The Internet is a vital part of the "new," "knowledge" or "information" economy and there is concern some people use it and others don’t. Neoliberals are excited by the thought of having everyone shopping on the Internet. The global economy has no bounds. If an African tribesman can be persuaded to buy an axe from Arizona, so be it. Educators committed to eroding barriers that impede access to formal education are also keen on the Internet. For them, online (or distributed) learning constitutes a merry utopia unencumbered by university squabbles about parking places, food services, budgets, crumbling infrastructure and demoralized faculty members (Boshier & Chia, 2000).

There is allegedly a digital divide that separates Internet users from non-users. It’s flavour of the month in academic journals, election campaigns, scholarly conferences and government proclamations. In most discursive constructions, white privileged people stand on one side and poorer, often black-skinned people or ethnic minorities, hover on the other. The divide metaphor contains echoes of being reared on the "wrong" side of the (railway) tracks, the "bad" side of town or, in early Shanghai, poor side of the river. In all these discursive formations, there is a valued centre and maligned other.

In this study the task was to compare and contrast:

• The extent to which different ethnic groups (as indicated by the Language Spoken at home) recently used the Internet.
• Examine ways in which different ethnic groups recently used different attributes (e.g. email, games, chat) of the Internet
• Measure the frequency to which different ethnic groups used the Internet.

Comparative educators committed to equity and learning in out-of-school settings need to know who’s using the Internet and for what. Internet use is not entirely determined by bits, bytes and bandwidth. It’s also constructed by cultural factors. Hence, a review essay for Comparative Education Review (Lindsay & Poindexter, 2003) boldly asked how "race and other demographic variables curtail students’ and citizens’ abilities to participate in and subsequently benefit from technology?" and complained about the lack of data on "racial politics and racial identity" as factors framing access to technology. Even the corporatist British Columbia (B.C.) Premier’s Technology Council agreed with this exhortation. "Despite great leaps made in rates of connectivity, there remains a digital divide – a line demarcated by … ethnicity …that determines who is online in British Columbia and who is not." (Digital Divide in B.C. 2002).

The notion of a digital divide has heuristic appeal but, like most things in life, is overly simplistic. Moreover, will it hold up in a postmodern city like Vancouver? If it refers only to the presence or absence of bandwidth, then Canada is digitally divided. But when authors claim it is primarily a psycho-cultural divide based on ethnicity or socio-economic status, simple-minded binaries lose explanatory power.

Vancouver is, in many ways, the end of the road, but different to other North American cities. What was once an outpost of British colonialism devoted to finding fish and furs, mining minerals and felling forest has become a bustling multicultural gateway to Asia. Almost everyone in Vancouver was born somewhere else and has strong ties to their homeland.

Vancouver is one of the most multicultural cities in the world. Even Her Majesty the Queen thought it appropriate to comment on the diversity of B.C. "You reflect the changing face of Canada, and in a broader sense, the rich cultural diversity of the Commonwealth … here in British Columbia, as elsewhere in Canada, you are crafting a multicultural society that provides a model for the rest of the world." ("You reflect," 2002) There are few African-Canadians in Vancouver. In 2001, the most visible ethnic minority was comprised of Chinese people who constituted 17.69 percent of Vancouver’s population (Statistics Canada, 2001a). Chinese is Canada’s third most common mother tongue but the first language in Richmond, B.C. where more than half the residents are from Asia or descended from Asian-Canadians.

Vancouver is a favoured destination for immigrants. Most used to be from Europe, but since about 1967, the focus has shifted to Asia. At the dawn of the 21st century, 1,967,475 residents inhabited the City of Vancouver. Of this number, 1,184,495 (60.20 percent) cited English as their mother tongue while 293,065 (14.90 percent) told the census-taker a Chinese language was their mother tongue (Statistics Canada, 2001b).

This is a good place for Internet-related research. British Columbia is the most connected of the Canadian provinces – with a telephone poll suggesting more than six out of ten people have access to the Internet. Three out of four small businesses are connected, along with all schools, universities and colleges.

According to The Digital Divide in British Columbia (2001), 53.30 percent of rural inhabitants and 62.70 percent of urban dwellers claim they use the Internet. In the B.C. Lower Mainland, various ethnic groups appear to use the Internet for different purposes. Casual observation, coupled with sporadic visits to Internet gaming parlours, suggested Chinese-speaking residents might be bigger Internet users than English speakers. What would it mean to have a "digital" divide between Chinese and English-speaking people? Would it excite as much interest as a black/white divide?

This study was not the first effort to research the digital divide in B.C. However, by abandoning telephone polling, in favour of administering a questionnaire and speaking (face-to-face) with a large number of citizens, it charted new territory. It was also designed to rectify problems in other surveys designed to measure Internet use in British Columbia. These problems are as follows:

• To gauge Internet use, Statistics Canada depends on telephone surveys covering the population aged 15 and over living in private households. Telephone interviews have merit but the reliability and validity of data thus derived is often questionable. Moreover, it is hazardous to call people at dinner time.
• Even more astonishing is the Statistics Canada criterion used to assess Internet use. "Respondents were identified as Internet users if they had used the Internet at least once in the 12 months preceding the survey." (Statistics Canada, 2001d) Hence, respondents were expected to remember what they had done up to a year prior to a telephone conversation from a call centre. In other procedures the telephone interviewer asks if anyone in the house is using the Internet. What Statistics Canada calls "regular users" are those who responded "yes" to the question "In a typical month, does anyone in your household use the Internet?" (Statistics Canada, 2003c) In this study, the focus is in recent Internet use (today or yesterday), type of and frequency of Internet use.
• Statistics Canada made large generalizations based on conversations with 3,300 respondents. In the study reported here, authors administered a questionnaire to and talked with 3,537 respondents.

Purposes

Having regard to the foregoing, the purposes of this study were to:

• Examine the extent to which recent Internet use (today or yesterday) varied as a function of ethnic origins (as measured by language spoken at home)
• Measure the extent to which persons speaking Chinese, English and Other languages had recently made use of different elements of the Internet
• Compare the "frequency" with which persons speaking Chinese, English and Other languages use the Internet.

There were three dependent variables.

• Recent Internet Use (RIU) - using (or not using) the Internet (today or yesterday). Respondents were asked "Have you used the Internet today or did you use it yesterday?"
• Type of Internet Use (TIU): If respondents answered "yes" to the question on recent Internet use, they were then asked "What For?" There were six choices - "E-mail", "Used the Web", "File Transfer", "Games", "Chat-room" and "Other Uses".
• Frequency of Internet Use: The question for this variable was "On average, how often do you use the Internet?" Choices included "Several times every day", "Once or twice a day", "Once or twice a week", "Once or twice every two weeks", "Once or twice per month", "Once or twice a year" and "Never use it".

Age was measured by having respondents note the year of their birth. Gender was measured by asking "Are you a woman or a man?" With regard to Language Spoken, respondents were presented with an open-ended question asking "What language do you mostly speak at home?"

In contrast to Age, Gender and Language Spoken at home, Socio-Economic Status (SES) was more complex. Like most places, the Lower Mainland of B.C. is divided into postal codes or forward sortation areas. Using Statistics Canada census data, researchers can secure a detailed profile of people residing in different postal codes. Like electoral districts, some postal codes embrace a large chunk of territory while others are smaller. For example, the University of British Columbia is in V6T, in contrast to the sprawling V4P zone in South Surrey. Forward sortation areas (FSAs) identify the first three digits (e.g. V6T) of the postal code. It is these that typically get used by market analysts, political pollsters and academics.

For each FSA, Statistics Canada collects data on many variables, some of which were imported for the purposes of this study. Examples included Literacy Rate (in quintiles), average Number of Bedrooms in dwellings, Median Family Income, Employment Rate, percentage of population who Own Property within this postal code. A composite measure of socio-economic status (SES) was created by summing data from six variables described above along with others such as Average Value of Dwelling, Median Individual Income, and percentage of residents with University Education. The composite index was calculated by first recoding each variable into five categories. Hence, for Median Family Income, respondents residing in areas earning the lowest 20 percent were assigned 1; the next 20 percent of postal codes were coded 2, and so on up to code 5 which identified postal codes with the highest 20 percent of incomes.

The questionnaire was administered to 3,537 citizens in educational, community, family and other settings in the Lower Mainland of British Columbia over a period of seven months. This yielded 3,502 useable responses. Respondents were found in the less affluent or more outlying areas such as Richmond, Maple Ridge, Squamish, Delta, Abbotsford, Surrey and Coquitlam as well as in higher SES areas such as the Vancouver-Westside.

The routine was to approach people seated at, for example, Calhoun’s or Tim Horton’s coffee shops, and say "Excuse me, I’m from UBC … We’re interested in … the Internet. Could you complete this? It takes about four minutes." The questionnaire was placed on the table with a pen or pencil pointing at the words "Your Name is Not Required." Most people appeared happy to take the 4 to 6 minutes needed to answer the questions. The questionnaire was professionally printed and attractive. Many respondents commented on its high face validity.

Of the 3,537 respondents, 737 (20.84 percent) were contacted in family settings, 710 (22.07 percent) were secured in Community Colleges, 525 (14.84 percent) from the University of British Columbia, 496 (14.02 percent) of questionnaires were completed in public spaces (such as Vancouver International Airport, Granville Island market, various shopping malls), 433 (12.24 percent) were done in coffee shops, 414 (11.71 percent) by people at work, 222 (6.28 percent) from high schools.

Out of 3,502 respondents to the question on Gender, 1,531 (43.72 percent) were men and 1,971 (56.28 percent) women. Concerning Age group, 16.48 percent were teenagers (13 – 19 years), 39.45 percent young adults (20 to 34 years old), 29.41 percent mature adults (35 to 50 years old) and 14.68 percent seniors (51 years and older). The mean age of women was 33.73 years (SD = 14.19 years) and men 33.97 years (SD = 14.98 years).

Just on 71.30 percent of respondents reported speaking English at home and 18.60 percent Chinese. The 10.10 percent of respondents reporting speaking languages such as Spanish, French, Japanese, or Hebrew were recoded as speakers of Other languages. Most of the remaining analysis focuses on respondents speaking English, Chinese and Other languages.

Of the 3,413 who responded to the question on Occupation, 24.70 percent were coded professional/technical; 16.70 percent were university students; 13.00 percent college students; 12.20 percent high school students; 9.10 percent were deemed to be involved with clerical/sales; 6.60 percent were skilled workers; 6.40 percent said they were retired; 4.50 percent were categorized as managerial; 3.80 percent unskilled workers; 1.90 percent homemakers; and 1.10 percent noted they were unemployed.

The first dependent variable concerned Recent Internet Use (today or yesterday). The first task was to examine the relationship between Recent Internet Use, Age and Gender. Non-users were coded zero and Recent Internet users (today or yesterday) one. Fig. 1 shows the mean use scores (from zero to one) plotted by Age and Gender. Overall, younger people were significantly more likely than older adults to have used the Internet (today or yesterday) (F = 46.12, df = 3, p < .001). There were 580 teenagers, 1,390 young adults, 1,031 mature adults and 521 seniors in this part of the analysis. Although older men were slightly more likely than older women to have used the Internet, gender differences were not significant. There was also no significant interaction between age and gender.

INSERT FIGURE 1

Fig. 1 clearly shows age matters but gender doesn’t. At all ages, women know less about the Internet (Boshier, Kow & Huang, 2003) but, contrary to feminist worries (e.g. Jackson, Ervin Gardner & Schmitt, 2001, Gilligan, 1982, Bimber, 2000) and assertions, at least among this large group of B.C. Lower Mainland residents, women and men were using the Internet to the same extent.

Fig. 2 shows the extent to which different-aged Chinese, English and speakers of Other languages had used the Internet (today or yesterday). Of the 652 Chinese speakers, 579 (88.80 percent) said they’d used the Internet (today or yesterday). This was in contrast to 79 percent of English speakers and 77.70 percent of those speaking Other languages. These differences were statistically significant (x² = 34.24, df = 2, p < .001). With the exception of teenagers, Chinese speakers in Lower Mainland, B.C. were significantly more inclined than were English or speakers of Other languages to have reported using the Internet (today or yesterday) (F = 3.68, df = 2, p < .02).

INSERT FIGURE 2

Concerning age, 93.59 percent of Chinese-speaking teenagers and 67.86 percent of seniors had used the Internet (today or yesterday) (a difference of 25.73 percent). For English speaking teenagers and seniors there was only a 21.35 percent difference. However, as shown in Fig. 2, the biggest difference in Internet use by age was for speakers of Other languages. Here, 97.30 percent of teenagers but only 56.67 percent of seniors reported using the Internet (for a generational-difference of 40.63 percent).

Critics might argue comparisons between speakers of Chinese, English and Other languages are spurious because the crucial variable in the digital divide is SES. This notion was examined by calculating mean Recent Internet Use scores for respondents residing in very low, low, medium, high and very high SES areas in the Lower Mainland. The mean composite SES for recent Internet users was 2.99 (SD = .98); the mean composite SES for non-users was 2.96 (SD = 1.01). Hence, there was no significant difference in the composite SES of the users and non-users. That satisfactorily toppled at least one pillar of conventional thinking about the digital divide. Despite this, authors proceeded with an analysis involving specific components of SES.

Using the respondents’ place of residence (as indicated by a postal code), authors were able to establish median incomes of persons in each area. Incomes were divided into quintiles with 1 indicating the lowest 20 percent and 5 the highest 20 percent of incomes. The 744 respondents residing in areas where the average family income was very low (i.e. less than $38,000) had a mean Recent Internet Use score of .83 (SD = .37); the 424 respondents in the low income areas (mean incomes around $38,000) had a mean Recent Internet Use score of .79 (SD = .40); the 663 in areas with a median family income around $45,000 had a mean Recent Internet Use score of .82 (SD = .82). The 657 respondents living in high income areas (median family income = $54,000) had a mean Recent Internet Use score of .83 (SD = .40). The 568 respondents living in the very high income areas (median family income ($58,000) had a mean Recent Internet Use score of .80 (SD = .40). In other words, there was no significant difference in the extent to which those living in very low, low, medium, high or very high income areas had recently used the Internet.

There was also data on the value of dwellings. The 45 respondents living in areas containing properties with very little value had a mean Recent Internet Use score of .69 (SD = .46). In the next "dwelling value" quintile were 294 respondents with a mean Recent Internet Use score of .76 (SD = .42). The 544 in the next highest quintile scored .78 (SD = .41). The 1,093 in the second highest dwelling value quintile scored .83 (SD = .37). The 65 respondents living in areas with the most expensive dwellings had a mean Recent Internet Use score of .48 (SD = .50). There were significant differences in Internet use when plotted against dwelling value (F = 14.18, df = 4, p < .001). However, those living in areas with the most expensive dwellings were less likely to have used the Internet (today or yesterday) than those in more modest districts! Does having a posh place to live suppress the urge to go online?

"Literacy rate" also showed mixed results. The 500 respondents (M = .80, SD = .39) living in areas with a "very low" literacy rate were just as inclined to have recently used the Internet as the 528 in "low" (M = .83, SD = .37) "medium" (n = 502, M = .80, SD = .40) or "high" literacy (n = 559, M = 80, SD = .40) areas. The exception was 705 respondents living in "very high" literacy rate areas. Their mean Recent Internet Use score was .87 (SD = .33). Although these differences in Internet use were statistically significant (F = 4.24, df = 4, p < .001) the relationship was confused and there was little to suggest "literacy rate" was a good predictor of Recent Internet Use.

Having established Chinese language speakers recently used the Internet more than English or speakers of Other languages, the next task was to discover which variables (in addition to Language Spoken) best explained Recent Internet Use. This task was accomplished by doing a discriminant function analysis of Recent Internet Use.

Recent Internet Use was predicted by combining the following variables: respondent Age, Language Group, Gender and composite SES. Discriminant function was well suited to this task because it predicts group membership (having recently "used" or "not used" the Internet). Unlike regression, discriminant analysis may be used with a nominal or dichotomous dependent variable. During the first step, an initial F-ratio is calculated which shows the relationship of each independent to the dependent variable. Next, standardized function coefficients are calculated. These are comparable to a beta weight in regression and show the extent to which each independent variable contributes to the function that predicts group membership when working with other independent variables. Functions are similar to factors (in factor analysis). Whether one, two or three functions are required depends on the structure of the correlation matrix. Discriminant function tests its equations by attempting to classify respondents into their correct group on the dependent variable (recently used or did-not-use the Internet).

For discriminant function purposes, it was necessary to convert Language Spoken into a dichotomous variable. Hence, in the analysis that follows, the comparison is between 1,968 respondents speaking English at home and the 579 speaking Chinese. As noted above, each respondent provided a postal code and SES was calculated by summing over census data to build a composite measure. Hence, as in the earlier analysis, SES was inferred from data pertaining to place of residence.

Table 1 lists variables in the order in which they entered the equation, the initial F-value and standardized function coefficient that resulted from the analysis. The larger the coefficient, the bigger the effect of the variable when working with others. The sign in front of the coefficient must be considered along with its size.

TABLE 1. Extent to which Age, Language, Gender and Composite SES Predicted Recent Internet Use.

Used Net Didn’t Use Net Wilks Lamba Univariate F Std function Coeffient.

This table shows Recent Internet users were younger than non-users. Recent Users were more inclined to be Chinese rather than English speakers and a bit more likely to be men than women. With respect to SES, there was almost no difference in the quality of areas inhabited by Recent Internet users and non-users. The standardized discriminant function coefficients show that, compared to speaking English or Chinese, SES or Gender, Age (.93) made the most powerful contribution to the equation. This was confirmed by its correlation with the function. Age (r = .95) made the largest contribution while Language Spoken (English or Chinese) (r = -.33), Gender (r = -.16) and composite SES (r = -.06) had a smaller impact. Hence, those who recently used the Internet (today or yesterday) were more inclined to be younger, male, Chinese-speaking and living in a slightly higher SES area. This analysis confirmed what was already known. Compared to Age, none of the available socio-demographic characteristics were good predictors of Internet use. If there’s a divide in the Lower Mainland of B.C., it mostly concerns age.

Of the 3,494 respondents, 80.70 percent said they’d recently used the Internet (today or yesterday). Of users, 73.00 percent had done e-mail; 62.00 percent had been on the Web; 24.00 percent had transferred a file; 16.00 percent had played games; 13.00 percent had visited chat rooms, and 24.30 percent had used the Internet for other purposes.

Table 2 shows the extent to which speakers of English, Chinese and Other languages reported using various elements of the Internet.

TABLE 2. Extent to which English, Chinese and Speakers of Other Languages Reported Using Internet Elements Yesterday or Today.

English Chinese Other

* denotes highest percentage for each type of Internet use

The most noticeable is Chinese speakers reported having used the Internet for various purposes more than those speaking English or Other languages (Table 2). Chinese speakers (30.30 percent) were more inclined to play games than were English (12.30 percent) or speakers of Other languages (16.30 percent). It was the same with chat-rooms. Only 8.70 percent of English speakers, in contrast to 24.90 percent of Chinese and 20.60 percent of Other language speakers, reported using a chat room (today or yesterday).

Chinese speakers reported using almost all elements of the Internet (e.g. email, Web) more than English or speakers of Other languages. But was this the case amongst all ages? Popular wisdom asserts Chinese-speaking teenagers are avid users of Internet chat and email. Are older Chinese similarly inclined? With these and related questions in mind, authors used the SPSS "select cases" routine to examine the Type of Internet Use in four age groups.

For teenagers, ethnic origin is a potent correlate of the Type of Internet Use. With the exception of email – where just over 91.00 percent of teenaged speakers of Other languages, 77.80 percent of Chinese and 76.40 percent of English speakers claimed to have used it – there were significant differences (between language groups) concerning web use, file transfer, games, chat and other uses of the Internet. The largest differences concerned file transfer where Chinese-speaking teenagers were the biggest users (x² = 22.11, df = 2, p < .001).

With regard to the Web, 78.50 percent of Chinese and 67.00 percent of English and Other language speakers claimed they’d used it (x² = 6.96, df = 2, p < .03). Chinese-speaking teenagers were also significantly more likely than English or speakers of Other languages to have played games (x² = 21.14, df = 2, p < .001). It was the same with chat rooms where Chinese-speaking teenagers (38.00 percent said "yes") and Other language speakers (37.00 percent) led English-speaking teenagers (25.70 percent) (x² = 9.20, df = 2, p < .01). The situation for "other" uses of the Internet was similar. Chinese (38.00 percent) and Other language speakers (54.10 percent) used them more than English speakers (38.00 percent) (x² = 16.80, df = 2, p < .001). Respondents in the three language groups differed significantly with respect to five of the six Internet elements.

Among young adults, there were significant differences between language groups concerning all six elements of the Internet. More than 86.00 percent of Chinese-speaking young adults, in contrast to 78.90 percent of English and 79.90 percent of Other language speakers, claimed to have recently used email (today or yesterday). Hence, Chinese-speaking young adults were significantly more inclined than the others to have used email (x² = 9.91, df = 2, p < .007). It was much the same for using the Web. Just on 75.00 percent of Chinese speakers, but only 67.20 percent of English and 63.40 percent of Other language speakers, claimed to have used it today or yesterday (x² = 8.21, df = 2, p < .01).

Chinese speakers were also more likely to say they’d transferred files. Just over 40 percent of Chinese, 22.50 percent of English speakers and 21.50 percent of those speaking Other languages, reported doing file transfers (x² = 38.71, df = 2, p < .001). It was the same with games where 28.60 percent of Chinese, in contrast to 17.70 percent of Other language and only 9.00 percent of English speakers said "yes" (x² = 70.93, df = 2, p < .001). These differences were even more profound for chat among young adults. Here, 30.80 percent of Chinese, only 8.30 percent of English and 24.20 percent of Other language speakers said they’d recently participated (x² = 98.72, df = 2, p < .001). There were similar large differences for "other uses" of the Internet with Chinese speakers leading the way (x² = 25.44, df = 2, p < .001).

For mature adults, there were no significant differences in the extent to which they reported using email or file transfers. However, there were significant differences in their use of the four other Internet elements. With respect to the Web, 66.10 percent of mature Chinese speakers, in contrast to 59.00 percent of Other and 58.60 percent of English speakers, reported using it. This difference was significant (x² = 8.49, df = 2, p < .01).

Mature adults were less likely to play games than were younger respondents. Nevertheless, there were differences among language groups. Just over 12.00 percent of mature Chinese speakers, 6.20 percent of English and 12.40 percent of Other language speakers claimed to have played games. This difference was significant (x² = 10.24, df = 2, p < .001). It was much the same for chat. Overall, this age group was not inclined to use the Internet for chat. However, 11.30 percent of Other language speakers, 7.90 percent of Chinese speakers and only 4.50 percent of English speakers reported using chat. These differences were significant (x² = 9.46, df = 2, p < .009). With regard to other uses of the Internet, 25.70 percent Chinese, only 13.40 percent of English speakers but 20.60 percent of Other language speakers said "yes" (x² = 17.90, df = 2, p < .001).

All these differences disappeared when this analysis was applied to seniors who, editors and readers will be appalled to note, were (gasp!) respondents aged 51 years and above. Nearly 60 percent of them reported having recently done email (today or yesterday), but there were no differences among language groups. Regarding the Web, 43.40 percent of seniors claimed to have used it but no language group was dominant. Just over 14.00 percent of seniors said they’d transferred files but there was no significant difference among language groups. There were also no significant differences in the extent to which seniors speaking different languages played games, used the Internet for chat or "other" purposes.

Respondents reporting they used the Internet "several times a day" were coded 6; those claiming they "never" used it were coded zero. Of all self-reported Internet users, 37.70 percent used it several times every day; 32.20 percent used it once or twice a day; 16.40 percent once or twice a week; 3.80 percent once or twice every two weeks; 3.20 percent once or twice per month; 2.00 percent once or twice every year. Only 3.80 percent said they had never used the Internet.

Among different age groups, the 1,382 young adults (M = 5.03, SD = 1.16) claimed to use the Internet more frequently than the 579 teenagers (M = 4.98, SD = 1.07), the 1,021 mature adults (M = 4.64, SD = 1.61) or 507 seniors (M = 4.12, SD = 2.05). These differences were statistically significant (F = 54.97, df = 3, p < .001).

The 645 Chinese speakers (M = 5.02, SD = 1.22) reported using the Internet significantly more frequently than the 2,474 English (M = 4.72, SD = 152) or 343 speakers of Other languages (M = 4.68, SD = 1.54) (F = 11.26, df = 2, p < .001).

The higher the Frequency score, the greater the Internet use. Unlike the question that asked "Have you used the Internet today or did you use it yesterday?" the one on frequency of use asked "On average, how often do you use the Internet?" Responses were arrayed on a seven-point scale ranging from "Never" to "Several times every day." The Recent Internet Use variable required a discriminant analysis but, being a seven-step equal interval variable, Frequency of Internet Use (FIU) was amenable to regression.

Four independent variables were used to predict Frequency of Internet Use – respondent Age, Gender, Language Spoken and composite SES. Working together, these variables produced a multiple r of .24. The most powerful predictor of Frequency of Internet Use was Age (standardized β = -.22). Compared to Age, the remaining independent variables made minor contributions to the regression equation as follows: Gender (β = .08), SES (β = .06), Language Spoken (β = .03).

Statistics Canada (2001) claimed "Young people are far more enthusiastic surfers than people in their 60’s and 70’s." They said 92 percent of Canadian teens used the Internet at least once during the 1999-2000 year. In the Statistics Canada survey, boy and girl teenagers were equally likely to use the Internet. Among 20 to 24 year olds, men (82 percent) were slightly more likely to do so than were women (77.00 percent). These findings came from a telephone survey where someone could be classified as a "user" if they said they had accessed the Internet only once in the previous year. In the present study, 89.23 percent of teenagers checked the "yes" box on a questionnaire that asked whether they had used the Internet today or yesterday. Why do teenagers use the Internet more than older adults? We posit these explanations:

• Homework: Most teenagers are at school and teachers have become increasingly prone to set homework that requires use of the Internet.
• Identity Formation: The Internet enables teenagers to form private relationships and pursue projects outside the purview of parental authority. It is an ideal instrument for securing identity, crossing boundaries, gaining independence and learning how to think in an abstract fashion.
• Free Goodies: The Internet delivers free music and other forms of entertainment. It’s the digital equivalent of the family car which, for earlier teenagers, was an instrument for fun, identity and independence.
• Active Participation: Unlike the passivity of television viewing, the Internet provides active participation.
• Parental endorsement: The Internet seems harmless and might be educative. Hence, Internet use is tacitly endorsed by parents whereas, in many circles, television is thought to rot the brain.
• Babysitter: Teenagers playing with the Internet don’t bother busy parents. Although parents are apprehensive about what’s being said, looked at and down- loaded, the kids are quiet. Particularly in hard-pressed professional families (of which there are many in B.C.), the Internet seems like a better babysitter than television.

Canada is a nation of immigrants. Apart from the First Nations (Native-Canadians), everyone arrived from somewhere. Immigration patterns and processes shape the digital divide in B.C. Although First Nations (i.e. native-Canadians) have difficulties securing Internet access because they live in remote areas, in the more affluent Lower Mainland of B.C., the largest ethnic minority (i.e. Chinese speakers) use the Internet more than the English-speaking majority. Moreover, SES does not appear to explain much variance in Internet use. So why do Chinese use the Internet more than English speakers? We posit these explanations:

• Education: Chinese speakers in B.C. are very committed to education. The Internet is now a necessary part of educating a child or adult. Hence, Chinese are prone to use it.
• Ties with home: Unlike labouring classes brought in to open mines and build the transcontinental railway, many recently arrived Chinese in B.C. are involved in business or professions back home.
• Social Solidarity: The Internet also helps build solidarity. Many Chinese new immigrants in B.C. have few local friends and largely depend on information from home. Ties are maintained by reading Chinese newspapers (from China and elsewhere), viewing websites, sending email and participating in chat on the Internet.
• No Accent: New immigrants sometimes have difficulties getting goods, services or contacts by phone. For them, the Internet has no accent.
• Job Search and Advancement: New and established Chinese-speaking immigrants are anxious to get their first and secure advancement in later jobs. The Internet enables a Chinese speaker to work at their own pace, seek clarification of English language sites and find information that lays behind the surface of an advertised vacancy. Amongst certain Chinese immigrant communities, the Internet is the only way to find a job.
• Social Habits: Young Chinese speakers are inclined to use the Internet for friendship, romance and fun. As shown above, young Chinese speakers are significantly more inclined than others to use the Internet for chat and games. More than 57 percent of Chinese teenagers, in contrast to only 36.70 percent of English speaking teenagers, used it for games. Thirty-eight percent of Chinese in contrast to 25.70 percent of English speaking teenagers, reported using the Internet (today or yesterday) for chat.

Earlier large-scale studies of Internet knowledge demonstrate Chinese in B.C. know less about the Internet than English speakers (Boshier, Kow & Huang, 2003). Yet, as shown here, Chinese speakers use it more than others. Here are possible explanations:

• Like driving a car, one doesn’t need to know much about the Internet in order to use it. Its use is utterly pragmatic. Provided it delivers the needed information, knowing about it doesn’t matter. This appeals to Chinese pragmatism.
• Some Chinese-speaking teenagers hail from "one-child" China and are living in Vancouver as an only child. Only children know less about the Internet than those with siblings (Boshier & Huang, 2003). But the relationship between Internet knowledge and Internet use is weak. Knowing little about the Internet doesn’t inhibit its use.
• Young Chinese speakers are highly social and use technology for social purposes – friendship, romance, fun and games. Technology is a tool. There is little reflexive examination of its use. It is to be used, not dissected.
• Internet connections and computers in Vancouver cost significantly less than in China so, after arrival in B.C., the first priority is getting a phone and Internet provider.

In the U.S., the digital divide allegedly separates white privileged Internet users from poorer black non-users (Digital Divide Network, 2003). In the Lower Mainland of B.C., Chinese are a very visible minority in and, as demonstrated here, it is they, not Anglo-European Canadians, who make greatest use of the Internet. Some U.S. data suggests the Asian love affair with the Internet percolates across the Canada/U.S. border. Hence, as Hubbard (2000) noted when referring to the U.S., "Instead of showing a predictable black-white gap… Asian-Americans, not whites, have the highest Internet and computer use. And while blacks at most income levels lag behind whites and Asians, it's Latinos, not blacks, who are the least likely to be wired. But no one's worrying aloud about an Asian-Latino digital divide."

So what happened to the digital divide? If its discursive construction dwells on connectivity and bandwidth, those in the affluent Lower Mainland are better off than people in beleaguered (and impoverished) B.C. coastal communities and hinterlands. However, the psycho-cultural construction of the digital divide is fascinating because Vancouver is a case where an ethnic minority (Chinese) uses the Internet more than the English-speaking (overwhelmingly white) majority.

Critics could claim the analysis presented here is misleading. After all, many Chinese immigrants to Vancouver are affluent and well-educated. Hence, they’re not like ethnic minorities in U.S. or European cities. Moreover, as Vancouver writers Doug Coupland, (2000) William Gibson (1984) and Paul Delany (1994) assert, Vancouver is different. What’s important is SES, not ethnicity.

The fact SES might explain differences in Internet use was certainly a worry. Yet, the composite measure of SES built from census data did almost nothing to predict Internet use (in the discriminant function equations). Moreover, the mean composite SES for Chinese speakers (M = 2.66, SD = .89) was significantly lower than that for English (M = 3.10, SD = .98) or speakers of Other languages (M = 2.88, SD = 1.00) (F = 52.27, df = 2, p < .001). Hence, avid Internet use by Chinese speakers did not stem from a superior SES. Their SES was significantly lower than that for English or Other language speakers. Nor were individual attributes of SES (e.g. Median Family Income, Literacy Rate, Value of Dwelling) good predictors of Internet use.

Critics might complain these analyses still don’t settle the matter because SES was not directly measured but inferred from place of residence (as indicated by the postal code). This is fair comment and, in future studies of the digital divide, it would be better to get SES data (on, for example, education, income, employment, literacy) by asking the respondent. The present data, while limited, strongly suggest the divide might be more complex than first thought.

It might comfort criticalists to portray the digital divide in black/white binary oppositions. This may work well in the United States. But in Canada, a place much admired by would-be-immigrants (Ng, 1999) committed to multiculturalism and characterized by Ralston-Saul (1997) as a "triumph of imagination", being a member of an (particularly Chinese) ethnic minority is not a bad thing.

If there’s a digital divide in Vancouver, it concerns age. It’s a gap between seniors and young people. With regard to ethnicity, Chinese speakers use the Internet the most. In typical Vancouver fashion, the digital divide is wrapped in won ton or dipped in dim sum. Exuding a distinctly oriental fragrance, it helps make Vancouver nonconforming, vibrant and intriguing.

Digital Divide in British Columbia (2002). Retrieved April 6, 2002 from the World Wide Web: http://www.gov.bc.ca/prem/popt/technology_council

Digital Divide Network. Retrieved June 3rd, 2003 from the World Wide Web: http://www.digitaldividenetwork.org/content/sections/index.cfm

Hubbard, L. (2000). Is the Digital Divide a Black Thing? Salon Magazine, Retrieved 31^st May, 2000 from the World Wide Web: http://www.salon.com/news/feature/2000/03/02/digital/index/html

Statistics Canada (2001a). Population by selected ethnic origins, 2001 census. Retrieved June 3^rd, 2003 from the World Wide Web: http://www.statcan.ca/english/Pgdb/demo43x.htm

Statistics Canada (2001b). Population by mother tongue, census metropolitan areas, 2001 census. Retrieved June 3^rd, 2003 from the World Wide Web: www.statcan.ca/english/Pgdb/demo18i.htm.

Statistics Canada (2001d). Internet Use in Canada, Retrieved June 2, 2003 from the World Wide Web: http://www.statcan.ca/Daily/English/020725/d020725a.htm