Comparing Patterns of Error in Generation 1.5, L1, and L2 First-Year Composition Writing

https://doi.org/10.1016/j.jslw.2016.11.002Get rights and content

Highlights

  • Generation 1.5 compared to L2 texts contained significantly fewer errors.

  • No significant differences between Generation 1.5 and L1 texts emerged.

  • The author suggests a shift from the term Generation 1.5 to resident L2 learner.

Abstract

The purpose of the current study is to determine whether language use patterns in early arrival Generation 1.5 writing more closely resemble the patterns demonstrated by L1 or L2 writers, or perhaps demonstrated patterns unlike either of the two groups. Writing samples were obtained at a South Texas 4-year public institution from First-Year Composition (FYC) classes, and essays were separated into 3 student groups for further analysis: early arrival Generation 1.5 (n = 133), L1 (n = 94), and L2 (n = 48) writing. Holistic writing quality and 11 error variables were compared across the 3 groups. Results indicate no significant holistic quality differences between early arrival Generation 1.5 writing and either L1 or L2 writing. In terms of error variables, no significant differences were found between early arrival Generation 1.5 and L1 writing; however, 9 of the 11 error variables demonstrated significantly lower rates of errors in early arrival Generation 1.5 writing when compared to L2 texts, with the same 9 error variables significantly lower in L1 writing compared to L2 writing. Implications are discussed generally as they relate to defining Generation 1.5 writers. Finally, following Roberge, Losey, and Wald (2015) and di Gennaro (2015), a shift away from the term Generation 1.5 is proposed.

Introduction

Second language (L2) writing research regularly makes distinctions between Generation 1.5 writers (i.e., resident L2 writers) and international student L2 writers (Belcher, 2012, Ferris, 2009, Ferris, 2011, Matsuda, 2008). The current state of the field conveys a strong sense that there are important differences between these groups, but that they are, nonetheless, both types of L2 writers. The sociocultural differences are well-documented and persuasive when comparing educational, cultural, and social backgrounds (Roberge, 2002, Roberge, 2009), but there is less consensus in terms of the patterns of difference in language use produced by these two groups. Before undertaking more applied research such as that related to pedagogy, it is worthwhile to establish expected and observable language use similarities and differences between resident L2, L1, and more traditional international L2 writing from a theoretical perspective. That is, before we begin from the premise that Generation 1.5 writers are a type of L2 student – and go on to explore issues such as pedagogy, curriculum development, and assessment – the more fundamental question of whether Generation 1.5 writers form a distinct group based on language use issues should be addressed. While student language use is only one piece of what informs pedagogy, curriculum development, and assessment, it is an important piece, as evidenced by the preponderance of scholarly work focusing on Generation 1.5 language use (Bunch, Endris, Panayotova, Romero, & Llosa, 2011; de Kleine & Lawton, 2015; Ferris, 2009; Reid, 2005).

This study advances our understanding of early arrival Generation 1.5 writing research by adding to a growing body of empirical research on differences and similarities between three student groups: early arrival Generation 1.5, L1, and L2 student writers enrolled in First Year Composition (FYC) classes, and improving on methods from Doolan (2013; previous research on this topic). For the purposes of this study, the following definition of an early arrival Generation 1.5 student will be used: (1) he/she regularly spoke a language other than English with his/her family (or in the house where he/she spent most of his/her childhood), (2) he/she has received five or more years of formal education in the United States, (3) he/she is less than 22 years old, and (4) he/she graduated from a U.S. high school or passed his/her General Education Development (GED) test (a high school equivalence test). This definition is largely consistent with (and even more narrow than) previous definitions of Generation 1.5 students (see Table 1). That said, because post hoc analysis of survey results revealed that 93% of the Generation 1.5 writers reported being in the U.S. educational system more than 10 years, it was determined that “early arrival Generation 1.5 student” was the more precise descriptor to use throughout the current study. Furthermore, the current study, along with all of the studies in Table 1, focus exclusively on U.S. undergraduate students, though important work with like populations at the kindergarden through 12th grade (K-12) level (Calderón & Minaya-Rowe, 2011) and internationally (Schecter, Arthurs, Sengupta, & Wong, 2015) could certainly inform larger discussions of resident L2 writing issues.

In essence, this study speaks to the theoretical issue of how to (and whether to) define subcategories within the larger category of the “L2 writer.” While applying labels of all kinds has the potential to “oversimplify within group distinctions” and overshadow individual differences, labels such as these are a well-established part of L2 writing scholarship and, when used responsibly, can have practical/pedagogical and political advantages (Ferris, 2013). Any discussion of categories necessarily involves generalization, which must be done responsibly and with respect for individual variation. In fact, even defining L2 writing is not entirely settled due to the complexity of accounting for such a wide range of individual differences (cf., the Disciplinary Dialogues section of the Journal of Second Language Writing, Atkinson, 2013). Yet, while the boundaries of L2 writing remain in flux, the three categories explored in this study (Generation 1.5, L1, and L2) have become rather common distinctions in the field of L2 writing studies (Ferris, 2009, Ferris, 2011; Roberge, Losey, & Wald, 2015). However, as these categories have become more commonplace, it remains an open question whether or not early arrival Generation 1.5 writers in undergraduate FYC classes produce writing with language use patterns that resemble more traditional L2 writing or if their writing can be distinguished empirically from the L1 writing.

One of the greatest challenges of conducting meaningful Generation 1.5 student writing research involves arriving at an agreed upon definition of Generation 1.5 students. Criticism of the term Generation 1.5 students are widespread, with the most prevalent critique being that the term is too amorphous and binds students to a label associated with language deficit (Benesch, 2008; Matsuda, Canagarajah, Harklau, Hyland, & Warschauer, 2003; Matsuda & Matsuda, 2009; Ortmeier-Hooper, 2008). As a result, many authors have shifted to other terms, at least in part to sidestep the criticisms leveled more directly at Generation 1.5.1 Of course, subtle differences do exist in defining Generation 1.5-like populations; however, the overlap in the many Generation 1.5-like terms is substantial. Issues of terminology will be revisited at the end of this study, but for this study, the term Generation 1.5 student will be retained because it has not yet been replaced by a widely agreed upon alternative.

In comparing definitions of Generation 1.5 students from previous empirical studies, four categories for group-inclusion are most prevalent: (1) a language criterion, (2) an educational criterion, (3) an age criterion, and (4) a time/residence in the United States criterion. Table 1 includes many relevant sources and criteria used. All cited studies identify students as speakers of non-English languages and almost all studies require Generation 1.5 students to have graduated from a U.S. high school. Very few studies include an age criterion, though median or mean ages (when included) indicate tertiary students of a traditional age range (e.g., students enrolling in post-secondary education beginning around 18 years old). Another benefit for including the age criterion in the current study is that it serves as a secondary indication that the Generation 1.5 student was enrolled in the U.S. K-12 educational system. Because the students must be less than 22 years old and in the U.S. educational system for five or more years, this ensures that they were in the U.S. educational system as minors. Requirements for time/residence in the United States fluctuate considerably, from none to seven or more years, though all but one study in Table 1 sets the minimum time/residence at three or more years. In light of these previous empirical studies on Generation 1.5 students, the definition adopted in the current study can clearly be seen as narrower than these previous Generation 1.5 student definitions.

To narrow the focus of the current study still further, data fitting the Generation 1.5 criteria for this study indicated 93% with 10 or more years in the U.S. educational system (see Appendix A for more survey information). In some cases, scholars have equated the term “Generation 1.5” with “long-term” in opposition to late-arriving resident L2 learners (Ferris, 2009), or immigrant students (de Kleine & Lawton, 2015). Yet, as the survey of empirical studies above suggests, students with as little as one year in the U.S. educational system (provided that they are high school graduates) could be considered Generation 1.5 students in much previous empirical research. As such, adopting the term early arrival Generation 1.5 writing in the current study more precisely describes the writing samples gathered herein. While some second language scholars may focus on the inclusion criteria of (1) students of a traditional age range, and (2) ten or more years in the U.S. educational system, and assume that English language proficiency would be advanced, perhaps citing the critical period hypothesis, the prevailing narrative of Generation 1.5 writing research is very much a story of long-term English learners. According to Calderón and Minaya-Rowe (2011), approximately 6,000,000 children in middle school and high school were classified as Long-Term English Learners in 2008. These students had been considered language learners since kindergarten. This number grew to 8,000,000 in 2009. A common definition of a long-term English learner in K-12 education is a student who has been in the U.S. educational system for more than six years, does not appear to be increasing language proficiency, and is struggling academically. Research efforts under the umbrella of Generation 1.5 writing does not typically include K-12 settings, but does extend this long-term language learner conversation into post-secondary settings. Yet, because experience in the U.S. educational system certainly can make a difference in ultimate attainment, efforts are made to recognize early arrival versus late arrival Generation 1.5 writing when sufficient information is available.

Generation 1.5 became a more commonly used term in L2 writing studies following work from the sociologists Rumbaut and Ima (1988) and writings about bilingual minority students by Valdés (1992) which foreshadowed the rise of immigrant L2 students in the U.S. educational system. The recognition of Generation 1.5 as a term associated with L2 writing is most often credited to edited volumes by Harklau, Losey, and Siegal (1999) and Roberge, Siegal, and Harklau (2009). Both of these major Generation 1.5 writing publications, along with the most recent work by Roberge et al. (2015) are teacher-driven (i.e., in response to teachers trying to more effectively cope with changing classroom demographics), and while Generation 1.5 students are by no means new to the U.S. educational system (Slager, 1956), this interest in Generation 1.5 writing was (and continues to be) an attempt to address the needs of an inadequately served segment of the post-secondary L2 writing classroom.

The prevailing narrative largely holds that though Generation 1.5 writers have spent considerable time in the U.S. educational system, they still demonstrate patterns of language use consistent with a form of language learning. That said, Generation 1.5 writers, when they begin their undergraduate studies, are presumably a distinct type of language learner, with both sociocultural and language use patterns that set them apart from both L1 and international L2 students (Ferris, 2009, Roberge, 2009). One influential hypothesis is found in Reid (2005), which claims that international learners are “eye learners,” having foreign language instruction that relies more heavily on reading and writing. In contrast, Generation 1.5 students are “ear learners,” with their oral and aural communication more heavily influencing the English acquisition process. For Reid, these ear learners could be born in the United States, or could have spent as little as three years in the U.S. educational system. This narrative has played an important role in attracting much needed attention to an underserved student population. Appropriately, as this area of interest has gained attention, it has developed a respectable number of empirical studies attempting to confirm, refute, and advance this area in both theory and practice.

The most robust area of research on Generation 1.5 writing and writers focuses on the sociocultural challenges and differences between Generation 1.5 students and more traditional L2 students (Roberge, 2009). Persuasive articles, chapters, and books have been published demonstrating challenging transitions between high school and college (Allison, 2009), identity-related challenges of negotiating multiple cultures (Chiang & Schmida, 1999), and tensions regarding placement in mainstream or ESL writing classes (Ortmeier-Hooper, 2008). This research speaks persuasively to the different educational and cultural backgrounds that Generation 1.5 students bring to post-secondary education when contrasted with more traditional, international L2 students.

While sociocultural research on Generation 1.5 students has developed a relatively robust and very persuasive body of research, a meaningful body of empirical research on the language use patterns of Generation 1.5 writing is only recently beginning to emerge. A relatively large number of theoretical publications and qualitative studies have pointed to linguistic development features as possibly distinguishing among student groups. Examples of these language features include those which are more typically associated with spoken language (e.g., contractions, that-deletion, personal pronouns) and written language (e.g., noun phrase incidence, post-nominal modification, nominalization). However, while a large amount of empirical research has been conducted in this area, a focus on individual linguistic features has not yet demonstrated consistent patterns of distinguishing groups of writers within the same writing class (see Jarvis, Grant, Bikowski, & Ferris, 2003 for a review of this literature). What limited work has been done with individual linguistic development features on large amounts of Generation 1.5 writing has also not distinguished writing well (Connerty, 2009, Doolan, 2011b, Doolan, 2013, Doolan, 2014). Investigations of error variables, however, have much more effectively distinguished Generation 1.5 from L1 and/or L2 writing.

Early research on error patterns in Generation 1.5 writing suggested that Generation 1.5 students in an Intensive English Program (IEP) produced higher composite error scores than international students (Muchisky & Tangren, 1999), but these Generation 1.5 students would more precisely be considered “late arrival” with often no more than three years in the U. S. educational system. Levi (2004) found that a group of L1 Spanish Generation 1.5 students (with an average age of arrival of 8.53) produced patterns of error for a range of variables (e.g., verb form, preposition, subject/verb, article errors) that were significantly greater than the L1 sample but significantly less than the L2 students. Early qualitative work by Frodesen and Starna (1999) describe two students who could also be considered late arrival Generation 1.5 students. Their written production also suggested error patterns such as verb tense, word forms, idiomatic usage, and appropriacy of function words contributing to below grade level writing in Generation 1.5 writing.

A few recent studies have confirmed these earlier works. A follow up study by Frodesen (2009) tracked a late arrival Taiwanese Generation 1.5 student whose college entrance exam contained errors in every sentence. Two studies by di Gennaro, 2009, di Gennaro, 2013 compared Generation 1.5 (with median length of residence at 5 and 6 years respectively) and L2 writing samples from college entrance exams using a RASCH analysis and found comparable error scores as measured by a holistic “grammatical control” variable. Mikesell (2007) focused exclusively on verb errors, but also found quantitative similarities between Generation 1.5 (with an average length of residence at 13.3 years) and L2 students, though the types of verb errors (when analyzed qualitatively) were different. Finally, a study by Doolan and Miller (2012) investigated Generation 1.5 writers, of whom, 83% had lived in the United States 10 or more years. When compared to L1 writers, these Generation 1.5 writers demonstrated significantly more errors in terms of verb errors, prepositional phrase, word form, and total identified errors. The 2012 study also highlighted qualitative differences in the types of errors present in a small sample of Generation 1.5 writing compared to L2 texts. Taken as a whole, these studies suggest that while Generation 1.5 and L2 writing may not demonstrate identical patterns of error, Generation 1.5 writing includes error patterns that more closely resemble the writing of their L2 than their L1 classmates.

Yet, not all research on Generation 1.5 writing has demonstrated language use patterns indicating that Generation 1.5 writers are clearly a type of language learner. In Doolan (2013), FYC writing from three institutions across the United States was compared across three student groups (Generation 1.5, L1, and L2) on a host of error and linguistic development variables. Of the Generation 1.5 students, 88% had been in the U.S. educational setting for more than 10 years. Writing procedures were well controlled and no significant differences were found between Generation 1.5 and L1 student texts. In contrast, Generation 1.5 texts, when compared to L2 text produced significantly fewer word errors, word class errors, and total identified errors. Investigating whether the findings from Doolan (2013) may extend to developmental writing classes, Doolan (2014) largely confirmed the 2013 study, this time demonstrating significantly fewer errors by Generation 1.5 writers when compared to L2 writers on word errors, word class errors, verb errors and total identified errors. A total of 86% of these Generation 1.5 students had been in the U.S. educational system for more than 10 years. The combined work of Doolan (2013) and Doolan (2014) suggests that the narrative of Generation 1.5 writers as a type of language learner, at least for primarily early arrival Generation 1.5 writers, may be less certain than previous studies have indicated. The study by Doolan (2013), however, was not without limitations. The error categories of word errors, verb errors, and word class errors are all composite variables (combined to improve the robustness of the statistical analysis), making it harder to decipher the impact of individual variables. Additionally, Doolan (2013) involved both paper and pencil and computer lab data collection, leading to questions about how differences in data collection may have affected written products. Finally, collecting data from multiple institutions raises questions about how different placement procedures at different institutions might compromise the comparability of the writing samples.

Further exploring the issue of comparability, it is worth considering the degree to which writing proficiency is being accounted for in discussions of Generation 1.5 writing research. Broadly speaking, there are countless examples of L2 writing research and L2 text analyses that provide no indication of holistic writing quality. Without any measure of L2 writing proficiency, scholars are left with a less methodologically rigorous understanding of how the writing from one group compares to the writing from another group. While holistic writing measures can rightfully be criticized on reliability and validity grounds, it seems beyond criticism that it is better to have some sense of writing proficiency than none at all. The RASCH analyses conducted by di Gennaro, 2009, di Gennaro, 2013 are admirable examples of measuring Generation 1.5 writing proficiency, and research by Doolan, 2011b, Doolan, 2013, Doolan, 2014 has provided at least holistic quality scores to give the reader a better understanding of writing proficiency prior to subsequent analysis. That said, most Generation 1.5 writing research does not provide any writing proficiency information, leaving the reader to speculate about Generation 1.5 writing proficiency as it compares to the writing of their classmates based on class placement, error patterns, sociocultural variables, or researcher/teacher/student impressions. Certainly the prevailing narrative of Generation 1.5 writing, at least at the FYC level, is that their writing is less proficient than the writing of their classmates.

The current study adds to the existing body of empirical work comparing early arrival Generation 1.5, L1, and L2 written error patterns at the FYC level by drawing on new data, and improved research methods from Doolan, 2013, Doolan, 2014). The research questions guiding the current study are:

  • Will holistic quality scores distinguish early arrival Generation 1.5 expository writing from L1/L2 writing across FYC classes?

  • Will error patterns distinguish early arrival Generation 1.5 expository writing from L1/L2 writing across FYC classes?

Section snippets

Method

The current study can be considered an “approximate replication” (Porte & Richards, 2012) of Doolan (2013). According to Porte and Richards, approximate replications become especially important when the original study “yielded unexpected or unexplained results” (p. 287), which was certainly the case for Doolan (2013) when one considers that almost all previous research on Generation 1.5 writing (discussed above) had found differences between L1 and Generation 1.5 writing. As such, the following

Results

Using the research questions to guide the analysis, inferential statistics were employed to determine whether differences among the three student groups emerged in terms of language use. Results are presented for each research questions.

Discussion

The purpose of the current study was to determine whether language use patterns in early arrival Generation 1.5 writing more closely resemble the patterns demonstrated by L1 or L2 writers, or perhaps demonstrated patterns unlike either of the two groups. Results from the current study suggest that, in terms of holistic quality, early arrival Generation 1.5 texts were not significantly different from the writing of their L1 or L2 classmates. Additionally, at least for this data set, early

Conclusion

The current study provides additional evidence that early arrival Generation 1.5 writers, at least broadly defined, produce writing with holistic quality similar to both their L1 and L2 classmates, but error patterns that more closely resemble the writing of their L1 than their L2 classmates. The fundamental lesson from these findings is a theoretical issue of more precisely defining the student population. This foundational issue should be resolved prior to discussions of placement and

Acknowledgements

Thank you to Dr. Shannon Fitzsimmons-Doolan for providing thoughtful feedback on this manuscript. A sincere thanks to the blind reviewers and editorial staff for suggestions that improved the quality of this manuscript. I thank Texas A&M University—Corpus Christi for internal funding in support of this research. I also extend my appreciation to the raters and coders who assisted in this analysis, and finally, thank you to my students for helping advance my thinking of L2 writing.

Stephen Doolan is an Assistant Professor of English at Texas A&M – Corpus Christi. Dr. Doolan’s areas of expertise and interest include Generation 1.5 writing, L2 writing, and the reading – writing connection. His recent publications on Generation 1.5 writing in Written Communication (2013, 2014) won the John R. Hayes Award for Excellence in Research and honorable mention for the same award in the following year. His work has also appeared in the CA-TESOL Journal and the Journal of Second

References (52)

  • D. Belcher

    Considering what we know and need to know about second language writing

    Applied Linguistics Review

    (2012)
  • S. Benesch

    Generation 1.5 and its discourses of partiality: A critical analysis

    Journal of Language, Identity, & Education

    (2008)
  • G.C. Bunch et al.

    Mapping the terrain: Language testing and placement for US-educated language minority students in California’s community colleges

    (2011)
  • M.E. Calderón et al.

    Preventing long-term ELs: Transforming schools to meet core standards

    (2011)
  • Y.D. Chiang et al.

    Language identity and language ownership: Linguistic conflicts of first-year university writing students

  • M. Connerty

    Variation in academic writing among Generation 1.5 learners, native English-speaking learners and ESL learners: The discoursal self of G1.5 student writers

    (2009)
  • C. de Kleine et al.

    Meeting the needs of linguistically diverse students at the college level

    (2015)
  • K. di Gennaro

    Investigating differences in the writing performance of international and Generation 1.5 students

    Language Testing

    (2009)
  • K. di Gennaro

    The heterogeneous second-language population in US colleges and the impact on writing program design

    Teaching English in the Two-Year College

    (2012)
  • K. di Gennaro

    What do they mean?: Comparing international and U.S. resident second language students’ use of sociopragmatic markers in writing

    Writing & Pedagogy

    (2015)
  • S.M. Doolan

    A comparison of Generation 1.5, L1, and L2 tertiary student writing

    (2011)
  • S.M. Doolan

    A language-related comparison of Generation 1.5 and L1 student writing

    The CATESOL Journal

    (2011)
  • S.M. Doolan

    Generation 1.5 writing: The relationship between self-reported social factors and textual variables

  • S.M. Doolan

    Generation 1.5 writing compared to L1 and L2 writing in first-year composition

    Written Communication

    (2013)
  • S.M. Doolan

    Comparing language use in the writing of developmental Generation 1.5, L1, and L2 tertiary students

    Written Communication

    (2014)
  • D.R. Ferris

    Teaching college writing to diverse student populations

    (2009)
  • Cited by (0)

    Stephen Doolan is an Assistant Professor of English at Texas A&M – Corpus Christi. Dr. Doolan’s areas of expertise and interest include Generation 1.5 writing, L2 writing, and the reading – writing connection. His recent publications on Generation 1.5 writing in Written Communication (2013, 2014) won the John R. Hayes Award for Excellence in Research and honorable mention for the same award in the following year. His work has also appeared in the CA-TESOL Journal and the Journal of Second Language Writing.

    View full text