 53:35-49.

________ and Thomas S. Robertson. 1985. A Propositional Inventory for New Diffusion Research. Journal of Consumer Research 11 (March):849-867.

Gilbert, Dennis and Joseph A. Kahl. 1982. The American Class Structure: A New Synthesis. Homewood, IL: Dorsey Press.

Gilly, Mary C. and Ben M. Enis. 1982. Recycling and the Family Life Cycle: A Proposal for Redefinition. In Advances in Consumer Research, Vol. 9, ed. Andrew A. Mitchell, 271-276. Ann Arbor, MI: Association for Consumer Behavior.

Goldthorpe, John H. and Keith Hope. 1973. Occupational Grading and Occupational Prestige. In Social Stratification and Career Mobility, eds. Walter Muller and Karl Ulrich Mayer, 2-73. Paris, France: Mouton and Encole Pratique.

Grasmick, Harold G. 1976. The Occupational Prestige Structure: A Multidimensional Scaling Approach. The Sociological Quarterly 17 (Winter):90-108.

Green, Paul E. 1978. Analyzing Multivariate Data. Hinsdale, IL: Dryden Press.

Hair, Joseph F., Rolph E. Anderson, Ronald L. Tatham, and William C. Black. 1995. Multivariate Data Analysis. 3d ed. New York, NY: Macmillan Publishing Company.

Heeter, Carrie. 1989. Implications of New Interactive Technologies for Conceptualizing Communication. In Media Use in the Information Age, eds. Jerry L. Salvaggio and Jennings Bryant, 217-235. Hillsdale, NJ: Lawrence Erlbaum Associates.

Hisrich, Robert D. and Michael P. Peters. 1974. Selecting the Superior Segmentation Correlate. Journal of Marketing 38 (July):60-63.

Hirschman, Elizabeth C. 1979. An Alternate Diffusion Paradigm Especially Relevant for Innovations Possessing High School Visibility. In Developments in Marketing Science: Proceedings of the Third Annual Conference of the Academy of Marketing Science, ed. Robert L. King, 2, 300-304. Richmond VA: Academy of Marketing Science.

________. 1981. Symbolism and Technology as Sources for the Generation of Innovations. In Advances in Consumer Research, ed. Andrew Mitchell, 9, 537-541. St. Louis, MO: Association for Consumer Research.

Hodge, R. W. and P. M. Siegel. 1968. The Measurement of Social Class. In International Encyclopedia of the Social Sciences, ed. David L. Sills. New York: The Macmillan Company.

Hodge, R. W., P. M. Siegel, and P. H. Rossi. 1966. A Comparative Study of Occupational Prestige. In Class, Status, and Power, ed. R. Bendix and S. M. Lipset, 309-321. New York: Free Press.

Houston, Douglas A. 1983. Implicit Discount Rates and the Purchase of Untried, Energy-Saving Durable Goods. Journal of Consumer Research 10 (September):236-246

Kraykamp, G., L. B. Van-Snippenburg, W. C. Ultee. 1989. Occupational Class and Occupational Status as an Explanation of Income: Attitudes Toward Inequality, Political Confidence, and Political Apathy. Mens-en Maatschappi, 64 (January): 42-65.

Levy, Sidney. 1966. Social Class and Consumer Behavior. In On Knowing the Consumer, ed. J. W. Newman, 146-160. New York, NY: John Wiley.

Lin, Nan, Walter M. Ensel and John C. Vaughn. 1981. Social Resources and Strength of Ties: Structural Factors in Occupational Status Attainment. American Sociological Review 46 (August):393-405.

Mahajan, Vijay, Eitan Muller, and Fram M. Bass. 1990. New Product Diffusion Models in Marketing: A Review and Directions for Research. Journal of Marketing 54 (January):1-26.

Martin, James. 1978. The Wired Society. Englewood Cliffs, NJ: Prentice Hall.

Martineau, Pierre. 1958. Social Classes and Spending Behavior. Journal of Marketing 23 (October):121-130.

Marting, Elizabeth. 1964. New Products, New Profits. New York, American Management Association.

McQuarrie, Edward F. 1989. The Impact of a Discontinuous Innovation: Outcomes Experienced by Owners of Home Computers. Computers in Human Behavior 5:227-240.

_______. and Kichiro Iwamoto. 1990. Public Opinion Toward Computers as a Function of Exposure. Social Science Computer Review 9 (2):221-233.

Mirabito, Michael M. and Barbara L. Morgenstern. 1990. The New Communication Technologies. Boston, MA: Focal Press.

Montagna, Paul D. 1977. Occupations and Society: Toward a Sociology of the Labor Market. New York: John Wiley and Sons.

Mundorf, Norbert, Nikhilesh Dholakia, Stuart Westin, and Winifred Brownell. 1992. Reevaluating Gender Differences in New Communication Technologies. Communication Research Reports 9 (Summer):171-181.

Murdock, G. and P. Golding. 1989. Information Poverty and Political Inequality: Citizenship in an Age of Privatized Communication. Journal of Communication 39 (3):180-195.

Myers, James H. and J. Guttman. 1974. Lifestyle: The Essence of Social Class. In Lifestyles and Psychographics, ed. William D. Wells, Chapter 10. Chicago: American Marketing Association.

________, and John. F. Mount. 1973. More on social Class vs. Income as correlates of Buying Behavior. Journal of Marketing 37 (April):71-73.

________, Roger R. Stanton and Arne F. Haug. 1971. Correlates of Buying Behavior: Social Class vs. Income. Journal of Marketing 35 (October):8-15.

Nakao, Keiko, and Judith Treas. 1994. Updating Occupational Prestige and Socioeconomic Scores: How the New Measures Measure Up. Sociological Methodology 9:1-72.

Nelson, Debra L. and Marilyn G. Kletke. 1990. Individual Adjustment During Technological Innovation: A Research Framework. Behavior and Information Technology 9(4):257-271.

Nowotny, H. 1982. The Information Society: Its Impact on the Home, Local Community, and Marginal Groups. In Information Society: For Richer, For Poorer, eds. H. Bjorn Anderson, M. Earl, O. Holst and E. Mumford. North Holland: Elsevier.

Ostlund, Lyman E. 1974. Perceived Innovation Attributes and Predictors of Innovativeness. Journal of Consumer Research 1 (June):23-29.

Otto, Luther B. 1975. Class and Status in Family Research. Journal of Marriage and the Family (May):315-332.

Parker, Betty and Teresa Holder. 1996. Information Haves and Have-Nots: The Diffusion of New Information and Communication Technologies. In Proceedings of the 1996 Public Policy and Marketing Conference.

Pelton, Joseph. 1981. Global Talk. Rockville, MD: Sijthoff and Noordhoff.

Philliber, William W. and Dana V. Hiller. 1978. The Implication of Wife's Occupational Attainment for Husband's Class Identification. The Sociological Quarterly 19 (Summer):450-458.

Porat, Marc. 1977. The Information Economy. Washington, DC: US Department of Commerce.

Rachman, David J., Michael H. Mescon, Courtland L. Bovee, and John V. Thill. 1993. Business Today. New York: McGraw-Hill.

Ram, S. and Hyung-Shik Jung. 1990. The Conceptualization and Measurement of Product Usage. Journal of the Academy of Marketing Science 18 (Winter):67-76.

Reiss, Albert J. 1961. Occupations and Social Status. New York, NY: The Free Press.

Reynolds, W. H. 1965. More Sense About Segmentation. Harvard Business Review, 43:107-114.

Rice, Ronald E. and G. Love. 1987. Electronic Emotion: Socioemotional Content in a Computer-Mediated Communication Network. Communication Research, 14:85-108.

_______ and F. Williams. 1984. Theories Old and New: The Study of New Media. In The New Media, ed. R. Rice, 55-80. Beverly Hills: Sage.

Rich, Stuart A. and Subhash C. Jain. 1968. Social Class and Life Cycle as Predictors of Shopping Behavior. Journal of Marketing Research 5 (February):41-49.

Robertson, Thomas S. 1967. Determinants of Innovative Behavior. In Proceedings of the American Marketing Association, ed. R. Moyer, 328-332. Chicago: American Marketing Association.

_______. 1971. Innovative Behavior and Communication. New York: Holt, Rinehart and Winston.

Robins, K. and M. Hepworth. 1988. Electronic Spaces: New Technologies and the Future of Cities. Futures April:155-176.

Rogers, Everett M. 1983. The Diffusion of Innovations. New York, NY: The Free Press.

________. 1985. The Diffusion of Home Computers Among Households in Silicon Valley. Marriage and Family Review 8:89-100.

________ and Marcel M. Albritton. 1995. Interactive Communication Technologies in Business Organizations. Journal of Business Communication 32 (April):177-195.

________ and S. Chaffee. 1983. Communication as an Academic Discipline: A Dialogue. Journal of Communication 33(3):18-30.

________ and Floyd F. Shoemaker. 1971. Communications of Innovations: A Cross-Cultural Approach. New York, NY: The Free Press.

Rosen, Larry D. and Michelle M. Weil. 1995. Adult and Teenage Use of Consumer, Business, and Entertainment Technology: Potholes on the Information Superhighway? The Journal of Consumer Affairs 29 (Summer):55-84.

Rosenberg, Morris. 1953. Perceptual Obstacles to Class Consciousness. Social Forces 31: 22-27.

Schuler, Doug. 1994. Community Networks: Building a New Participatory Medium. Communications of the ACM 37 (January):39-51.

Schiller, H. I. 1985. Privatizing the Public Sector. In Information and Behavior, Volume 1, ed. B. D. Ruben. New Brunswick, NJ: Transaction.

Seron, Carroll and Kerry Ferris. 1995. Negotiating Professionalism(The Gendered Social Capital of Flexible Time. Work and Occupational 22 (February): 22-47.

Siegel, P. M. 1970. The American Occupational Prestige Structure. Unpublished Ph.D. Dissertation. Department of Sociology. University of Chicago.

Sheth, Jagdish N. 1977. Multivariate Methods for Market and Survey Research. Chicago: American Marketing Association.

________. 1981. Psychology of Innovation Resistance: The Less Developed Concept (LDC) in Diffusion Research. In Research in Marketing, ed. J. N. Sheth, 273-282. Greenwich, CT: JAI Press, Inc.

Steinfield, Charles W., William H. Dutton, and Peter Kovaric. 1989. A Framework and Agenda for Research on Computing in the Home. In Media Use in the Information Age, eds. Jerry L. Salvaggio and Jennings Bryant, 61-85. Hillsdale, NJ: Lawrence Erlbaum Associates.

Stoneman, Paul, and Myung-Joong Kwon. 1994. The Diffusion of Multiple Process Technologies. The Economic Journal 104 (March):420-431.

Taylor, James W. 1977. A Striking Characteristic of Innovators. Journal of Marketing Research 14 (February):104-107.

Tornatzky, L. G. and K. J. Klein. 1982. Innovation Characteristics and Innovation Adoption-Implementation: A Meta-Analysis of Findings. IEEE Transactions on Engineering Management 29 (1):28-45.

Treiman, Donald J. 1977. Occupational Prestige in Comparative Perspective. New York, NY: Academic Press.

Verity, John W. 1994. The Information Revolution: How Digital Technology is Changing the Way We Work and Live. Business Week, May 18:10-18.

Wells, William D. and George Gubar. 1966. Life Cycle Concept in Marketing Research. Journal of Marketing Research 3(November):353-363.

White, R. 1983. Mass Communications and Culture: Transition to a new paradigm. Journal of Communication 33(3):279-301.

Wilensky, Harold. 1966. Class, Class Consciousness, and American Workers. In Labor in a Changing America, ed. William Haber, 12-28. New York, NY: Basic Books.

Wilkes, Robert E. 1995. Household Life-Cycle Stage, Transitions, and Product Expenditures. Journal of Consumer Research 22 (June):27-42.

Wilkie, William L. 1986. Consumer Behavior. New York, NY: John Wiley.

Williams, Frederick. 1982. The Communication Revolution. Beverly Hill: Sage Publications.

________. 1987. Technology and Communication Behavior. Belmont, CA: Wadsworth Publishing Company.

________, R. E. Rice and H. S. Dordick. 1985. Behavioral Impacts in the Information Age. In Information and Behavior, Vol. 1, eds. B. D. Reben. New Brunswick, NJ: Transaction.

________, Sharon Strover, and August E. Grant. 1994. Social Aspects of New Media Technologies. In Media Affects: Advances in Theory and Research, eds. Jennings Bryant and Dolf Zillmann, 463-483. Hillsdale, NJ: Lawrence Erlbaum Associates.

Wind, Yoram. 1978. Issues and Advance in Segmentation Research. Journal of Marketing Research 15, (3):317-337.

Wresch, William. 1996. Disconnected: Haves and Have-Nots in the Information Age. New Brunswick, N.J.: Rutgers University Press.

Zaltman, Gerald and Melanie Wallendorf. 1977. Consumer Behavior: Basic Findings and Management Implications. New York, NY: John Wiley.

Zeithaml, Valarie A. and Mary C. Gilly. 1987. Characteristics of Affecting the Acceptance of Retailing Strategies: A Comparison of Elderly and Nonelderly Consumers. Journal of Retailing, 63 (1): 49-68.

© 1999 Daniel J. Davied, James E. Fisher, Mark Arnold, and David Johnsen. Published with permission of the copyright holder.