The Effect of Native Language on Internet Usage

Neil Gandal, Carl Shapiro

arXiv (arXiv: 0109009v1)

Generated on April 27, 2025

Abstract
Our goal is to distinguish between the following two hypotheses: (A) The Internet will remain
disproportionately in English and will, over time, cause more people to learn English as second
language and thus solidify the role of English as a global language. This outcome will prevail even
though there are more native Chinese and Spanish speakers than there are native English speakers.
(B) As the Internet matures, it will more accurately reflect the native languages spoken around the
world (perhaps weighted by purchasing power) and will not promote English as a global language.
English's "early lead" on the web is more likely to persist if those who are not native English speakers
frequently access the large number of English language web sites that are currently available. In that
case, many existing web sites will have little incentive to develop non-English versions of their sites,
and new sites will tend to gravitate towards English. The key empirical question, therefore, is whether
individuals whose native language is not English use the Web, or certain types of Web sites, less than
do native English speakers. In order to examine this issue empirically, we employ a unique data set on
Internet use at the individual level in Canada from Media Metrix. Canada provides an ideal setting to
examine this issue because English is one of the two official languages. Our preliminary results suggest
that English web sites are not a barrier to Internet use for French-speaking Quebecois. These
preliminary results are consistent with the scenario in which the Internet will promote English as a
global language.

Language TPRC.doc, 9/5/01 10:31 PM, Page 1 of 1 The Effect of Native Language on Internet Usage
Neil Gandal (Tel Aviv University) Carl Shapiro (University of California at Berkeley) September 2001
ABSTRACT In this paper, we empirically explore the relationship between native language and use of
the Internet. The ultimate economic and social questions we explore are: (1) how native language
affects use of the Internet, both in total and by type of Web site; (2) whether English is likely to retain its
“first-mover advantage” on the Web in terms of the language employed by Web sites; and (3) whether
the Internet ultimately will accelerate the movement to English as a global language. Our goal is to
distinguish between the following two hypotheses: (A) The Internet will remain disproportionately in
English and will, over time, cause more people to learn English as second language and thus solidify
the role of English as a global language. This outcome will prevail even though there are more native
Chinese and Spanish speakers than there are native English speakers. (B) As the Internet matures, it
will more accurately reflect the native languages spoken around the world (perhaps weighted by
purchasing power) and will not promote English as a global language. English’s “early lead” on the web
is more likely to persist if those who are not native English speakers frequently access the large
number of English language web sites that are currently available. In that case, many existing web sites
will have little incentive to develop non-English versions of their sites, and new sites will tend to
gravitate towards English. The key empirical question, therefore, is whether individuals whose native
language is not English use the Web, or certain types of Web sites, less than do native English
speakers. In order to examine this issue empirically, we employ a unique data set on Internet use at the
individual level in Canada from Media Metrix. Canada provides an ideal setting to examine this issue
because English is one of the two official languages. Our preliminary results suggest that English web
sites are not a barrier to Internet use for French-speaking Quebecois. These results are consistent with
the scenario in which the Internet will promote English as a global language.
Language TPRC.doc, 9/5/01 10:31 PM, Page 2 of 2 1. Introduction: Network Effects, First-Mover
Advantages, and Language In recent years, English has become the de facto standard for business
and academic communication and has to some degree attained the status of a global language.
English is the official language of the Asian trade group ASEAN and the official language of the
European Central Bank, despite the fact that the bank is in (Frankfort) Germany and neither the U.K.
nor Ireland are members of the European Monetary Union. 1 Several public schools in Zurich,
Switzerland are now teaching some of the elementary school subjects in English. This is occurring in a
country where there are four official languages --French, German, Italian, and Romansch. In a recent
European Union survey, 70 percent agreed with the notion that everybody should speak English. 2 In
this paper , we examine how native language affects Internet use. The goal is to determine whether the
Internet is likely to remain disproportionately English and thus whether the Internet will accelerate the
movement to English as a global language. Currently there is much more Internet content available in
English than in other languages. A recent estimate by Global Reach indicates that nearly 70 percent of
all Internet content is currently in English. Japanese and German follow with approximately 6 percent
each.3 The Internet certainly is an effective instrument for circulating English around the world. On the
other hand, it is quite possible that several languages will have a large critical mass of Internet content,
so that English’s role as a global language will diminish: Although 44 percent of current Web users
(March 2001) are native English speakers,4 web use is currently growing faster among non-native
English speakers. Indeed, it is estimated that by 2003, only 29 percent of all Internet users will be
native English speakers. Additionally because of low transaction 1 Wallraff, B., What Global Language,
Atlantic Monthly , November 2000, p.53-66. 2 Daley, S., “In Europe, Some Fear National Languages
are Endangered,” NY Times, 4/16/01. 3 See http://www.euromktg.com/globstats/refs.php3 . 4 Source:
See Global Reach at http://www.glreach.com/globstats/. Following English, 9.0 percent of all Internet
users are native Chinese speakers, while 8.6 percent are Japanese, and 6.1 percent are German.
Language TPRC.doc, 9/5/01 10:31 PM, Page 3 of 3 costs, the Internet is ideal for bringing together
members of small groups like speakers of Frisian, which is spoken by approximately 500,000 people
throughout the world. Concern that national languages are becoming endangered might lead policy
makers to require websites to be in the domestic national language. France already has many laws in
place that protect the French language. Quebec requires all websites in that province be available in
French. In Brazil, which has the largest Internet industry in South America, a bill was recently
introduced that would prohibit the introduction and use of foreign words.5 Additionally, another
standards war is already brewing on Internet domain names, which until quite recently exclusively used
romance language alphabets. Currently two groups, Verisign Inc. and China’s Internet Authority now
issue Chinese language domain names using two incompatible systems. This means that in order two
operate in both cyber-spaces, businesses would have to register with both authorities.6 Pindar Wong,
the former vice president of the Internet Corporation for Assigned Names and Numbers (ICANN), the
global body that governs the Internet indicated that the two incompatible systems risk “Balkanization of
the Internet, dividing the Internet up into islands of connectivity.”7 A key determinant of whether the
Internet will move towards Balkanization (critical mass of content in many languages) or
standardization is whether the “first-mover advantage” (of significant web content in English) will
encourage non-English speakers to use English language web sites. In such a case, existing content
providers would have little incentive to offer non- English versions of their sites, and new sites would
have a strong incentive to provide their content in English. Such a first mover advantage may lead to a
“bandwagon” because there are network effects in language: learning a second language is more
valuable, the more widely that language is used. A network effect exists when the value that consumers
place on a particular product increases as the total number of consumers who use identical or
compatible goods increases. In the case of 5 NY Times, 5/14/01, “English is Spoken Here…Too Much,
Some Say,” by Larry Rohter. 6 Wall Street Journal, 11/30/2000, “Will Language Wars Balkanize the
Web”, by Gren Manuel and Leslie Chang.
Language TPRC.doc, 9/5/01 10:31 PM, Page 4 of 4 an actual (or physical) network, such as the
telephone or email network, the value of the network depends on the total number of subscribers who
have access to the network. Since languages are in part communication technologies, the value of a
language network increases in the number of speakers and users of that language. Languages are
perfect substitutes, but they are incompatible in the sense that two individuals can talk with each other
only if they both speak the same language. 8 Languages, as a type of communications networks,
clearly are subject to strong direct network effects. In the case of virtual networks, that are not linked
physically, a network effect arises from positive feedback from complementary goods.9 The positive
feedback mechanism works as follows: the value of the base product (such as a DVD players) is
enhanced as the variety of (compatible) complementary products (content available on DVD disc)
increases; hence consumers will be more likely to purchase a base product with many compatible
complementary products. The variety of complementary products, in turn, will depend on the total
number of consumers that purchase the base product. As the number of consumers that purchase the
base product increases, there is a greater demand for compatible complementary products. This
increases the profitability of supplying complementary products. Since there are typically fixed or sunk
entry costs, production of the complementary products is characterized by increasing returns to scale.
So, more complementary products will be produced or developed for a base product with a large share
of the market. This further enhances the value of the base product, causing positive feedback in the
system: an increase in the sales of the base product leads to more compatible complementary
products, which further increases (the value of and) sales of the base product. See Chou and Shy
(1990) and Church and Gandal (1992). Languages are also subject to such indirect or virtual network
effects. The value of speaking English increases as the number of English language web sites (or other
content, such as books, 7 Ibid. 8 Of course, when there is no common language, people can still
communicate in non verbal ways, such as gestures, expressions, etc. 9 Examples of virtual networks in
which the value of the “base” product increases as the variety of complementary products increases
include computer operating systems, videocassette recorders (VCRs), compact disc players (CD-
players), and Digital Versatile Disc players (DVD-players).
Language TPRC.doc, 9/5/01 10:31 PM, Page 5 of 5 magazines, or movies) increases. This will lead to
an increase in the number of non-English speakers learning English in order to have access to the
English language web sites, since individuals who speak English will have more web sites to use. This
in turn will lead to an increase in the number of English language web sites. Markets in which there are
network effects are often characterized by tipping: once a system has gained an initial lead, there is a
snowball effect. Katz and Shapiro (1994, p. 105) note that positive feedback means that there is a
“natural tendency towards de facto standardization.” Hence it is possible that the first-mover advantage
of English may result in English remaining the dominant language on the web, while other languages
will end up serving niche markets. This could occur even though there are more native Chinese and
Spanish speakers than there are native English speakers. The outcome will depend, in large part, on
the strength of the virtual network effect: Does the large number of English language web sites
encourage non English speakers to learn English so that they can access them? 10 The use of
language on the Internet can fruitfully be viewed as a co-adoption process .11 Here “adoption” means
use of a particular language; we are thus thinking of language training and use as comparable to
technology adoption decisions that have been extensively studied. The operator of a web site “adopts”
a language by offering its site in that language. Likewise, an individual “adopts” a language by learning
that language. More specifically, focusing on the decisions made by web sites and users, we can
examine the dynamics of language adoption over three time frames. In the short-term (day to day),
individuals decide – based in part on their language skills and in part on the available offerings in
different languages – which web sites to visit, how long to stay at these sites, and whether to engage in
commercial transactions. These decisions determine actual Internet usage by different 10 In the case
of language, translation is an ex-post substitute for compatibility. If translation utilities worked well, the
issue of language would likely be less important. Given the subtleties involved in language, translation
by artificial intelligence is in its infancy and currently works quite poorly. 11 Co-adoption processes are
common when virtual network effects are present. For example, when there were two rival incompatible
formats for 56k modems, Internet Service Providers and consumers selected modem formats, each
influenced by the other group’s decisions. This same dynamic arises as well in various server-client
Language TPRC.doc, 9/5/01 10:31 PM, Page 6 of 6 individuals and groups (such as the group of native
French speakers). In the medium term (over a period of several months to a year or two), operators of
web sites decide which language to use for their site, and whether to offer their sites in multiple
languages (if permitted this choice by their local governments). These decisions are driven in large part
by the amount of traffic that a site expects to attract in one language or another, plus the incremental
traffic that a site expects to attract by offering its content in multiple languages. Over the long term
(more than one to two years), individuals (and their parents and teachers) make decisions about which
languages to learn. This decisions are driven in part by the desire to access certain content, as well as
the desire to communicate directly with others speaking other languages. In subsequent research, we
will to develop a simple theoretical model that captures these three inter-related decisions (visiting web
sites, creating content in different languages, and learning languages) that take place continually but
over different time frames. We believe that this theoretical treatment, building on the literature on
technology adoption and network effects, will support the following general line of reasoning: if in the
short term non-native English speakers routinely and extensively use English-language sites, the
incentives over the medium term for web sites to make their content available in other languages is
reduced, and as a result the incentive over the long term for individuals to learn English as a second
language is enhanced, all of which would support the prediction that the Internet will promote English
as a global language. On the other hand, if non-native speakers use the Web less, or conduct fewer
transactions over the Web than their native-English counterparts (adjusting for other factors such as
income and education), web sites will have stronger incentives to offer sites in languages other than
English, and English’s first-mover advantage on the Web is more likely to dissipate. In this paper, we
explore this issue empirically. To that end, we have obtained a unique data set on Internet use at the
individual level in Canada from Media Metrix. Why Canada? Canada provides an ideal setting to
examine this issue because English is one of the two official languages. French is, of course, the other
official language of Canada. 12 If French speakers are architectures, and influenced the battle between
the Netscape Navigator and Microsoft Internet Explorer browsers, as well as the adoption and use of
Sun’s Java. 12 Indeed, many of the studies in the economics of language focus on Canada in general,
and Quebec in particular. See Grin and Vaillancourt (1997).
Language TPRC.doc, 9/5/01 10:31 PM, Page 7 of 7 likely to use English content sites where there are
few substitutes in French, this suggests that the “first mover advantage” is important and that English
may indeed remain the global standard on the Internet. The attractiveness of beginning with a single
country (and a single region within a country) is that there is typically greater heterogeneity across
countries than within a single country. 13 Hence in this paper, we focus primarily on Canada and
Quebec. But the techniques discussed are applicable to other regions of the world.14 Our preliminary
results suggest that in most categories, native French speakers in Quebec are not less likely than
native English speakers to use the Internet. There are some slight differences in Internet use patterns:
native French-speaking Quebecois are somewhat more likely than their English counterparts to use
government sites, while English speaking Quebecois are somewhat more likely to spend time at search
sites. We also find that there are some differences in the percent of time spent at English language
websites between native French and native English speakers. The differences between the groups in
this dimension are less significant for the youngest users (age less than 15) and for the next youngest
group of users (ages 15 to 24). This is despite the fact that the youngest native French speakers in
Quebec are the least likely to have knowledge of English. While these results are quite preliminary,
they suggest that language is much less of a barrier for younger users. These results are consistent
with the scenario in which non-native English speakers extensively use English-language sites, the
incentives for web sites to make their content available in other languages is reduced, the incentive for
individuals to learn English as a second language is enhanced, and that the Internet will promote
English as a global language. 13 For example, in some countries, local phone calls are metered, while
in other countries (such as the U.S. and Canada), there is a fixed monthly charge for local service.
Additionally, Internet access speed might differ widely by country. It can be difficult to find data on and
control for these variables. 14 At a later stage, we envision examining data from many other countries.
There is some casual evidence that the most popular online destinations in Mexico originate in the U.S.
and do not target Spanish Speakers. See Heft, D. “Who Rules the Internet in Mexico? Why It’s
America, The Standard, June 12, 2001, at http://www.thestandard.com/article/0,1902,27096,00.html .
Language TPRC.doc, 9/5/01 10:31 PM, Page 8 of 8 Our work complements and make use of the
growing literature on the economics of language. Grin and Vaillancourt (1997) provide an overview of
the literature; a nice survey is provided by Grin (1996). The major research area within this field is the
empirical relationship between earnings and language attributes. Two recent papers are Chiswick and
Miller (1999), and Zavodny (2000). Grin (1990) and Church and King (1993) examine rational language
choice and public policy toward bilingualism using theoretical models. Rauch (1999) shows that
common language facilitates international trade in differentiated products. Freund and Weinhold (2000)
find that increased access to the Internet increases trade flows among developed countries. To the best
of our knowledge, there is no work on the relationship between Internet use and language, which is the
focus of this study. The data are described in Section 2. In Section 3, we describe the empirical
methodology. Our preliminary results and conclusions are provided in Section 4. 2. Data The project
employs a unique data set on Internet use at the individual level in Canada, which comes from Media
Metrix, the industry leader in the measurement of Internet use. The data include information on
demographics of the user such as income, education, family size, province, etc. Additionally, and this is
key for the study, the mother tongue of the user – English or French – is known. The data on Internet
use is very detailed. Complete click-stream data are available for the December 2000 period. These
data include a separate entry for each URL that is visited, and include the URL domain, as well as the
number of active seconds spent at each URL location. 15 Data were not collected on the language of
the web site. Hence, a computer (spider) program was written to check out the language of each URL
domain. Although, there were more than 4 million URL full pages, there are “only” approximately
100,000 unique URL domains; an 15 Data on total time is available as well. If a user does not enter a
key for 60 seconds, the active time count is halted. Hence for less than sixty seconds at a web site,
active time is equal to total time. For time spent on a page beyond 60 seconds, active time is less than
total time.
Language TPRC.doc, 9/5/01 10:31 PM, Page 9 of 9 example of a URL domain is
http://www.sfgate.com and an example of a URL full page is http://www.sfgate.com/classifieds/rentals/.
Of course, many of the URL domains contain some “interior” pages that are in one language and some
interior pages that are in other languages. We employed a crude version of this program for the
purposes of obtaining preliminary results. 16 In order to obtain preliminary results, we employed the
“basic” spider program on the approximately 40,000 unique URLs for Quebec. Hence, by this method,
we are able to assign a language to websites. We also consider it important to categorize the “type” of
website accessed, so we can understand in greater detail how different types of Internet usage are
influenced by language . Media Metrix did this categorization for approximately 2/3 of the observations,
but this only accounted for 25% of the websites. Research assistants classified a large portion of the
remaining websites in Quebec, so that nearly 90 percent of all observations in Quebec have been
categorized. Further research assistance will be needed to classify the remaining unclassified web sites
from Quebec (and the rest of Canada). Using the Media Metrix descriptions of the categories, we use
the following categories: (1) Retail, Business, Finance; (2) Entertainment, News, Sports, Technology;
(3) Education; (4) Search/Portals/Directories; (5) Services (Careers, Community, Hobbies, ISPs,
Mailboxes, Storage); (6) Government; (7) Adult. The following variables are available for the study: •
Active Time – This is the total time (in seconds) that the user was active in each of the seven categories
described above. 16 Our preliminary spider program classified all unique URL domains that have ASCII
characters above 192 (this includes all characters with accents marks such as “é” and “û”) as French. If
such characters were not present, the website was characterized as English. Although, this is a fairly
crude mechanism, it probably works reasonably well as a first cut, since most of the URL domains in
the data set are in either English or French. Of course, precision demands that we employ a more
sophisticated spider program or perhaps a program based on standard language identification
schemes. See Grefenstette (1995) for an overview. A more sophisticated spider program will eventually
have to characterize all of the approximately one million unique URL full pages.
Language TPRC.doc, 9/5/01 10:31 PM, Page 10 of 10 • Other Time – This is the total active time (in
seconds) that the user was active in all other categories17 • Age – Age of the user • Gender – A dummy
variable that takes on the value 1 if the user is female and 0 if the user is male. • Language – A dummy
variable that takes on the value 1 if French is the mother tongue of the user and 0 if English is the
mother tongue of the user. • Size – Equal to the number of members of the household, up to a
maximum of five. All households with 5 or more members have size equal to five. • Income – The
variable takes on the value 1 if the household income is less than $24,000, 2 if household income is
between $25,000 and $40,000, 3 if household income is between $40,000 and $60,000, 4 if household
income is between $60,000 and $75,000, 5 if household income is between $75,000 and $100,000,
and 6 if household income exceeds $100,000. • Kids – This is a dummy variable equal to 1 if there are
children under age 18 in the household. • Education – The variable takes on the value 1 if the individual
has completed middle school or less, 2 if the individual has attended but not completed high school, 3 if
the individual has completed high school, 4 if the individual has attended but not completed college or
university, 5 if the individual has an undergraduate degree from a college or university, 6 if the
individual has done some post graduate work, but does not have a post-graduate degree, and 7 if the
individual has a post-graduate degree. 17 As mentioned above approximately 33 percent of the active
time has not yet been categorized. These data are included when computing “Other Time.”
Language TPRC.doc, 9/5/01 10:31 PM, Page 11 of 11 • Pereng – This variable is defined to be the
percent of the active time that was spent on websites whose content is in English out of the total active
time spent at websites whose content is either English or French.18 3. Methodology Are there
differences in Internet use between native French and native English speakers? We initially focus on
Quebec. The reason for doing so is that there may be significant differences among provinces on
variables for which we have no control, such as speed of Internet service. Hence, it makes sense to
look at Quebec, which is the only province in Canada with significant proportions of native speakers of
both English and French in the Media Metrix sample. According to Statistics Canada, there are 602,865
native English speakers in Quebec and 5,728,290 native French speakers in the province. There are
50,585 people who are both native French and English speakers. Approximately 15 percent of our
Quebec sample are native English speakers. The first step is to determine whether native language
affects Internet use, where Internet use is defined to be active time spent on the Internet, regardless if
the active time is spent on French language or English Language websites. We look at this by category,
using the seven categories defined above. The next step involves examining what factors determine
the percent of the time that each user spends at English language websites. Initially, we examine this
by age group. 4. Preliminary Results and Conclusions Descriptive and summary statistics for the
Quebec data are contained in Tables 1-3. Table 1 provides summary statistics for all the variables used
in the analysis. The table shows overall, English speakers spent approximately 35 percent more time
on the Internet. Table 1 also shows that, on average, native English speaking Quebecois accessed
English content websites 87 percent of the time, while native French speaking Quebecois accessed
English content websites 18 The spider program could classify not all websites. This is due primarily to
the following reasons: (i) the unclassified websites were services that required the user to enter his/her
personal ID or information, (ii) the website
Language TPRC.doc, 9/5/01 10:31 PM, Page 12 of 12 64 percent of the time. This already suggests
that Quebecois are using the web intensively in English.19 Table 2 delineates active time by category.
Overall, users spend 36 percent of their active time at search/portal sites, followed by services (21
percent of active time) and retail/business (18 percent of active time). Users spend less than two
percent on average at government and educational sites combined. Table 3 breaks down the active
time by age category and shows that there are significant differences among age groups. In the under
15 age group, native French speakers spent approximately 66 percent more time on the Internet than
native English speakers, while in the 25-44 age group, native English speakers spent approximately 77
percent more time on the Internet than native French speakers. 20 In order to examine whether the
differences in Active Time vary across categories, Table 4 presents preliminary ordinary least squares
(OLS) regression results with Active Time in seconds as the dependent variable. Table 4 shows that
native French-speaking Quebecois are somewhat more likely than their English counterparts to use
government sites, while English speaking Quebecois are somewhat more likely to spend time at search
sites. 21 There is virtually no difference between native French and English speakers in the other
categories.22 These preliminary results suggest that in most categories, native French speakers in
Quebec are not less likely than native English speakers to use the Internet.23 Table 5 examines what
factors determine the percent of the time spent at English language websites. Table 5, which reports
ordinary least squares (OLS) regressions, shows that language is much less of a barrier for the
youngest users (age less than 15) and for the next youngest group redirected the user more than four
times. 19 There was little difference in this measure among the various age groups. 20 The age group
classification was chosen to match the data from Statistics Canada. See table 6. 21 The differences are
significant at (approximately) the 90 percent level of confidence. 22 English speakers spent significantly
more active time than French speakers in the adult category. 23 These results are still quite preliminary
b ecause 10% of the active time data for Quebec has not yet been categorized.
Language TPRC.doc, 9/5/01 10:31 PM, Page 13 of 13 of users (ages 15 to 24). This is despite the fact
that the youngest native French speakers in Quebec are the least likely to have knowledge of
English.24 It goes without saying that these results are also preliminary and further research needs to
be undertaken.25 While these results are quite preliminary, they suggest that language is much less of
a barrier for younger users. These results are consistent with the scenario in which non-native English
speakers extensively use English-language sites, the incentives for web sites to make their content
available in other languages is reduced, the incentive for individuals to learn English as a second
language is enhanced, and that the Internet will promote English as a global language. The next step
will be to incorporate data on the rest of Canada. It will be extremely interesting to compare native
English-speaking residents of Quebec to English speaking Canadians in other major provinces, since
so many of the native English-speakers in Quebec are bilingual. According to Statistics Canada,
Quebec has a much larger portion of bilingual speakers than other major provinces. Are there
differences in Internet use patterns between English speakers in Quebec, and other English speakers
from other parts of Canada? The same techniques we described in section 3 will be employed to
answer this question. 24 See Tables 5 and 6. Of course, our group of under-15 users primarily consists
of youths in the 10-15 age group. This group is likely to be much more bilingual than others in the
under-15 group in the population. 25 Given that the dependent variable ranges between 0 and 1, logit
regressions will eventually be employed here. The OLS regressions just provide a quick look at the
data.
Language TPRC.doc, 9/5/01 10:31 PM, Page 14 of 14 References Chiswick, B., and R. Miller, “The
‘Double Negative’ Effect on Earnings of Limited Language Proficiency Among Immigrants in Canada,”
University of Chicago Center for the Study of the Economy and the State Working Paper: 148, February
1999. Chou, C. and O. Shy, 1990, “Network Effects without Network Externalities,” International Journal
of Industrial Organization , 8: 259-270, 1990. Church, J., ad I. King, “Bilingualism and Network
Externalities,” Canadian Journal of Economics , 26(2):337-345, 1993. Church and Gandal, “Network
Effects, Software Provision and Standardization,” Journal of Industrial Economics , XL: 85-104, 1992.
Freund, C., and D. Weinhold, “International Trade and the Internet: Some Preliminary Evidence,”
mimeo, 2000. Grefenstette, G., 1995, “Comparing Two Language Identification Schemes,” 3rd annual
conference on Statistical Analysis of Textual Data, Grin, F., “The Economic Approach to Minority
Languages,” Journal of Multilingual and Multicultural Development , 11, 153-171, 1990. Grin, F., “The
Economics of Language: Survey, Assessment and Prospects”, International Journal of the Sociology of
Language , 121, 17-44, 1996. Grin, F. and F. Vaillancourt, 1997, “The Economics of Multilingualism:
Overview of the Literature and Analytical Framework", in W. Grabe (ed.), Multilingualism and
Multilingual Communities. Cambridge: Cambridge University Press, 43-65. Katz, M., and C. Shapiro,
“Systems Competition and Network Effects,” Journal of Economic Perspectives , 8: 93-115, 1994.
Rauch, J., “Networks versus Markets in International Trade,” Journal of International Economics , 48(1):
7-35, 2000. Zavodny, M., “The Effects of Official English Laws on Limited-English-Proficient Workers,
Journal of Labor Economics, 18(3): 427-52, 2000.
Language TPRC.doc, 9/5/01 10:31 PM, Page 15 of 15 Appendix: Tables Table 1. Descriptive Statistics
French speakers, N=819 VARIABLE MEAN STD. DEV MINIMUM MAXIMUM Age 36.02 16.15 2 26 99
Female 0.495 .500 0 1 Income 3.27 1.51 1 6 Size 2.83 1.24 1 5 School 3.97 1.58 1 7 Kids 0.44 0.50 0 1
Active time 34,154.43 53,852.03 1 849,625 Pereng 0.87 0.17 0 1 English Speakers N=163 VARIABLE
MEAN STD. DEV MINIMUM MAXIMUM Age 35.91 15.56 7 74 Female 0.54 0.50 0 1 Income 3.36 1.71
1 6 Size 3.01 1.41 1 5 School 4.11 1.62 1 7 Kids 0.38 0.49 0 1 Active Time 46,133.61 66,164.52 9
402,622 Pereng 0.64 0.25 0 1 26 Three French speakers are listed as “99” years old. No one else is
over 79 years old. Similarly there is a small group of French speakers under the age of 5. Nothing
qualitative changes if we ignore these two groups in the analysis.
Language TPRC.doc, 9/5/01 10:31 PM, Page 16 of 16 Table 2. Percent of Total Active Time by
Category Percent of Total Active Time Retail, Business 18.9 Entertainment 10.6 Education 0.9 Search
35.7 Services 21.1 Government 1.9 Adult 10.9
Language TPRC.doc, 9/5/01 10:31 PM, Page 17 of 17 Table 3: Active Time by Age, Language, and
Category (E=English, F=French). Age group Mean Active Time By Category Overall Retail, Business
Entertain Edu Search Services Govt Adult Under 15 E F 9,564 15,900 890 2119 2162 3804 104 308
5848 5215 1876 4389 192 340 676 1168 15-24 E F 39,741 27,707 3890 3877 5052 3810 1628 1029
14010 9176 6533 5343 485 1691 4792 4130 25-44 E F 64,632 36,483 14520 7880 9094 4672 851 732
17072 10501 11678 8408 988 1309 7586 4035 45-64 E F 54,976 48,707 10524 9004 4671 6681 698
712 16937 14659 7257 8575 1381 1423 11422 4347
Language TPRC.doc, 9/5/01 10:31 PM, Page 18 of 18 Table 3: Preliminary Regression Results The
Dependent Variable is Active Time in the Category. The t-values are in parentheses. (Bold fonts mean
significance at the .05 level.) Retail, Business Entertain Edu Search Services Govt Adult Constant
-2073.94 (-0.73) 2793.33 (1.15) 71.33 (0.11) 11482.38 (3.46) -4623.1 (-1.30) 1157.5 (1.87) 10509.7
(3.84) Age 21.26 (0.59) 35.36 (1.17) -6.02 (-0.76) 67.39 (1.64) 54.80 (1.23) 3.01 (0.39) -35.58 (-1.00)
Female 214.39 (0.21) 705.22 (0.81) -43.21 (-0.19) 425.14 (0.36) 2528.63 (1.97) -136.16 (-0.62) -5593.7
(-5.66) Language -912.37 (-0.68) -510.64 (-0.45) -118.24 (-0.41) -2441.23 (-1.56) 1810.15 (1.06)
494.67 (1.62) -3303.7 (-2.51) Size -1574.23 (-2.66) -278.50 (-0.55) 142.65 (1.19) -392.10 (-0.57)
-163.88 (-0.22) -48.99 (-0.40) -181.09 (-0.32) Income 1139.70 (3.26) 236.23 (0.77) 2.81 (0.04) -941.40
(-2.31) -1314.1 (-2.99) -204.18 (-2.75) -148.57 (-0.45) School 729.20 (2.11) -436.01 (-1.46) 216.23
(2.68) -513.71 (-1.29) 654.03 (1.51) 50.91 (0.69) 14269 (0.41) Kids 2180.96 (1.48) 860.86 (0.68)
-600.55 (-2.01) -263.67 (-0.15) 5758.6 (3.12) -60.81 (0.20) -1572.1 (-1.11) Other time 0.18 (18.05)
0.066 (8.80) -.00005 (-0.03) 0.21 (16.65) 0.21 (15.98) .0052 (3.44) .027 (3.83) # of Obs 888 820 330
934 913 505 472 Adj R2 0.29 0.09 .01 0.25 0.24 0.04 0.10
Language TPRC.doc, 9/5/01 10:31 PM, Page 19 of 19 Table 4: Preliminary Regression Results
Dependent Variable is Pereng. These regressions are done at the level of the individual by agegroup.
The t-values are in parentheses. (Bold fonts mean significance at the .05 level.) Age Less than 15 Ages
15 to 24 Ages 25 to 44 Ages 45 to 64 Constant 0.53 (2.87) 1.24 (6.46) 0.94 (9.28) 0.73 (3.88) Priage
0.022 (2.49) -0.019 (-1.94) -0.0037 (-1.67) 0.0028 (0.89) Active Time 3.78e-07 (0.32) 2.26e-07 (0.48)
4.88e-08 (0.25) 1.79e-07 (0.75) Female -0.045 (-0.75) -0.032 (-0.85) -0.0074 (-0.30) -0.043 (-1.49)
Language -0.14 (-1.49) -0.18 (-3.69) -0.22 (-6.70) -0.24 (-6.31) Size 0.025 (0.56) -0.0044 (-0.22)
-0.0075 (-0.48) -0.012 (-0.68) Income -0.019 (0.87) -0.0020 (-0.15) 0.012 (1.33) 0.012 (1.06) School
0.034 (0.66) 0.0097 (0.45) 0.0028 (0.31) -0.00042 (-0.04) Kids 0.022 (0.14) 0.0025 (0.05) 0.017 (0.44)
0.045 (0.90) # of Obs. 100 162 418 269 Adj R2 0.08 0.09 0.09 0.13
Language TPRC.doc, 9/5/01 10:31 PM, Page 20 of 20 Table 5: Mother Tongue by Region Source:
Statistics Canada (1996) Region English French Eng & French % English BC 2,827,325 54,020 6,655
98 Ontario 7,825,360 484,620 33,940 94 Quebec 602,865 5,728,290 50,585 10 Other 5,884,610
405,570 17,925 94 Total 17,140,160 6,672,500 117,165 72 Table 6: Percent of Quebec Population with
Knowledge of Other Official Language (By Age Group) Source Statistics Canada (1996). Mother
Tongue Overall Less than 15 15-24 25-44 45-64 English 40 38 66 48 35 French 32 4 39 41 36