26
The Social Science Journal 48 (2011) 371–396 Available online at www.sciencedirect.com Public opinion on bilingual education in Colorado and Massachusetts Jennifer Fitzgerald Department of Political Science, University of Colorado, UCB 333, Boulder, CO 80309, USA Received 13 April 2009; received in revised form 2 June 2010; accepted 6 July 2010 Abstract In 2002, voters in Massachusetts and Colorado faced identical ballot initiatives to remove bilingual education programs from the public schools. The measure passed in Massachusetts and failed in Col- orado. This paper investigates the debates over the issue in these two states. It provides insight into how people reason with respect to minority politics. It also helps to make sense of the states’ divergent outcomes. Content analysis of letters-to-the-editor reveals that voters are motivated by ethnic compe- tition and fiscal concerns, as existing theories would predict. Additionally, citizens debate which kinds of programs work best for English learners and take stands on how these youngsters can be successful in life. The inter-state comparison reveals that a major factor distinguishing the two statewide debates was ethnic paternalism, a logic often used by members of ethnic majorities to justify restrictive policy decisions on the basis of what they think is best for the affected population. The analysis shows that themes related to doing what is best for English learners were significantly more salient in Massachusetts than Colorado. This finding implies that where public debate over this issue is framed in terms of helping minority youth, the fate of bilingual education is not secure. © 2011 Published by Elsevier Inc on behalf of Western Social Science Association. Keywords: Social issues; Public policy; Public opinion; Bilingual education; Ethnic politics; Paternalism; Content analysis In 2002 Colorado and Massachusetts voters faced a choice with implications for English learners in public schools: uphold legal provisions for transitional native language instruction or require that all teaching be conducted in English. This round in America’s decades-long battle over bilingual education policies, traditionally conceptualized as a civil rights issue for The author is indebted to Jason Barnosky, Joe Coleman, Eric Gonzalez-Juenke, Eric Kisskalt, Jennifer Wolak and two anonymous reviewers for their helpful comments on earlier versions of this paper. Tel.: +1 303 492 6235. E-mail address: Jennifer.fi[email protected] 0362-3319/$ – see front matter © 2011 Published by Elsevier Inc on behalf of Western Social Science Association. doi:10.1016/j.soscij.2010.07.001

Public opinion on bilingual education in Colorado and Massachusetts

Embed Size (px)

Citation preview

Page 1: Public opinion on bilingual education in Colorado and Massachusetts

The Social Science Journal 48 (2011) 371–396

Available online at www.sciencedirect.com

Public opinion on bilingual educationin Colorado and Massachusetts

Jennifer Fitzgerald ∗

Department of Political Science, University of Colorado, UCB 333, Boulder, CO 80309, USA

Received 13 April 2009; received in revised form 2 June 2010; accepted 6 July 2010

Abstract

In 2002, voters in Massachusetts and Colorado faced identical ballot initiatives to remove bilingualeducation programs from the public schools. The measure passed in Massachusetts and failed in Col-orado. This paper investigates the debates over the issue in these two states. It provides insight intohow people reason with respect to minority politics. It also helps to make sense of the states’ divergentoutcomes. Content analysis of letters-to-the-editor reveals that voters are motivated by ethnic compe-tition and fiscal concerns, as existing theories would predict. Additionally, citizens debate which kindsof programs work best for English learners and take stands on how these youngsters can be successfulin life. The inter-state comparison reveals that a major factor distinguishing the two statewide debateswas ethnic paternalism, a logic often used by members of ethnic majorities to justify restrictive policydecisions on the basis of what they think is best for the affected population. The analysis shows thatthemes related to doing what is best for English learners were significantly more salient in Massachusettsthan Colorado. This finding implies that where public debate over this issue is framed in terms of helpingminority youth, the fate of bilingual education is not secure.© 2011 Published by Elsevier Inc on behalf of Western Social Science Association.

Keywords: Social issues; Public policy; Public opinion; Bilingual education; Ethnic politics; Paternalism; Contentanalysis

In 2002 Colorado and Massachusetts voters faced a choice with implications for Englishlearners in public schools: uphold legal provisions for transitional native language instructionor require that all teaching be conducted in English. This round in America’s decades-longbattle over bilingual education policies, traditionally conceptualized as a civil rights issue for

� The author is indebted to Jason Barnosky, Joe Coleman, Eric Gonzalez-Juenke, Eric Kisskalt, Jennifer Wolakand two anonymous reviewers for their helpful comments on earlier versions of this paper.

∗ Tel.: +1 303 492 6235.E-mail address: [email protected]

0362-3319/$ – see front matter © 2011 Published by Elsevier Inc on behalf of Western Social Science Association.doi:10.1016/j.soscij.2010.07.001

Page 2: Public opinion on bilingual education in Colorado and Massachusetts

372 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

ethnic minorities, ended in an unexpected way. Colorado, a state with a strong conservativetradition, a voting history of defending English as the dominant language, and a relatively largepopulation of Spanish-speaking students, voted to retain bilingual education. Massachusettsvoters, notorious for liberal positions on social issues, opted to place their relatively few Englishlanguage learners into 1-year English immersion programs.

Inspired by these puzzling events, this paper asks: what considerations underpin voters’views on bilingual education? And, under what issue frames is a vote against bilingual educationmore likely? Answering these questions is important for a number of reasons. Empirically, thiskind of ballot initiative has major implications for young English learners and their families, andwe will likely see more of them in the future. Theoretically, these debates can help to enhanceour understanding of public opinion on issues that pertain to minorities. And in broader terms,they can also provide insight into the ways in which people reason on public policy issues thatare unlikely to affect them personally.

Through content analysis of Colorado and Massachusetts letters-to-the-editor published inthe months leading up to the 2002 elections, this paper examines and compares public opinionthemes in these states. The goals are twofold. The first is to provide a fuller representationthan currently exists of the themes guiding in citizens’ bilingual education judgments. Thesecond is to generate hypotheses to explain why bilingual education bans (and other minorityrelated ballot initiatives) succeed in some contexts and fail in others. The set of letters examinedprovides a glimpse into how citizens made up their minds on the issue in the two state contexts.To be clear, I do not offer direct causal claims, but rather provide preliminary evidence thatcan enrich future studies.

The main finding offered here is that the Colorado and Massachusetts debates were mostcentrally focused on the effectiveness of the different education plans. Another prominenttheme is the impact that different educational regimes can have on English learners’ presentand future lives. Stark, practical considerations, such as cost and legality, are also relevant.Themes related to ethnic competition—such as negative perceptions of English learners andtheir reluctance to assimilate—are present as well. In terms of comparing the debates, thetwo themes that distinguish these cases from one another are fiscal matters (more prominent inColorado) and the consideration of how these plans will affect English learners (more prominentin Massachusetts).

The main theoretical contribution that this paper makes to our understanding of publicsentiment on minority politics issues is centered on the concept of ethnic paternalism. I arguethat when engaged ethnic majorities (Anglos) see it as their responsibility to determine what isbest for minorities, they often adopt a paternalistic mentality that leads their policy preferencesin a restrictive direction. The evidence for this claim comes in the form of opinion statementsthat reveal paternalistic reasoning. Examples of such reasoning include concerns about whetherEnglish learners will drop out of school, whether they will get into college, what kinds of jobsthey’ll be qualified for and where they will be able to afford to live. Assertions about what is inthe best interests of English learners, I argue, are especially paternalistic. This theme is moreprominent in Massachusetts than in Colorado, and in Massachusetts it is more common thanstatements revealing antagonism toward minorities.

Furthermore, one can detect paternalism in debates over which type of language programis best for Limited English Proficiency (LEP) students. For example, letter writers claim

Page 3: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 373

that immersion works, that bilingual instruction does not (often supported by personal orsecond-hand accounts), and that other states (notably California) have had positive resultsafter scrapping bilingual education. I argue that this theme, too, belies paternalistic tendencieson two grounds. First, education research points quite conclusively to the benefits of bilingualprograms (see, for instance Burnham-Massey & Pina, 1990; Greene, 1997; Slavin & Cheung,2005), though this information is rarely found in public debates on the matter (McQuillian &Tse, 1996).

Second, members of non-English speaking communities, especially Spanish speakingminorities, are overwhelmingly supportive of bilingual education (Citrin, Wong, & Duff, 2001;Citrin, Kiley, & Pearson, 2003; Capetillo-Ponce & Kramer, 2006). And so, when public majori-ties take a stand against bilingual programs, they do so without the support of education researchand they override the preferences of communities directly influenced by the resulting legisla-tion. These two themes—concern over effects of bilingual education on LEP students’ livesand assertions about which type of language program is most effective—signal a paternalisticway of thinking about minority youth that is not identified in existing studies. As such, thispaper adds much needed nuance to our current understanding of how people arrive at theiropinions about minority politics issues.

1. Theoretical perspectives

Existing work conceptualizes the bilingual education debate as an example of a broaderclass of ethnic or racial issues in American politics. Citrin, Sears, Muste, and Wong (2001), forinstance, use support for bilingual education (and other issues such as immigration and minorityrepresentation) as an indicator of openness to multiculturalism. And Branton and Jones (2005)approach bilingual education as a racial politics issue. They use the same model to explainvariance in support for bilingual education and other policies that people associate with ethnicand racial minorities, such as affirmative action and welfare. As such, the hypotheses testedin these studies stem from the current academic dialogue over how to account for outcomesin minority politics. The main debate focuses on whether views toward ethnic or racial issuesare fundamentally about group competition for economic resources or socio-cultural status.Put more simply, minority politics studies tend to ask: Are tensions among groups rooted inrational interests or cultural reflexes (Huddy & Sears, 1990, 1995)?

The rationalist perspective predicts that voters’ views are shaped by interests, which aretypically understood in terms of monetary resources. Views on minority issues, therefore, arelinked to perceived financial competition. Anglo parents with children enrolled in schools withbilingual programs would be expected to object to the diversion of school funds to speciallanguage programs for minority youth. Or taxpayers without English-learning children in thepublic schools would be expected to hold negative views of such policies based on perceptionsof high costs for low personal benefits. Huddy and Sears (1995) and Citrin et al. (2003) findthat fiscal considerations play a role in Anglos’ opinions on bilingual education. Houvouras(2001) also finds evidence that rational interests help to predict attitudes toward these pro-grams. A debate characterized by rational interests will emphasize costs and competition forresources.

Page 4: Public opinion on bilingual education in Colorado and Massachusetts

374 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

From the cultural perspective, the forces driving public opinion are linked to the symbolicnature of ethnic issues (Sears, Van Laar, Carrillo, & Kosterman, 1997). The argument withrespect to bilingual education is that established Americans have throughout history perceivednew immigrants as resistant to assimilation and as threatening to the American way of life(Perlmann, 1990). Language is an important symbol in this conceptualization, as English isconsidered to be a key component of Americanism (Citrin, Reingold, Walters, & Green, 1990;Chandler & Tsai, 2001; Citrin, Sears, Muste, & Wong, 2001b; Theiss-Morse, 2009). Publicopinion research generally supports the case that the symbolic importance of English as ourcountry’s one common language underlies most support for eliminating bilingual educationfrom the public schools (Citrin et al., 2003; Houvouras, 2001). With respect to debate themes,ethnic competition motivations would be evident in opinion statements that characterize mattersof immigration, assimilation and multiculturalism in terms of conflict and threat.

Though scholars interested in the role of inter-group bias typically conceptualize majoritymembers’ attitudes toward minority issues in terms of cultural competition, there is more tothese views. Much research on out-group attitudes reveals that these dispositions are more mul-tifaceted than studies of minority politics attitudes acknowledge. And unfortunately, existingsurvey measures used to explain attitudes toward bilingual education are designed to tap anti-minority sentiments in general and so cannot illuminate the precise mechanisms at work. Thepresent study considers an alternative attitudinal component to the standard ethnic antagonismmeasures: ethnic paternalism. This ideological dimension has not received much attention incontemporary research on minority policy attitudes. And furthermore, none of the evidence ofthis attitudinal component takes the form of public opinion statements. Examining minoritypolitics attitudes as they are justified in people’s own words provides an opportunity to deepenour understanding of how members of mass publics arrive at and defend their stated policypositions.

In existing work, paternalism has been examined as a component of racial attitudes. Forinstance, racial paternalism (or paternalistic racism) has been identified in a wide range ofstudies. Myrdal (1964), Van den Berghe (1970), Johnson (1957) and Alston and Ferrie (1985)provide historical studies that identify a paternalistic ideology underlying southern whites’views toward blacks in the American south. One policy oriented manifestation of that pater-nalism came in the form of arguments against desegregation, which many whites argued wasnot in the best interest of African Americans. From the field of communications comes Meyers’(2004) study that identifies racial paternalism in media coverage of “crack mothers”. Legalscholarship yields Delgado and Yun’s (1994) argument that paternalism is central to manyarguments against racial hate speech regulation. And sociologist Trepagnier (2001) tracespaternalism in the life histories of white women through focus groups that discuss inter-racialexperiences.

Jackman (1994) studies paternalistic sentiments based on race, class and gender as theyrelate to public policy attitudes. She explains:

Dominant attitudes that are paternalistic combine positive feelings toward subordinates withdiscriminatory policy dispositions. By taking over the definition of subordinates’ interests,dominant groups bring themselves to believe that the inequalities that they seek to perpetuateare actually to everyone’s benefit. This frees them to practice the inequality without any sense of

Page 5: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 375

unpleasantness or hostility. Instead, they can discriminate against subordinates while professingwarm feelings toward them and indulging in a gratifying sense of duty and obligation” (1994:271).

Jackman distinguishes racially oriented paternalistic attitudes from conflictive ones(described above as ethnic competition), which are viewed as less efficient for maintain-ing group dominance due to their antagonistic nature. She reports that approximately fortypercent of her white subjects exhibit racially conflictive motivations, and that an equalproportion present, instead, paternalistic ideologies. Dietrich (2005) replicates Jackman’sstudy and finds evidence of Anglos’ paternalism toward Latinos in addition to blacks.Below, letters-to-the-editor are examined to determine whether people’s views on bilin-gual education are shaped by ethnic paternalism. Supporting evidence would be foundin opinion statements that signal an interest in helping minority youth to be success-ful.

The analysis below uses existing theories of economic considerations and out-group attitudesas a departure point for examining people’s views on bilingual education. It also casts a widernet to identify other factors that are relevant to public opinion on the issue. This strategy isone advocated by Sniderman and Piazza (1993), who argue that not all public judgments onracial issues can be understood in terms of orientations toward minority groups. Instead, theysuggest, attitudes toward each issue must be studied as a public policy matter in and of itself,so that additional considerations also factor into the analysis. And so here, a broader set ofpublic considerations is also examined.

2. The issue

The electoral contests over bilingual education policies in Colorado (Amendment 31, orA31) and Massachusetts (Question 2, or Q2) came in the wake of similar debates in Californiain 1998 (Proposition 227) and Arizona in 2001 (Amendment 203). In all four states, the issuelanded on ballots as popular initiatives sponsored by English for the Children, a pro-Englishpolitical lobby. The proposals called for discontinuation of bilingual education, where studentslearn academic subjects in their native tongue while developing English fluency, usually overseveral years. These transitional programs would be replaced in favor of sheltered immersion,where all instruction is in English with minimal native language assistance. The initiativescalled for a maximum transition period of one academic year, devoted to intensive study ofthe English language. After this sheltered time, they would be mainstreamed into regularclassrooms to learn core subjects in English. Only in Colorado did the measure to dismantlebilingual education structures fail. Massachusetts handed the measure its strongest electoralvictory to date.

The controversy has incited passions from critics and defenders of bilingual education. InMassachusetts, protestors against Q2 chanted from the steps of the State House that the leaderof the campaign against bilingual education, Ron Unz, was a Nazi (“Bilingual Ed Advocates”,2002). More extreme behavior characterized the controversy in Colorado, where protestorsagainst A31 resorted to harassment, intimidation and vandalism against the measure’s support-

Page 6: Public opinion on bilingual education in Colorado and Massachusetts

376 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

ers. Anti-31 fanatics went so far as to torch the car of a vocal advocate of the initiative (Sanko,2002).

3. The cases

In Colorado 46 percent of the voters supported A31, and 68 percent of Massachusetts voterssupported Q2. Studying these statewide debates together reveals how two publics can wrestledifferently over the same issue, while also offering insight into the commonalities. A pairedcomparison of the disputes is particularly useful as it offers natural controls for some potentialexplanatory factors. For instance, elections were held simultaneously, and the wording wasthe same for both propositions. Both initiatives were drafted and sponsored by English forthe Children, headed by Ron Unz, a California software entrepreneur who embarked on thisproject with the successful referenda campaign in his home state.

Of course, there are differences between the situations in Colorado and Massachusetts. Thetwo states have very different political cultures. Whether they are termed Moralistic (CO)and Individualistic (MA) (Elazar, 1972; Erickson, McIver, & Wright, 1987), conservative andliberal (Erickson et al., 1987), or labeled by their dominant political affiliation—Republicanand Democratic—Colorado and Massachusetts are far from one another along a number ofconceptual continuums. A state culture perspective, however, only makes the outcome seemmore puzzling. In fact, Unz cited Massachusetts’s renowned liberal reputation as one reasonfor bringing his campaign to Bay State voters. He remarked, “Massachusetts is recognizedas the most liberal, best-educated state and the state where bilingual education began. Wethink passage of the bill here will have a considerable impact” (Hayward, 2002). Unz usedMassachusetts as a political experiment and considered that particular campaign his measure’stoughest test. He predicted that a win in what has been called the People’s Republic of Mas-sachusetts would foreshadow his movement’s chances of success at the national level (Citrinet al., 2003; Election, 2002).

A second factor that makes the outcome more puzzling is the linguistic portrait of the twostates’ populations. Though large proportions of the Colorado and Massachusetts populationsaged 5–17 speak a language other than English at home (14.5 percent in Colorado and 18.8percent in Massachusetts), over eighteen percent of these youth in Colorado speak English“not well” or “not at all.” In Massachusetts the parallel statistic is nine percent. Further, ofthose young people who speak a language other than English at home in Colorado, over77 percent are Spanish speaking, compared to the 48 percent who are Spanish speaking inMassachusetts (though Capetillo-Ponce and Kramer (2006) find that Massachusetts voterstended to associate Q2 with Latinos). In general, youth who speak a foreign language inMassachusetts are characterized by greater diversity of native language—47 percent speakIndo-European languages other than Spanish, and fifteen percent speak Asian or Pacific Islandlanguages.1 If one takes the perspective that resistance to bilingual programs stems from aneed to defend English as the dominant language, one would expect resistance to be strongerin Colorado than in Massachusetts.

Still, there are other differences between the Colorado and Massachusetts cases that makethe outcome less surprising. For instance, the campaign against the initiative in Colorado

Page 7: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 377

was disproportionately well-funded compared to its counterpart in Massachusetts. Heiress PatStryker, in an effort to preserve her daughter’s two-way bilingual education school in FortCollins (where native English speakers and native Spanish speakers learn in both languagesside-by-side), contributed $3 million to fight A31 (“Two states”, 2002). No other party to thisdebate in either state raised even $1 million (Citrin et al., 2003).

A second difference between the two cases is the organization of decision-making in thepublic school systems. School districts in Massachusetts had less flexibility in choosing theirown programs for educating English learners than in Colorado, where public education is moredecentralized. Massachusetts law at the time required that a public school with at least twentyLEP students with the same native language offer bilingual education. As discussed below, thedesire to retain local control of schools was a key factor in the Colorado debate. A third criticaldifference relates to elite support. In Colorado, most politicians and media outlets publiclycriticized the measure. In Massachusetts, however, there was significantly less elite consensus.The two main gubernatorial candidates took opposing stances. The victor, Republican MittRomney, made support for Q2 a part of his campaign. Finally, each of the two major Bostonnewspapers, The Boston Globe and The Boston Herald, advocated a different side of the issue.

4. Method

Opinion statements offer insight into public deliberation in these two statewide contests.Here, 211 letters-to-the-editor published in Colorado (n = 104) and Massachusetts (n = 107)are codified thematically, and editorial pieces and citizen quotations from newspaper articlesare used as supplemental resources. These opinion statements were published in the monthsleading up to the vote—most are from September and October, though a few are from the earlydays of November. The newspapers used are listed in Appendix B. The list of newspapersoffers variation along ideological and geographical lines, which is important when seekinga representative set of opinions (Sigelman & Walkosz, 1992). See the appendix for the dis-tribution of sources for the letters, which are all the available letters-to-the-editor printed intheir pages that deal explicitly with A31 or Q2. The sample is all of the letters made availablethrough Lexis-Nexis that addressed the two ballot measures. The unit of analysis is the opinionpiece, so the findings are presented as percent of letters expressing each theme. Letters are cat-egorized as pro-initiative (for eliminating bilingual education) or anti-initiative (for retainingbilingual education).2 The breakdown of pro- and anti- letters in Colorado and Massachusettsis presented in Fig. 1. These patterns parallel the outcome of the 2002 vote in that there weremore anti-initiative letters in Colorado, where the measure failed, than in Massachusetts, whereit succeeded. This implies that—in comparative terms—the letters reflect the weight of publicsentiment for and against the ballot measures.3

For the content analysis, specific issues were identified, coded, and combined into thematiccategories. On the pro-initiative side the five key themes are: Cost, Ethnic competition, Effectson LEP youth, the Education plan itself, and the Opposition. The anti-initiative letters includethese five themes plus three more: Legal issues, Control over education, and Effects on non-LEPyouth. The detailed process is diagrammed in Appendix A. The coding was performed manuallyby the author and duplicated by a colleague. Because the goal was to identify characteristics

Page 8: Public opinion on bilingual education in Colorado and Massachusetts

378 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

Fig. 1. Pro- and anti-initiative letters-to-the-editor.

of manifest content rather than underlying concepts, the coding was very straightforward.Therefore, there was very little discrepancy between the two coders, but in the five instancesof discrepancy the letters were dropped from the analysis.

Fig. 2 reveals the breakdown of themes among the pro-initiative letters (panel A) and thosepresent in the anti-initiative letters (panel B). Asterisks mark instances in which inter-statethematic differences are statistically significant (at 99% confidence levels). These patterns arediscussed in the sections below.

5. Strengths and limitations of a letters-to-the-editor approach

The letters-to-the-editor approach has its advantages and disadvantages. Its main strengthis that it allows people to express their views in their own words and on their own terms. Thisis valuable for the present study, which seeks to uncover nuances in the way people make uptheir minds about bilingual education and similar issues. These letters are also an importantresource due to the relatively sparse survey data that exists on bilingual policy preferences.Beyond the glimpse these letters offer into people’s choices on bilingual education, they also

Page 9: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 379

Fig. 2.

Page 10: Public opinion on bilingual education in Colorado and Massachusetts

380 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

give us a sense of the issue frames that shape public deliberation. By looking at a collectionof these letters in Colorado and Massachusetts, we can gain a sense of what aspects of theissue drove the debates. And because this procedure reveals the characteristics of each state’scontroversy, it also suggests answers to the question of why the outcome differed in the twostates.

Of course, there are drawbacks to this particular approach. Mainly, the use of letters-to-the-editor raises questions about how representative letter writers are of public opinion ingeneral. Existing evidence of the parallels between public opinion and letters is mixed onthis point. Hill (1981) finds no significant differences between letter writers and generalpublic opinion on the Equal Rights Amendment. Buell (1975) argues that some bias doesexist, though he also finds that letters-to-the-editor represent reasoned political discussion.Sigelman and Walkosz (1992) review existing work in this area and propose that examininghigh-salience issues and drawing from a range of newspapers significantly enhances the rep-resentativeness of the opinions expressed in letters-to-the-editor. This is relevant to the presentstudy. The high-salience of the bilingual education issue4 and the use of many newspapers ineach state should enhance the reliability of these letters as gauges of public opinion, broadlyconstrued.

Still, it is possible that the letters are skewed toward a sub-set of viewpoints, and this mightinfluence the findings for each state in that some frames are more present in the letter samplethan in the broader public discourse. For instance, people with certain viewpoints might beespecially likely to take the time to write and submit a letter. To this point, Wahl-Jorgensen(2001) reports that letter writers are typically males who are relatively advanced in age, andso issues of import to older men might be over-represented in editorial pages. However, this isunlikely to have a major impact on the cross-state comparison since those same biases shouldcompel both Colorado and Massachusetts residents alike.

Another issue related to the representativeness of the letters is that newspaper editors havethe final say in terms of the content that is published. Editorial privilege, therefore, can skewthe thematic distribution of letters. The evidence on this critique, as well, is mixed. The classiccitations of studies that uncovered editorial bias and that did not are Renfro (1979) and Grey andBrown (1970), respectively. Other research tells us that editors across the board tend to publishletters with emotion and personal content rather than letters containing research findings ortechnical details (Wahl-Jorgensen, 2001), and that brief and entertaining submissions are alsogiven priority (Wahl-Jorgensen, 2002). Some bias, therefore, can be expected as letters maketheir way through the editorial process. But this might lead us to expect these editorial biasesto suppress the appearance of letters about educational programs and favor ethnic competitionthemes. And therefore the direction of the bias would make it less likely to find evidence ofpaternalistic ideologies. Moreover, as with letter-writer bias, we can expect that these effectsmight influence the content of letters in each state debate, but that the inter-state comparisonshould still be valid.

Beyond considerations of the strengths and weakness of the specific approach used in thispaper, is the issue of relying on a single dataset to examine these state debates. This is notideal. However, the best companion data for this study would be identical, representative publicopinion surveys in Colorado and Massachusetts (preferably in Fall 2002). And unfortunately,this kind of data that asks the relevant questions of respondents does not exist. Therefore,

Page 11: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 381

the hypotheses generated from this analysis can be used to design better survey instrumentsthat more adequately reveal the nuances of public reasoning on bilingual education and otherminority politics issues.

6. Pro-Initiative arguments in Colorado and Massachusetts (down with bilingualeducation)

This section and the one following it compare pro- and anti-initiative themes, respectively.The purpose is to identify the common and distinctive dimensions of debate in the two statecontexts. The final empirical section, then, collapses pro- and anti- themes to create a big-picturecomparison of what each state debate—over all—was fundamentally about.

Cost: 25 percent of the Colorado pro- letters addressed the costs of bilingual education.They argued that: bilingual education is expensive; it uses tax money; Colorado’s social serviceswere already strained; and bilingual education amounts to social welfare. A Broomfield residentargues: “Properly managed, the change required by Amendment 31 will not have any additionalcost. The resources currently used to provide the current multiplicity of bilingual languageprograms can easily be reallocated to cover the relatively small demands of English immersion,if any” (Johnson, 2002). Cost was far less salient in Massachusetts. Only four pro-lettersaddressed this theme, one stating: “We should do away with bilingual education because it isa waste of money. . .” (Mariano, 2002).

Ethnic competition. Symbolic politics was present in pro-initiative letters in Colorado andMassachusetts. Sixty-one percent of the Colorado pro- letters made general points about immi-grants. These include: immigrants need to assimilate; Americans should speak English; learnor leave; new immigrants get preferential treatment; past immigrant groups learned; and bilin-gual education supports immigrant cultures. A woman asks, “Why do the Hispanic childrenhave to be catered to?” (Burdan, 2002) In Massachusetts, symbolic politics themes were raisedin 53 percent of the pro-Q2 letters. A Boston hair stylist put it this way: “If you’re going to livehere, you should speak the dominant language” (Bombardieri, 2002). An established immi-grant makes a case for Q2 based on her own experiences: “If you choose to come to the UnitedStates, you must learn the English language, U.S. laws and ways, and the earlier you learn thebetter. I’m a foreigner who came here at 14 with my brothers, sisters, and parents. We wentto immigration school for a year. . .we all made out okay” (Perault, 2002). And an educatorfrom East Sandwich plays the history card, “Our country was founded by British Pilgrims whospoke English. It has been our national language ever since. All immigrants come because theywant to, and it was reasonable to assume they intended to become part of America” (Stebbins,2002).

Effects on LEP youth: 64 percent of A31 proponents in Colorado wrote that the initiative’sgoals would better immigrant youth lives. Specific arguments include: immigrant kids shouldlearn English for their own good; not learning English limits the opportunities available tominorities; bilingual education traps LEP kids in foreign language ghettos; bilingual educationcontributes to the high dropout rates among Hispanic youth; and that bilingual education seg-regates LEP kids from their peers and offers them inferior educational and social experiences.Arguing in favor of A31, a Boulder resident writes:

Page 12: Public opinion on bilingual education in Colorado and Massachusetts

382 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

Effectively, bilingual education programs function as a parallel school system. . .These studentswill subsequently spend years in bilingual programs, struggling with translation problemsstemming from various levels of English proficiency while never gaining the strong Englishlanguage skills that lead to better grades, higher standardized test scores, and the ability to goon to college. The net result is that opportunities to advance into white-collar suburbia are cutoff (most notably for Hispanic kids). To your Hispanic readership I say, if you really love yourchildren and want them to be successful in America, keep them as far away from bilingualprograms as possible (Kelley, 2002).

This theme was even more salient to Massachusetts pro-writers, appearing in 84 percent ofthe letters supporting Q2. As articulated by a resident of Holyoke, “If you don’t speak English,you’re going to make less money, you’re going to have less success.” This letter further arguedthat eliminating bilingual education “absolutely is the moral thing to do” (Kurtz & Rogriguez,2002). These two statements not only rest on an apparent desire to do what is best for Englishlearners, but each also asserts paternalistically what the best actually is.

Educational plan. The most common theme among supporters of the proposals in both stateswas the educational plan itself. In Colorado, three-fourths of the pro-31 letter writers arguedcases based at least in part on pedagogy. The letters argued that bilingual education had been aproven failure and/or that 1-year of immersion would work. The letters often cited the positiveresults reported from the introduction of immersion in California or Arizona. In Massachusetts,the educational plan itself dominated the debate. 88 percent of pro-letters touted the benefitsof immersion and/or highlighted the failures of bilingual education. A South Boston residentwrites:

I wish the “activists” who decry the end of bilingual education would look at the irrefutableevidence: immersion works. . .and the school my children attend is proof. In South Bostonyou can find St. Mary’s Catholic School. What was once a primarily Polish elementary schoolis now one of the community’s most diverse schools. Kindergarteners come to school withlittle command of the English language. They not only adapt but excel at a rate equal to theirAmerican-born counterparts. If the Polish children at St. Mary’s School can do it, so can anyother of the many ethnic groups clamoring for bilingual support. It works and not playing thevictim goes a long way to the success of such a method (Lombard, 2002).

In similar terms, a Colorado resident evokes memories from his youth to argue that immer-sion works and that it is better for English learners in the long run:

My parents were part of a group of rural farmers in Minnesota who sponsored 30 to 40 familiesof displaced persons from refugee camps in Germany in the late 1940s. There were manyyoung children in these families who spoke Russian, Polish, Ukranian, etc. I was a junior highstudent at the time, and can specifically recall how quickly and effectively the kids learnedEnglish and became excellent students and citizens. It was an immersion program because itassumed that everyone’s best outcome was made possible by jumping right in. The parentsclearly had a more difficult time, but the accomplishments of their kids were a source of greatjoy and encouragement as well as a great stimulus to the parents to work on their English skills.(Carlson, 2002)

These two statements take strong stands on what works when it comes to English instruction.Such claims are more prevalent in Massachusetts than Colorado.

Page 13: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 383

Opposition. Half of the pro- letters in Colorado criticized the main opponents of A31: teach-ers’ unions, the educratic (or bilingual) machine; Democrats who pander to the unions; andheiress Pat Stryker. A Boulder resident makes this case, “. . .bilingual education programs func-tion as a parallel school system, with all the attendant spending and staff. Unions love the extrajobs, while the Democratic Party enjoys the political funding that flows almost unanimouslyto them via union dues” (Kelley, 2002). In Massachusetts, 37 percent of the anti-letters castdispersions on the opposition, arguing that the anti-Q2 camp would do or say anything to keepthe bilingual gravy train rolling. A Canton resident states, “The most vocal opponent [to Q2]in Massachusetts was the teachers’ union. Was it the students’ welfare or jobs and control ofthe educational process that motivated them?” (Levin, 2002).

Over all, the most common arguments on the pro- side in Colorado and Massachusettswere related to immigrant youth, providing evidence of paternalistic ideologies. Initiativeproponents maintained that English immersion is a superior instruction method to bilingualeducation. They also argued that it would be the best choice for the sake of LEP students.These paternalistic themes were especially salient in Massachusetts. It should be noted that fornone of these pro- themes was there a statistically significant difference (at the conventional 95percent level) between the Colorado and Massachusetts debates. But the statistical significanceof Effects on LEP youth just misses the cutoff (t = 1.98, significant at 94.8%), and surpasses theless demanding 90% hurdle. This demonstrates that paternalistic concern for minority youthrepresents the clearest distinction between the two states’ controversies.

7. Anti-initiative arguments in Colorado and Massachusetts (save bilingualeducation)

Cost. Cost was a salient theme in Colorado’s debate with 28 percent of anti- letters raisingthe issue. Though it was not clear whether there would be a financial cost associated withA31 (“Heated Battle”, 2002), concerns along this line were expressed. A Longmont residentexplained: “Amendment 31 is costly. It mandates more testing and requires new materials andextensive staff training. It is irresponsible and unsound financial practice to add additionaltaxpayer burden with no allotted funding source during a time of budget cuts” (Wren, 2002).Cost was not a significant feature in arguments against Q2. A resident of Newton links costand effects on non-LEP youth themes, “The Unz plan also hurts both mainstream kids andtaxpayers. Bilingual students who are put into mainstream classes before they have adequateEnglish and literacy skills will put an enormous strain on mainstream teachers. As a result, theteachers are likely to refer them to special education which is much more costly than bilingualeducation” (Soares, 2002).

Ethnic competition. Among opponents of the initiative, minority rights issues were notheavily emphasized in Colorado. Less than eight percent of these letters contained symbolicthemes. Still, some argued for protecting minority language rights and touted the benefitsof a multi-lingual population. An Evergreen resident equated the implementation of A31 to:“. . .stripping [native Spanish speakers] of their language, culture and heritage” (Fitzpatrick,2002). Symbolic themes did not dominate in Massachusetts, either, though they were morepresent. Twenty percent of the anti-letters in Massachusetts highlighted ethnic themes. A

Page 14: Public opinion on bilingual education in Colorado and Massachusetts

384 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

Framingham woman writes, “Telling immigrants that they have to learn English overnightonly tells them that they are not really welcome in Massachusetts” (Hurwitz, 2002).

Effects on LEP youth. Specific concerns about the impact of A31 on immigrant schoolchil-dren were represented in the Colorado letters. Some argued that replacing bilingual educationwith immersion would limit future opportunities for minority youth. A teacher of English as aSecond Language claimed, “If Amendment 31 passes. . .my students might end up as secondgeneration busboys or hotel maids” (Tracy, 2002). Some letters claimed that the “sink or swim”approach was needlessly stressful, and that it would cause non-English speakers to drop out ofschool. A student at Basalt High School wrote to The Aspen Times: “Another problem [withAmendment 31] would be the increase of school dropouts. Why would you want to go to schoolif you couldn’t understand any of it?” (Calderon, 2002).

In Massachusetts, several anti-Q2 letters reflected concerns about LEP youth. 31 percentof opinion pieces expressed sentiments about the negative effects the proposed plan wouldhave on children’s lives. Most of these predicted that eliminating bilingual education wouldbe detrimental to the kids’ futures. Additional comments on behalf of kids were that a sink-or-swim program would harm children in the long run more than it would help them, and alsothat separating LEP students from mainstream youth for an entire year would have negativeeffects on the young immigrants. As argued by a writer from Somerville, “[another letterwriter] reminds us that students are human beings and that traumatizing students in the nameof test scores and nativist fears will neither speed up learning nor help teachers understand thestudents they face” (Zucker-Conde, 2002).

Education policy. Though the Colorado outcome was a victory for supporters of bilingualeducation, it did not reflect a ringing public endorsement of the educational program itself.Only 61 percent of commentaries against the initiative directly addressed the merits of bilingualeducation or the problems with immersion. Still, the pedagogical issue was the most commontheme represented in Colorado’s anti-initiative letters. Some writers buttressed their argumentsby emphasizing the negative results of immersion in California or Arizona. A resident of Eriewrites, “The results from California are not encouraging. For 4 years in a row since this methodwas imposed there. . .the test scores of English learners there have lagged further and furtherbehind those of their fluent peers” (Meyer, 2002).

On the anti-Q2 side, the education plan was the most salient theme by far. 74 percent ofthe 66 anti- opinion pieces either made a case for keeping bilingual education on its mer-its or criticized its proposed replacement. Negative results from the transitions in Californiaand Arizona were often cited in support of these statements. Among the related argumentsdirected against Unz’s plan, the most popular was the notion that the one-size-fits-all nature ofthe 1-year immersion plan is not sound education policy. Nearly as prevalent were the argu-ments that a year is not enough time to ensure that all LEP kids reach English proficiency. Arelated sub-theme stems from research on language learning that distinguishes conversationalfrom academic fluency—again arguing that the 1-year limit on learning English before main-streaming is overly restrictive. A Westfield resident shares her own experience with Englishimmersion.

Having come from an immersion English program, I went through elementary and middleschool with apparent success. My conversational English skills were excellent. In high school, I

Page 15: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 385

noticed that I was not as skilled in English composition, vocabulary and reading comprehensionas my native English-speaking peers. . .It’s impossible to provide in just one year the kind ofinstruction necessary for successful mastery of the English language (Parrilla-Sweeney, 2002).

These comments are especially interesting theoretically because they reject the idea thatall young English learners are the same. This reasoning stands in contrast to those that wouldlump members of this out-group together and assume similarities that do not exist.

Opposition. Nineteen percent of Coloradoans’ letters against A31 attacked the initiative’sbackers. Most of the negativity was directed at Ron Unz. Unz’s critics in Colorado focusedmainly on the fact that he is an outsider whose expertise is in business rather than education.Some of the anti-letters also cast doubt on the true intentions of the initiative; two called Unz aracist. A resident of Lafayette argues: “Colorado, where is our pride? Why let Unz, a wealthyCalifornia computer geek, use us to bolster his own political career by dictating how we shouldeducate our children?” (Miller, 2002) In Massachusetts, Unz was criticized in 28 percent of theanti- letters. Writers commented on his millions, origins, questionable intentions, and perceivedanti-immigrant bias. A Dedham resident asks: “Why are we going to let Ron Unz, a Californiamillionaire with nothing else to do but further his wacky ideas, come to Massachusetts and tellus how to educate our children?” (Adams, 2002).

Legal issues. Legal aspects of the amendment were also present in appeals to fight theinitiatives. Over half of the anti- writers in Colorado used this theme. Writers expressed concernover teachers facing legal action for speaking a foreign language in the classroom. Anotherlegal issue involved the granting of waivers for certain students if their parents requested thatthey remain in immersion for more than 1 year. Upon granting such waivers, public employeessubjected themselves to lawsuits from parents for the next ten years (Citrin et al., 2003). As aresident of Castle Rock put it:

So let me get this straight: a parent can petition for his child to spend additional time in a bilingualprogram, but retains the right to sue the schools if the parent later regrets his own decision? Whata shame this ridiculous clause is part of the amendment. As a supporter of English immersion,I would have been happy to vote for provisions to provide only that (Reed, 2002).

Another legal angle to the debate questioned whether education policy should be a con-stitutional issue. Many letter-writers found this unacceptable. A resident of Aurora writes,“Fitting all school districts with the same constitutional straightjacket counters some of themost successful operating principles in this country, mainly decentralization and flexibility. . .Ifavor allowing the field commanders of our educational system—principals—to alter Englishacquisition programs to fit shifting student patterns” (Tamayo-Gomez, 2002).

Legal matters were important in Massachusetts as well. The most common legal pointwas that the battle over Q2 was yesterday’s war. Several months before election day the Mas-sachusetts state legislature had passed a bill, which was then signed by the Governor, to overhaulbilingual education statewide. The Larkin-Antonioni Law set new standards, redesigned report-ing structures and established a stricter limit on the number of years a child could be educatedwith native language assistance prior to mainstreaming. Many letters stated that these changeswere more moderate and reasonable than Unz’s plan, and should have been given a chance towork. Like in Colorado, many anti-Q2 letters expressed concern about subjecting teachers andschools to lawsuits through the new plan.

Page 16: Public opinion on bilingual education in Colorado and Massachusetts

386 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

Control. Control of the public schools is another key issue in the two states’ debates. TheColorado initiative proposed a constitutional amendment to supersede a previous clause dis-tributing authority over school decisions to the local school districts. Forty-five percent ofAnti-A31 letter writers argued that local autonomy was in jeopardy. Parental choice over howtheir kids would be educated was the primary concern among these writers, followed closelyby defense of local school or district choice over how to best teach students. Writers alsodefended teachers’ discretion in the classroom. 35 percent of anti-Q2 letters in Massachusettsraised issues of control. Although Massachusetts state law mandated that bilingual educationbe offered under certain demographic circumstances, in most cases parents had the power toopt for mainstream instruction for their children. The threat that Q2’s plan would infringe onparental choice was addressed by twelve percent of anti-writers. More common were concernsthat school districts, which enjoyed some instructional flexibility, would be tied to one programregardless of local children’s needs. 31 percent of the letters against Q2 expressed displeasurewith this loss of local independence. Teacher’s choice and choice in general were also defendedby letter writers.

Effects on non-LEP youth. Another theme relates to the effects on mainstream or Anglostudents. Twenty-nine percent of the Anti- letters in Colorado fall into this category. Concernover this issue area came in two versions. First, some parents with children in two-way bilingualschools saw that the proposal threatened these programs. A resident of Boulder appeals to voterson behalf of his kids: “My two Anglo daughters attend a bilingual elementary school whereEnglish and Spanish are taught at the same time. . .this flexibility of mind will help them inpractical ways in problem solving, for example. But being fluent in other languages and cultureswill also help my daughters become citizens of the world, which is what they will need to beif they will excel within a global economy” (Patterson, 2002).

The second form of protecting mainstream kids came through in a few Colorado letters.According to John Britz, a political strategist who consulted for opponents of the initiative, itstruck a chord with Anglo parents in Colorado (‘‘Colorado Hands”, 2002). The argument is thatpremature mainstreaming of non-proficient LEP kids would require regular teachers to focuson the English learners at the expense of native English speakers. This, it was argued, wouldlead to “chaos in the classroom.” A resident of Fort Collins advances this argument (while alsoevoking the control and cost themes): “[A31] takes away local control of our schools, it willcreate unnecessary disruptions to our children’s learning environment, and it will cost all ofus, as taxpayers, tens of millions of dollars. . .” (Yates, 2002). In Massachusetts, Anglo-relatedconcerns, particularly about two-way programs that benefit native English speakers and aboutimmersion’s negative effects on mainstream learning environments, appeared in twenty percentof the anti-letters.

Most opponents of the initiatives in Colorado and Massachusetts advanced arguments basedon the merits of the different educational plans. Legal themes and control issues were alsorelevant, and cultural symbolism and cost were significantly less salient. When comparingacross state debates, one can see that the educational program was a more salient issue inMassachusetts than in Colorado, though the difference is not statistically significant. Whatfactors did matter at statistically significant levels were the Effects on LEP youth (t = 2.52,significant with 98.7% confidence) and the Cost of different policy choices (t = 3.15, significantwith 99.8% confidence).

Page 17: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 387

Fig. 3.

8. Comparing two state debates

To some extent, the Colorado and Massachusetts debates differed thematically. Thoughcausal claims are beyond the purposes of this analysis, examining the distinctions between thetwo state debates can help to develop hypotheses for what types of frames benefit proponentsand opponents of bilingual education. Fig. 3 merges the pro- and anti-initiative letters together,revealing the salience of the themes in each state. The themes with asterisks by their names,Effects on LEP youth (t = 3.54, significant at 99.9% confidence) and Cost (t = 3.97, significantat 99.9%), are statistically significant at the 99% confidence level. The significance of twoother factors, Education plan (t = 1.87, significant at 93.8%) and Choice (t = 1.8, significant at92.6%) surpasses the less stringent 90% hurdle.

Rational principles were present in both debates. Cost is the most obvious interest-basedtheme, though legal issues and control over education have their rational elements. This findingsignals that pragmatic matters were believed to be at stake in both states. In comparative terms,the cost theme was considerably more important in the Colorado than in Massachusetts. Thissuggests that efforts to ban bilingual education do not necessarily benefit from an emphasison interest-based principles. Initiative advocates are better off where practical matters take abackseat to other considerations.

Page 18: Public opinion on bilingual education in Colorado and Massachusetts

388 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

Ethnic competition in the form of symbolic conflict was also evident in both debates; defenseof minority language rights clashed against nativist rhetoric in predicable ways. This supportsthe perspective that attitudes toward bilingual education can be explained, in part, throughexisting theories of ethnic conflict. Concerns that immigrants are not open to integration anddefensiveness of English as the dominant language are important aspects of public skepticismtoward bilingual education. However, although this finding helps to understand how manypeople reason on the topic of bilingual education, it does not effectively distinguish the debatesfrom one another. This theme appeared in a quarter of the Colorado letters and one-third ofMassachusetts letters, but the difference is not statistically significant at conventional cutofflevels. Therefore, its role in the divergent outcomes of the two statewide votes cannot be firmlyestablished.

Letter writers in Colorado and Massachusetts debate what is best for minority youth, whichsignals that paternalistic considerations existed in both states. The stated desire to do what’s inminority youths’ best interests was twice as prevalent in Massachusetts as it was in Colorado.This suggests that a potential reason for the success of the measure in Massachusetts (and itsfailure in Colorado) is related to the levels of ethnic paternalism in the different populations,or perhaps to the extent to which such sentiments were activated in the course of the debate. Itmay be that once the door is opened to the discussion of young English learners’ futures, thepublic is invited to engage in debate over what is best for them.

The most salient theme in both statewide debates was the educational merits of bilingualinstruction. In both contexts, the foremost consideration found in letters-to-the-editor washow effective different language programs are.5 Looking at the states together, we see thatthese themes are more relevant in Massachusetts, where eighty percent of letters debate theeducational issue. As noted above, this difference between the two states is not significant atthe 95% level, but does surpass the 90% confidence threshold. This suggests that a focus oneducational details might ultimately doom bilingual education programs.

The question is: why? Past research by Huddy and Sears (1990) at the micro-level suggeststhat the more people know (or think they know) about bilingual education, the more likely theyare to be against it. If we consider the implications of their finding for aggregate public opinion,it suggests that the more people think about the specifics of bilingual education (perhaps asa function of elite cues), the more resistance to such programs one will see at the state level.Furthermore, the paternalism argument advanced here connects these dots more directly. Whenregular people devote attention to the particulars of the bilingual education debate, they take onthe responsibility of determining what is best for immigrants. This, then, evokes paternalistictendencies as people seek to make the best choice. The lack of elite consensus on the initiativein Massachusetts compared to the more consistent negative messages transmitted in Coloradomight have played a role here. When faced with conflicting views, voters take it upon themselvesto determine the best course of action.

9. Discussion

When viewed together, the two statewide debates over bilingual education present an inter-esting puzzle. The initiative’s success in Massachusetts appears especially odd when compared

Page 19: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 389

with its failure in Colorado. One would have expected that on ethnic competition grounds alone,the Colorado measure had a much better chance of passing. Immigration rates, Latino popu-lation sizes, English language skills and state culture characteristics all predict to the oppositeoutcome.

This is where ethnic paternalism comes into play. The analysis here shows that althoughattitudes toward minorities were relevant in Massachusetts, they were cast in terms thatdo not conform to the conceptualization advanced in most minority politics studies. Themore commonly investigated ethnic competition motivations (Sears and Huddy, 1990; 1995)were less evident, and their presence did not distinguish the Colorado and Massachusettsdebates from one another. Indeed, among Massachusetts letter writers, paternalistic consid-erations were significantly more prevalent than ethnic conflict motives (t = 2.7, significantat 99.3%). This result supports the work of racial and ethnic scholars who find that pater-nalism is a pervasive, if understudied, dimension of out-group attitudes (Dietrich, 2005;Jackman, 1994). And though this finding is only a single piece of the puzzle, it is anextremely important one because it at once makes sense of the divergent outcomes and itadds an important dimension to out-group attitudes that has not been identified in existingstudies of minority politics attitudes in general. Public opinion surveys and experimentsshould in the future be designed to address a broader set of dispositions toward out-groups.

A factor that ethnic politics studies do typically examine and that mattered to some letterwriters in both states was the financial cost associated with different programs. And costwas one of the main themes distinguishing the Colorado debate from the Massachusettsdebate: it was a key concern in Colorado. This finding supports the argument that economicconsiderations matter for public support of bilingual education (Citrin et al., 2003; Huddy& Sears, 1995) and other programs associated with minorities, such as affirmative action(Branton & Jones, 2005; Kluegel & Smith, 1983). But what the above analysis shows is thatthe cost factor is more closely associated with the anti-bilingual education measure’s failure.This suggests that when rational principles are central to people’s considerations on minor-ity politics issues, and when there is a perceived cost to dismantling a program that benefitsminorities, the bulk of public opinion may swing in a way that allows such a program tocontinue.

This study also shows that the debates over bilingual education in Colorado and Mas-sachusetts included other themes that ethnic politics studies do not typically address. Thisresult supports the argument of Sniderman and Piazza (1993) that policy issues directedtoward minorities draw out different sets of dispositions and considerations dependingon the specific matter at hand. Through letters-to-the-editor people express their viewson bilingual education in terms of what they think about educational institutions, con-stitutional alterations and various interest groups among other topics. This diverse setof considerations suggests that we cannot understand public opinion on minority issues(whether at the individual or the aggregate level) solely in terms of the usual ethnic politicsframes.

Understanding majority opinion on minority issues is extremely important. Yet the toolswe typically use to examine public opinion are insufficient to handle the complexity andunpredictability of people’s judgments. The case of bilingual education opinions is especially

Page 20: Public opinion on bilingual education in Colorado and Massachusetts

390 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

complex, in part because the affected groups are recent immigrants who are often ineligible toparticipate in the decision-making. This is one of the intriguing aspects of deciding languageof instruction by popular vote: a vast majority of the decision-makers have no clear personalstake in the outcome. And the chasm between voters and recent immigrants is defined not onlyby race or ethnicity, but also age, education level and income (Jones-Correa, 2005; Michelson,2006; Rosenstone & Hansen, 1985). This gap makes ethnic paternalism all the more salient.Voting publics see it as their responsibility to help young English learners who cannot (andwhose families cannot) help themselves.

If Unz continues to be successful, bilingual education will emerge as a national issue.Therefore, it is important to understand public opinion as it relates to language of instruction.Though the debates do involve rational elements and ethnic competition aspects, they are alsodriven by deliberation over the educational proposals and arguments about what is best forLEP youth. Among the American public, there appears to be much interest in doing what isdeemed best for these students. And this paternalism tends to yield anti-bilingual educationpositions.

Notes

1. All data from the U.S. Census 2000: www.census.gov.2. Three letters did not take a clear stand on the issue, and so were excluded from the

analysis.3. To be certain that no one newspaper biased the content analysis, a series of bivariate

logit models were run to predict whether a letter was pro- initiative. The independentvariables were dummy variables for each newspaper. The Boston Herald (coeff.=1.2,significant at 97.7%) and The Boston Globe (coeff.=−.96, significant at 99.6%) werethe only papers with statistically significant effects. But their opposing signs and similarlysized coefficients suggest that they balanced each other out. No other newspapers hadstatistically significant effects.

4. A focus group participant in Massachusetts stated that the bilingual education issue wasthe most important element of the election except for the race for governor (Capetillo-Ponce & Kramer, 2006: 281).

5. Moreover, this approach to understanding the debates via letters-to-the-editor probablyunderestimates the importance of the educational proposals. As noted above, interviewsof newspaper editors reveal that letters discussing research or technical details are notvery popular. Editors prefer to publish pieces that evoke emotion or recount personalstories (Wahl-Jorgensen, 2001).

Appendix A. Categorization of sub-themes into themes

Five pro-initiative themes

(1) Costs:

Page 21: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 391

• bilingual education is expensive, waste of money;• budgetary concerns: social services are already strained;• taxes fund the schools;• bilingual education amounts to social welfare.

(2) Ethnic competition:• Americans should speak English;• learn English or leave;• bilingual education unfairly favors “new” (Spanish speaking) immigrants;• past waves of immigrants learned English without bilingual education;• testimonials from past immigrants who learned without bilingual education;• testimonials from people whose families learned without bilingual education;• bilingual education supports Hispanic culture;• immigrants need to assimilate.

(3) Effects on LEP youth:• bilingual education hurts kids’ futures;• bilingual education traps kids in language ghettos, linguistic underclass;• bilingual education forces LEP kids to be segregated;• bilingual education contributes to the high Hispanic dropout rate;• kids should learn English fast;• mainstreaming is important for LEP kids.

(4) Educational plan- bilingual education vs. structured immersion:• immersion works;• bilingual education doesn’t work;• positive results with immersion in California (or Arizona);• bilingual education is a failed “experiment”;• kids learn languages fast—immersion will work.

(5) Opposition:• teachers’ unions have own agenda;• educratic machine has its own agenda;• Democratic politicians support bilingual education to pander to the unions;• research used to defend bilingual education cannot be trusted;• bilingual education is a liberal cause;• opposition uses scare tactics;• opposition calls us racist;• opposition is disingenuous;• (in CO only) Stryker’s millions have tainted this contest.

Eight anti-initiative themes

(1) Cost:• moving over to immersion would be expensive;• bilingual programs not expensive.

(2) Ethnic competition:• we must preserve and protect immigrant cultures;

Page 22: Public opinion on bilingual education in Colorado and Massachusetts

392 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

• US is multi-lingual and multicultural;• language rights are being threatened;• proposal is anti-immigrant, anti-Hispanic;• proposal is discriminatory against language minorities.

(3) Effects on LEP youth:• immersion will limit the prospects available to kids;• immersion will segregate LEP kids from other students;• sink-or-swim methods are tough on kids emotionally, psychologically;• immersion will exacerbate high Hispanic dropout rate.

(4) Educational plan- bilingual education vs. structured immersion:• bilingual education works;• immersion doesn’t work;• negative results with immersion in California (or Arizona);• immersion is a one-size-fits-all program;• 1 year not long enough to learn English;• conversational vs. academic English are very different;• proposed plan is limiting and inflexible.

(5) Legal issues:• lawsuits against teachers are bad idea;• proposal is too legally punitive in nature;• this issue should not be up for public vote;• (in MA only) new legislation on bilingual education should be given time to

work;• (in CO only) educational policy should not be written into state constitution.

(6) Choice:• proposal threatens choice in general;• proposal threatens school, district choice;• proposal threatens teacher choice;• proposal threatens parental choice;• proposal undermines local control over schools.

(7) Opposition:• Unz is a racist;• Unz is a businessman, millionaire;• Unz is a Californian;• Unz/opposition is sneaky, has ulterior motives.

(8) Effects on native English speaking kids• proposal threatens 2-way bilingual programs;• immersion threatens native English speaking kids’ educational experiences.

Page 23: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 393

Appendix B. Breakdown of newspaper sources

# Anti-letters # Pro-letters #Letters/paper

Colorado Papers

Aspen Times 2 0 2

Denver Post 35 14 49

Douglas County News-Press 5 2 7

Fort Collins Coloradoan 11 2 13

Haxtun-Fleming Herald 1 0 1

Pueblo Chieftain 2 1 3

Rocky Mountain News 17 8 25

Westword (Denver) 2 1 3

Total # letters 76 28 104

Massachusetts papers

Beacon Hill Times 1 0 1

Boston Globe 31 7 38

Boston Herald 11 16 27

Cape Cod Times 0 3 3

Cavalier Daily 1 0 1

Christian Science Monitor (Boston) 4 2 6

Enterprise (Brockton) 2 0 2

Groton Landmark 0 1 1

Herald News (online) 2 0 2

Lowell Sun 7 3 10

Patriot Ledger Quincy) 5 2 7

Worcester Telegram & Gazette 1 8 9

Total # letters 65 42 107

References

Adams, Margaret. (2002). Letter to the editor. “Editorials,” The Boston Herald June 6, 032.Alston, L. J., & Ferrie, J. P. (1985). Labor costs, paternalism, and loyalty in southern agriculture: A constraint on

the growth of the welfare state. The Journal of Economic History, 45(1), 95–117.

Page 24: Public opinion on bilingual education in Colorado and Massachusetts

394 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

Bilingual (2002). “Bilingual Ed Advocates Marshal Forces State House Rally Sparks War of Words,” The BostonGlobe Oct. 1, 2002, p. B2.

Bombardieri, M. (2002). Chelsea: In this precinct, question 2 was on their minds. The Boston Globe, 2002(November(6)), B15.

Branton, R. P., & Jones, B. S. (2005). Reexamining racial attitudes: The conditional relationship between diversityand socioeconomic environment. American Journal of Political Science, 49(2), 359–372.

Buell, E. H. (1975). Eccentrics or gladiators? People who write about politics in letters-to-the-editor. Social ScienceQuarterly, 56(3), 440–449.

Burdan, W. (2002). Letter to the editor. “The Open Forum,” The Denver Post Nov. 3, E-03.Burnham-Massey, L., & Pina, M. (1990). Effects of bilingual instruction on English academic achievement of LEP

students. Reading Improvement, 27(2), 129–132.Calderon, R. (2002). Letter to the editor. “Amendment 31 an impossible task” The Aspen Times Nov. 1.

http://www.aspentimes.com.Capetillo-Ponce, J., & Robert, K. (2006). Politics, ethnicity and bilingual education in Massachusetts: The case

of referendum question 2.” Ch. 13. In Andrés Torres (Ed.), Latinos in New England. Philadelphia, PA: TempleUniversity Press.

Carlson, R. E. (2002). Montero’s my hero. The Denver Post, 2002(Sept. 29), E03.Chandler, C. R., & Tsai, Y.-M. (2001). Social factors influencing immigration attitudes: An analysis of data from

the General Social Survey. The Social Science Journal, 38, 177–188.Citrin, J., Reingold, B., Walters, E., & Green, D. P. (1990). The ‘Official’ English movement and the symbolic

politics of language in the United States. Western Political Quarterly, 43(3), 535–559.Citrin, J., Wong, C., & Duff, B. (2001). “The Meaning of American National Identity.” In: Ashmore, R., Jussim, L.,

& Wilder, D. (Eds.), Social Identity, Intergroup Conflict, and Conflict Reduction. New York: Oxford UniversityPress.

Citrin, J., Sears, D. O., Muste, C., & Wong, C. (2001). Multiculturalism in American public opinion. British Journalof Political Science, 31, 247–275.

Citrin, J., Kiley, J., & Pearson, K. (2003). Direct democracy takes on bilingual education: Framing the debate infour state initiatives. In Paper prepared for presentation at the Annual Meeting of the American Political ScienceAssociation Philadelphia, August 28–31,

Colorado (2002). “Colorado Hands English Immersion Backer his First Loss,” The Rocky Mountain News Nov. 6,29A.

Delgado, R., & Yun, D. H. (1994). Pressure valves and bloodied chickens: An analysis of paternalistic objectionsto hate speech regulation. California Law Review, 82(4), 871–892.

Dietrich, D. (2005). Attitudes concerning race and ethnic paternalism. In Paper presented at the annual meeting ofthe American Sociological Association Marriott Hotel, Loews Hotel, Philadelphia, PA, August 12.

Elazar, D. (1972). American federalism: A view from the states. New York: Thomas Y. Cromwell Co.Election (2002). Voters Decide on Bilingual Ed; Unz hopes ballot win spurs Congress,” The Boston Herald Nov.

6:07.Erickson, R., McIver, J., & Wright, G., Jr. (1987). State political culture and public opinion. The American Political

Science Review, 81(3), 797–814.Fitzpatrick, C. (2002). Letter to the editor. “Open forum on bilingual education; Immersion ignored,” The Denver

Post Sept. 29, E-03.Greene, J. P. (1997). A meta-analysis of the Rossell and Baker review of bilingual education research. Bilingual

Research Journal, 21(2–3), 103–122.Grey, D. L., & Brown, T. R. (1970). Letters to the editor: Hazy Reflections of public opinion. Journalism Quarterly,

47(1), 450–456.Hayward, Ed. (2002). “Voters decide on bilingual education, Unz hopes ballot win spurs Congress,” The Boston

Herald Nov. 6, sports 007.“Heated Battle over English Immersion Intensifies.” The Boston Globe November 3, 2002, p. B6.Hill, D. B. (1981). Letter opinion on ERA: A test of the newspaper bias hypothesis. Public Opinion, 45,

384–392.

Page 25: Public opinion on bilingual education in Colorado and Massachusetts

J. Fitzgerald / The Social Science Journal 48 (2011) 371–396 395

Houvouras, S. K. (2001). The effects of demographic variables, ethnic prejudice, and attitudes toward immigrationon opposition to bilingual education. Hispanic Journal of Behavioral Sciences, 23(2), 136–152.

Huddy, L., & Sears, D. O. (1990). Qualified public support for bilingual education: Some policy implications. TheAnnals, 508, 119–134.

Huddy, L., & Sears, D. O. (1995). Opposition to bilingual education: Prejudice or the defense of realistic interests?Social Psychology Quarterly, 58, 133–143.

Hurwitz, P. (2002). Letter to the editor. “Reform in name only,” The Boston Herald Oct. 2, 2002: 028.Jackman, M. R. (1994). The velvet glove: Paternalism and conflict in gender, class and race relations. University

of California Press.Johnson, G. G. (1957). Southern paternalism toward Negroes after emancipation. The Journal of Southern History,

23(4), 483–509.Johnson, G. (2002). Letter to the editor. “Immersion foes put up ‘red herrings,”’ The Rocky Mountain News Oct.

22, 39A.Jones-Correa, M. (2005). Language provisions under the voting rights act: How effective are they? Social Science

Quarterly, 86(3), 549–564.Kelley, K. (2002). Letter to the editor. Westword.com Aug. 15, 2002. http://www.westword.com/issues/2002-08-

15/news/letters.html. Accessed March 11, 2003.Kluegel, J. R., & Smith, E. R. (1983). Affirmative action attitudes: Effects of self-interest, racial affect, and

stratification beliefs on Whites’ views. Social Forces, 61(3), 797–824.Kurtz, M., & Cindy R. (2002). “Debate over immersion roils Latino community,” The Boston Globe Nov. 2, A1.Levin, H. (2002). “Letters to the editor: It was a vote to help immigrants.” The Boston Globe Nov. 8, A22.Lombard, E.E. (2002). “Immersion works,” The Boston Herald April 29: 020.Mariano, L. (2002). Letter to the editor. “Editorial,” The Patriot Ledger Nov. 1: 16.McQuillian, J., & Tse, L. (1996). Does research matter? An Analysis of media opinion on bilingual education,

1984–1994. The Bilingual Research Journal, 20(1), 1–27.Meyer, B. (2002). “Proposal would end ESL programs,” The Rocky Mountain News Oct 5: 24B.Meyers, M. (2004). Crack mothers in the news: A narrative of paternalistic racism. Journal of Communication

Inquiry, 28(3), 194–216.Michelson, M. R. (2006). Mobilizing the Latino youth vote: Some experimental results. Social Science Quarterly,

87(5), 1188–1206.Miller, K. (2002). Letter to the editor. “The open forum; Bilingual education: Of foes and fools,” The Denver Post

Feb. 17, E-03.Myrdal, G. (1964). An American Dilemma. McGraw-Hill Publishers.Parrilla-Sweeney, A. (2002). Letter to the editor. “Letters,” The Boston Globe Oct 20, D12.Patterson, E. (2002). Letter to the editor. “The open forum; flexibility of mind,” The Denver Post Sept. 29, E03.Perault, C. (2002). Letter to the editor. “Bilingual waste,” The Boston Herald Feb. 5, 020.Perlmann, Joel. (1990). “Historical legacies: 1840-1920.” In: Courtney B., Cazden and Catherine E. Snow, Eds.

“English plus: Issues in bilingual education,” The Annals of the American Academy of Political and SocialSciences 508: 27–37.

Reed, A.L. (2002). Letter to the editor. The Denver Post Oct. 6, E-03.Renfro, P. C. (1979). Bias in selection of letters to editor. Journalism Quarterly, 56(4), 822–826.Rosenstone, S. J., & Hansen, J. M. (1985). Mobilization, participation and democracy in America. New York:

MacMillan.Sanko, J. (2002). “Bilingual feud turns fiery; arsonists bury car of leader behind amendment 31,” Rocky Mountain

News Oct. 19, 4A.Sears, D. O., Van Laar, C., Carrillo, M., & Kosterman, R. (1997). Is it really racism? The origins of white Americans’

opposition to race-targeted policies. The Public Opinion Quarterly, 61(1), 16–53.Sigelman, L., & Walkosz, B. J. (1992). Letters to the editor as a public opinion thermometer: Martin Luther King

holiday vote in Arizona. Social Science Quarterly, 73(4), 938–946.Slavin, R. E, & Cheung, A. (2005). A synthesis of research on language of reading instruction for English language

learners. Review of Educational Research, 75(2), 247–284.

Page 26: Public opinion on bilingual education in Colorado and Massachusetts

396 J. Fitzgerald / The Social Science Journal 48 (2011) 371–396

Sniderman, P. M., & Piazza, T. (1993). The scar of race. Cambridge, MA: Harvard University Press.Soares, S. (2002). “Question 2 will do irreparable harm.” The Herald News Online, October 29,

http://www.heraldnews.com.Stebbins, P. (2002). “Immersion is best way to teach new language,” letter to the editor, Cape Cod Times Oct. 12.Tamayo-Gomez, Dr. D.D. (2002). “Amendment 31: Americans play the fool when we fear languages,” The Denver

Post Oct. 25, B07.Theiss-Morse, E. (2009). Who counts as an American? The boundaries of national identity. New York: Cambridge

University Press.Tracy, A. (2002). Letter to the editor. “The open forum bilingual education; two kinds of fluency,” The Denver Post

Oct. 6, E-03.Trepagnier, B. (2001). Deconstructing categories: The exposure of silent racism. Symbolic Interaction, 24(2),

141–163.Two States (2002). “2 States, 1 Issue, Unexpectedly Different Outcomes; Bilingual Initiatives in Colo. and Mass.

Affected by Funding,” Rocky Mountain News November 7, 2002. p. 28A.Van den Berghe, P. L. (1970). Race and ethnicity: Essays in comparative sociology. New York: Basic Books.Wahl-Jorgensen, K. (2001). Letters to the editor as a forum for public deliberation: Modes of publicity and democratic

debate. Critical Studies in Media Communication, 18(3), 303–320.Wahl-Jorgensen, K. (2002). Understanding the conditions for public discourse: Four rules for selecting letters to

the editor. Journalism Studies, 3(1), 69–81.Wren, C. (2002). Letter to the editor. “Read the fine print,” The Denver Post Sept. 12, C-06.Yates, A. (2002). Letter to the editor. “The devil’s in the details,” The Fort Collins Coloradoan Oct 27, B6.Zucker-Conde, L. (2002). Letter to the Editor “Traumatizing students won’t help.” The Boston Globe Oct. 6.