Bases for Defining Language Proficiency

The conventional rationale for curriculum planning (e.g., outlined in Tyler, Reference Tyler1949) is that educators determine the nature of educational programs first by establishing relevant objectives, then by creating and sequencing appropriate teaching and learning activities, and then later evaluating the extent to which students have achieved the objectives intended as a result of participating in the activities prescribed. Education in second and foreign languages, however, has often stood apart from this paradigm because schooling is not the sole (or necessarily the most effective) basis by which people develop language proficiency. People may develop language proficiency independent of schooling through interactions at home or in local subcommunities or while traveling. Moreover, many language programs are for mature learners, either in higher education or for settlement or work-related purposes, who tend to differ greatly in their abilities, backgrounds, and levels of language proficiency.

In these latter circumstances, evaluation assumes roles beyond the conventional function of determining students’ achievements directly from an educational program: Language assessment commonly serves purposes of screening applicants for admission to programs of higher education or suitability for employment or immigration, monitoring outputs or variability within educational systems, supporting independent learning, or accommodating specific populations. These latter functions of language assessment have become prominent over the past two decades for a variety of reasons: increased international migration and mobility, the globalization of economic and educational opportunities in respect to a few lingua franca languages (such as English, French, Arabic, Chinese, and Spanish), the emergence of local varieties of these languages, concerns for the survival of other minority languages, and calls for accountability in education (Canagarajah, Reference Canagarajah2005, Reference Canagarajah2006; Elder & Davies, Reference Elder and Davies2006; Shohamy, Reference Shohamy2001, Reference Shohamy, Cummins and Davison2007). In short, language assessment itself has become a major policy issue in educational programs.

Aligning Tests, Curricula, and Pedagogy

Most educational systems have, over the past two decades, established new curriculum policies that define standards for students’ intended achievements in languages in terms of benchmarks, competencies, or attainment levels. A guiding rationale for these policies is to make explicit the results expected for students’ achievements at specific points in educational programs so that these can be taught, studied, assessed, reported, as well as monitored and accounted for, it is hoped, with increased clarity, consistency, and precision (Brindley, Reference Brindley1998, Reference Brindley2000; Council of Europe, 2001; McKay, Reference McKay, Cummins and Davison2007; McKay et al., Reference McKay, Coppari, Cumming, Graves, Lopriore and Short2001; McNamara & Roever, Reference McNamara and Roever2006; Nunan, Reference Nunan, Cummins and Davison2007; Trim, Reference Trim1998). These policies can permit the alignment of curricula and assessment through criterion-referenced forms of assessment. That is, curriculum standards set “criteria” for teaching and for student evaluation (for placement, diagnosis, and achievement), rather than the normative-based approaches that have dominated the design and development of most major tests of language abilities, founded on piloting and analyses with populations outside of educational programs, which inevitably present disjunctures with educational programs (Brown & Hudson, Reference Brown and Hudson2002; Lynch & Davidson, Reference Lynch and Davidson1994; Gipps & Cumming, Reference Gipps, Cumming, Bascia, Cumming, Datnow, Leithwood and Livingstone2005). In turn, new curriculum standards can promote functionally oriented conceptualizations of language abilities for teaching and assessments, focused on communicative competence in languages or arising from analyses of optimal teaching and learning practices (Brown & Hudson; McKay, Reference McKay, Cummins and Davison2007; Johnstone, Reference Johnstone2000; North, Reference North2000; Nunan, Reference Nunan, Cummins and Davison2007; Purpura, Reference Purpura, Shohamy and Hornberger2008; Trim, Reference Trim1998).

Research into the implementation of these curriculum policies, however, has exposed a variety of dilemmas. One dilemma is the relative imprecision of descriptive criteria, rating scales, and even sample benchmarks in view of the complexity of languages and human behavior. Moreover, most curriculum standards have been established through professional judgment or consensus rather than systematic empirical or theoretical inquiry. This dilemma challenges the transposition of curriculum standards into assessment instruments, as reported in many projects to implement the Common European Framework of Reference for Languages (CEFR) (Alderson, Reference Alderson2005; Byrnes, Reference Byrnes2007). Indeed, variability tends to characterize teachers’ interpretations and uses of language standards, rather than the expected consistency (Alabau et al., Reference Alabau, Bonnet, de Bot, Bramsbye, Dauphin and Erickson2002; Brindley, Reference Brindley1998, Reference Brindley2000; Brindley & Burrows, Reference Brindley and Burrows2001; Cumming, Reference Cumming, Silva and Matsuda2001a, Reference Cumming2001b; Davison, Reference Davison2004; North, Reference North2000). Moreover, fears exist that assessments fixed to language standards narrow the scope of teaching and learning, as in education at large (Cummins, Reference Cummins2000; Leung & Lewkowicz, Reference Leung and Lewkowicz2006; Linn, Reference Linn2000; Hamp-Lyons, Reference Hamp-Lyons, Cummins and Davison2007; Hillocks, Reference Hillocks2002). In the United States, policies and legislation for No Child Left Behind have required simply that school systems demonstrate “improvement,” resulting in great discrepancies in the ways and extent to which states and school boards have specified and acted on curriculum standards and relevant assessments to meet this policy (Harper, Platt, Naranjo, & Boynton, Reference Harper, Platt, Naranjo and Boynton2007; Rivera & Collum, Reference Rivera and Collum2006). This raises a second dilemma concerning the appropriate scope and legitimate political jurisdiction of curriculum standards, particularly when standards extend across vast geopolitical domains (such as all of Europe) or there is competition between educational authorities, teachers’ professional associations, and local school or community concerns (Alabau et al., Reference Alabau, Bonnet, de Bot, Bramsbye, Dauphin and Erickson2002; Byrnes, Reference Byrnes2007; Canagarajah, Reference Canagarajah2005, Reference Canagarajah2006; Darling-Hammond, Ancess, & Falk, Reference Darling-Hammond, Ancess and Falk1995; McKay, Reference McKay, Cummins and Davison2007; Moore, Reference Moore2005; Shohamy, Reference Shohamy, Cummins and Davison2007).

A third dilemma is that successful adoption of new curriculum standards requires enormous investments in resources for teachers’ professional development (Brindley & Burrows, Reference Brindley and Burrows2001; Murray, Reference Murray2008; North, Reference North2000; Nunan, Reference Nunan, Cummins and Davison2007; Slayter, Reference Slayter2003; Short et al., Reference Short, Gomez, Cloud, Katz, Gottlieb and Malone2000; TESOL, 1998, 2001; Yoshida, Reference Yoshida2003). A final dilemma arises implicitly from conceptualizations of “dynamic assessment,” which argue, from principles of sociocultural theory, that language teaching and learning need to be conceived as integrally interactive, jointly constructed, and negotiated processes between teachers and learners, which cannot be prescribed or predicted by general curriculum policies (Lantolf & Poehner, Reference Lantolf, Poehner, Shohamy and Hornberger2008; Leung, Reference Leung2007; Leung & Mohan, Reference Leung and Mohan2004; Poehner & Lantolf, Reference Poehner and Lantolf2005).

Pedagogical Practices

These curriculum innovations and dilemmas have prompted recent research to begin to describe classroom practices for language assessment. Descriptive case studies and surveys have started to document the functions and formats of teachers’ assessment practices in language courses and in view of curriculum standards (Breen, Hird, Milton, Oliver, & Thwaite, Reference Breen, Hird, Milton, Oliver and Thwaite2001; Brindley, Reference Brindley2000; Colby-Kelly & Turner, Reference Colby-Kelly, Turner, Cumming and Laurier2007; Cumming, Reference Cumming2001b; Davison, Reference Davison2004; Edelenbos & Kubanek-German, Reference Edelenbos and Kubanek-German2004; Grierson, Reference Grierson and Brindley1995; Leung & Mohan, Reference Leung and Mohan2004; Rea-Dickins, Reference Rea-Dickins2001). A rallying slogan has been that assessment should be for learning (i.e., formative in nature) rather than of learning (i.e., not summative, as in conventional tests, even though few tests of language achievement actually exist—in the sense of being linked directly to curricula taught) (Colby-Kelly & Turner, Reference Colby-Kelly, Turner, Cumming and Laurier2007; Gipps & Cumming, Reference Gipps, Cumming, Bascia, Cumming, Datnow, Leithwood and Livingstone2005; Rea-Dickins, Reference Rea-Dickins, Cummins and Davison2007). Despite these exploratory accounts, teachers’ practices for responding to second-language students’ writing remains one of the few areas that has been studied with sufficient scope to foster general principles, and even these are tentative and controversial (Ferris, Reference Ferris2003; Goldstein, Reference Goldstein2005; Hyland & Hyland, Reference Hyland and Hyland2006). Another development has been renewed interest in promoting learners’ self-assessment, facilitated by the explicit descriptive criteria in curriculum standards (Alderson, Reference Alderson2005; Ekbatani & Pierson, Reference Ekbatani and Pierson2000; Little, Reference Little2005; Ross, Reference Ross1998). Studies investigating the potential washback of language tests on teaching and learning have proliferated, but evidence for direct causal impacts from assessments remains diffuse given the complex range of interrelated variables associated with assessment in language education (Bailey, Reference Bailey1996; Cheng, Reference Cheng2005; Cheng, Watanabe, & Curtis, Reference Cheng, Watanabe and Curtis2004; Shi, Reference Shi, Cumming and Laurier2007; Wall, Reference Wall2000; Wall & Horak, Reference Wall and Horak2006). Perhaps most revealing in this regard are studies of how language assessments or curriculum standards are interpreted and acted upon either by students (Huhta, Kalaja, & Pitkanen-Huhta, Reference Huhta, Kalaja and Pitkanen-Huhta2006; Johns, Reference Johns1991; Shi, Reference Shi, Cumming and Laurier2007) or by teachers (Brindley, Reference Brindley2000; Coniam & Falvey, Reference Coniam, Falvey, Cummins and Davison2007; Gardner, Reference Gardner2008; Wall & Horak, Reference Wall and Horak2006).

Accommodations for Particular Populations or Purposes

Recent emphases on language assessment and curriculum standards have also given rise to two types of concerns for particular populations. For minority-language children learning English in mainstream education, numerous analyses have documented the unfairness inherent in applying curriculum standards and assessments designed for majority populations, and various adaptations and accommodations have been recommended for minority students (Abedi, Hofstetter, & Lord, Reference Abedi, Hofstetter and Lord2004; Artiles & Ortiz, Reference Artiles and Ortiz2002; Bailey & Butler, Reference Bailey and Butler2004; Butler & Stevens, Reference Butler and Stevens2001; Cheng, Klinger, & Zheng, Reference Cheng, Klinger and Zheng2007; Cummins, Reference Cummins2000; McKay, Reference McKay2005; Garcia, McKoon & August, Reference Garcia, McKoon, August, August and Shanahan2006; McKay, Reference McKay2005; Riveria & Collum, 2006; Solano-Flores, Reference Solano-Flores2008; Solano-Flores & Trumbull, Reference Solano-Flores and Trumbull2003; Stansfield, Reference Stansfield2003). A wholly different concern arises in uses of language assessments to set minimum performance standards for certification for employment, professional practice, or curricula. In this vein, researchers have followed principles established in Cizek (Reference Cizek2001) to formally set performance standards as minimum benchmark levels on major language tests, for example, among teachers or health professionals (Coniam & Falvey, Reference Coniam, Falvey, Cummins and Davison2007; Epp, Stawychny, Bonham, & Cumming, Reference Epp, Stawychny, Bonham and Cumming2002; O'Neil, Buckendahl, Plake, & Taylor, Reference O'Neil, Buckendahl, Plake and Taylor2007; Xi, Reference Xi2007).

Summary

Language assessment has, in recent decades, become a major policy issue in education around the world, raising numerous fundamental issues about the functions, conceptualizations, and applications of assessment in respect to curricula, teaching, and learning. As educational programs adopt curriculum standards that define explicit criteria for students’ expected language achievements, various challenges emerge: in aligning curricula and tests or the compatibility of different sets of language standards, in describing or promoting optimal pedagogical practices and conditions for learning, and in accommodating unique student populations. More fundamentally, questions arise about how constructs of language proficiency should be defined, by whom, in what contexts, and on the basis of what evidence. Although innovative, criterion-referenced assessments serve pedagogical purposes within educational systems, the authority of norm-referenced language tests persists—particularly where benchmark norms are required apart from those defined by or within educational programs—for such consequential policy decisions as selection into programs, professional licensure, or monitoring educational systems themselves.