Language learning and study abroad

Abstract

The aim of this review is to synthesize empirical studies on undergraduate language learners’ experience abroad during a time period of a year or less. To help provide a framework to this synthesis, we begin our review by tracing the recent evolution of empirical mixed-method research on the learner, identifying problems and characteristics that language learners generally encounter in the study abroad (SA) experience. We take a closer look at variables related to individual difference such as anxiety, motivation, and attitudes to more recent views of learner identity in language learning. We highlight the shift to language learner agency, a topic that merits more discussion in SA literature. We then review how the SA learning environments are treated. This review takes a closer look at research informed by socially grounded theories. Finally, we review the role that SA plays in undergraduate language curricula, where the objectives of the experience are aligned with at-home (AH) curricula, a topic that has not been fully discussed in SA literature. The conclusions offer suggestions for keeping pace with the broader field of applied/educational linguistics.

1. Introduction

Generated, perhaps, by the priority of institutions of higher education (HE) worldwide to prepare students to be global citizens in the international community, ‘[t]he global population of students who move to another country to study continues to rise . . . with an annual increase of 10%’ (International Trends in Higher Education 2015: 5). The internationalization of HE has been underway since the Middle Ages in Europe (de Ridder-Symoens 1992), having paused in the eighteenth and nineteenth centuries (de Wit 2009) and experiencing a rebirth after the two World Wars. Lewin (2009) traces the ideological history of SA during these time periods and he concludes that the current goals of SA go beyond language learning to the internationalization of students.

According to Knight (2008), ‘“Internationalization” is not a new term. [It] has been used for years in political science and governmental relations, but its popularity in the education sector has really soared only since the early 1980s’ (p. 12). The term Internationalization has existed since the 1970s and ‘was defined in terms of activities such as SA, language studies, institutional agreements, and area students’ (p. 13). Internationalizing campuses is interpreted differently in various geographical areas. The International Association of Universities (IAU 2014: 9) states that regional difference exists worldwide with regards to the importance of internationalization and expected benefits:

Students’ increased international awareness is the top-ranked benefit of an international experience abroad in Asia, the Pacific and North America. In Europe and the Middle East, the top-ranked benefit is improved quality of teaching and learning. For African respondents, the top-ranked benefit is strengthened knowledge production capacity, and for Latin American and the Caribbean institutions, the highest-ranked benefit is increased networking of faculty and researchers.

The nature of SA, the desired development outcomes associated with it, and the research itself are shaped by broader realities and sociopolitical agendas. As Kinginger (2009) pointed out in her critical review of the research literature on SA, apprehending SA in its entirety is extremely difficult. In fact, even the terminology associated with SA poses challenges to scholars who seek to provide even a broad overview. Coleman (1997) notes the following terms used in connection with the phenomenon: SA, residence abroad, student mobility, in-country study, overseas language immersion (IM), and academic migration. In addition, King & Raghuram (2013: 129) pose that the phenomenon of international student migration/mobility can be differentiated among other vectors as well, including the length of SA, the level of study (e.g., undergraduate vs. graduate), the participants themselves, and the experiences. Each of the above-mentioned terms describes particular types of study at foreign institutions, ranging from the short-term SA programs (four to 14 weeks) that are prevalent within the US, to long-term matriculation in a foreign institution, prevalent in countries such as Sri Lanka and Cyprus, where demand for HE outstrips supply (OECD 2009). Most students from Asia and Eastern Europe who choose to study abroad seek degrees in foreign universities, whereas only 2% of postsecondary students in Western Europe enroll in foreign degree programs (Kinginger 2009). SA in Europe, North America, and Mexico predominantly involves programs of less than one year in duration.

The data provided by UNESCO (2014) on the global flow of tertiary-level students helps to explain three trends in where students choose to go. The first trend points to the reality that mobile students overwhelmingly tend to study in countries where their first language (L1) is one of the official languages, implying little to no need to acquire a foreign language (FL). The next trend is to study in a destination country where English is one of the official languages, predominantly the US, UK, Canada, and Australia. This may be explained by the globalization processes implemented in many secondary and tertiary institutions worldwide. The globalization initiative, according to Haberland & Mortensen (2012), has ‘led to a situation where English has attained an unprecedented degree of globality’ (p. 1) and may explain the overwhelming presence of English-language education in language education policies and planning for many non-Anglophone countries (Kaplan & Baldauf 2003). This tendency can be clarified by the fact that the lingua franca of the natural sciences has become English (Lillis & Curry 2010; Mortensen & Haberland 2012; Söderlundh 2012), as compared to the beginning of the twentieth century when the lingua franca was German (Hamel 2013: 327; Kaypak & Ortaçtepe 2014).

The third trend includes students who choose to study in a country where the target language (non-English) is different from their L1. These students appear to be from countries where the FL education policies, at the secondary and tertiary level, may include a learning objective of acquiring a second language (L2) other than English. Research falling under the first trend will not be addressed in this paper, since the focus is neither on language acquired nor on how various factors might influence language use or acquisition. Yang (2016) recently highlighted that the majority of the SA research (49.2% of what Yang reviewed) falls under the second trend, though not necessarily labeled SA, and is typically reported in journals dedicated to the teaching and learning of English as a second or foreign language (ESL) or in publications addressing HE experiences, psychology, intercultural development, or other areas.

A fourth trend is that of heritage seekers, who study abroad because of an ethnic, (ethno)religious, linguistic, or national connection to a specific ancestral country or region, for example, African American students studying in Ghana, Jewish American students studying in Israel, and Mexican-American students studying in the Spanish-speaking world. See Shively (2016) for a review of research on heritage seeking in SA.

In the overview of educational policies regarding the implementation of an L2 within curricular program planning, the policies must be analyzed in the context of globalization. It appears that globalization, instead of being an advantage for promoting L2 learning and therefore promoting student mobility worldwide, in fact promotes English-language acquisition and thereby advantages students in the new global networks in which English has not only become the lingua franca but has also benefited from its value as a ‘form of cultural capital’ (Short et al. 2001).

Regardless of how internationalization is valued in respective countries, the OECD projects that participation in SA is likely to reach 8 million students per year by 2025, up from 5 million reported in 2014 (ICEF Monitor 2017). This growth is accompanied by a corresponding expansion of studies researching the benefits of studying abroad. The SA experience that is of focus in this review is one that involves language learning during a time period of a year or less, in which the student is enrolled in a host institution taking courses related to the target language and, mostly, in the target language. As will be shown in the review of literature in the following sections, research in this field focuses on what learners do differently during SA and how those actions affect their acquisition of the target language and has shown that not all experiences are equivalent. The dominant tendency has been to focus on second language acquisition (SLA), researching mainly developing bilingual participants in a target language community. Participant characteristics vary widely when we take into account aspects such as affective variables and preprogram proficiency levels. The context within which data are collected and methods are used can vary just as widely, not to mention the numerous aspects of linguistic competency that can be measured. In addition, the outcomes of SA are documented in various domains, including international mobility, ESL, and HE policy, to name a few. SA outcomes are also included within frameworks of intercultural competence development that provide evidence for theoretical models that clarify issues of intercultural sensitivity (see Chen & Starosta 1997) as well as intercultural transition. According to Zhou et al. (2008), the latter can be categorized into three theories: stress and coping model, culture learning model, and social identification theory. These theories from other fields highlight the complexity of SA research and underscore the challenge inherent in compiling an overview of the available SA literature.

The need for greater understanding of this experience has inspired a growing interest in mixed-methods approaches that go beyond measuring the linguistic development of the language learner to include research on changes in learner identity and agency and student perspectives about language learning that inform development of intercultural/transcultural competence (e.g., MLA 2007; Jackson 2010, 2013; Beaven & Spencer-Oatey 2016). Looking beyond linguistic gains abroad has also brought attention to a gap between the curriculum of SA programs and those back at home, revealing a need to calibrate traditional classroom goals with those skills needed to negotiate life abroad (Lord & Isabelli-García 2014). Relatedly, there is considerable interest in the pedagogical enhancement of language learning in SA (Lafford & Isabelli in press).

The aim of this review is to synthesize empirical studies on undergraduate language learners’ experience abroad during a time period of a year or less. To help provide a framework to this synthesis, we begin our review by tracing the recent evolution of empirical mixed-method research on the learner, identifying problems and characteristics that language learners generally encounter in the SA experience. We take a closer look at variables related to individual difference such as anxiety, motivation, and attitudes to more recent views of learner identity in language learning. We highlight the shift to language learner agency, a topic that merits more discussion in SA literature.

We then review how the SA learning environments are treated. Though SA is at times treated as a categorical label, recent studies situate SA as a Social Ecosystem (de Bot, Lowie & Verspoor 2007) where research interprets interacting, internal dynamic subsystems to explain the L2 learning process. This review takes a closer look at research informed by socially grounded theories (e.g., Pavlenko & Lantolf 2000; Lantolf & Pavlenko 2001; Block 2003). Finally, we review the role that SA plays in undergraduate language curricula, where the objectives of the experience are aligned with AH curricula, a topic that has not been fully discussed in SA literature. The conclusions offer suggestions for keeping pace with the broader field of applied/educational linguistics.

2. Research design

Research in the field of SA has primarily centered on the student, with earlier methodological approaches exploring the participants’ experiences. These mostly qualitative studies had, and still have, as the goal to ‘listen to participants and build an understanding based on what is heard’ (Creswell 2013: 29). As more research contributed to the conversation on the topic, we saw the application of more quantitative methods where the ‘researcher uses the [previous research] deductively as a framework for the research questions or hypotheses’ (p. 29). By 1989, educators Green, Caracelli & Graham challenged researchers to provide a conceptual framework for the, at the time, recent expansion of ‘methodological repertoire that include the use of both qualitative and quantitative methods’ (p. 255). Since then, mixed-method evaluation has gained a stronghold in SA research designs. Most of the mixed-method designs fall under Creswell & Garrett's (2008) definition of ‘collecting both quantitative and qualitative data (in response to quantitative and qualitative research questions), the merging, linking, or combining the two sources of data, and then conducting research as a single study or as longitudinal projects with multiple phases’ (p. 326).

In the following sections, we address how these three methodological approaches have helped move the field forward.

2.1 Quantitative research

Quantitative research, by definition, involves numbers: scores, statistics, and correlations. Broadly speaking, quantitative methods involve the ‘tallying, manipulation or systematic aggregation of quantities of data’ (Henning 1986: 702). Its main purposes are to examine patterns in the data and to determine relationships among variables. Through quantitative analyses, researchers can determine what is typical and how much variation there is in the results, e.g., ‘What is the average change in listening test scores over SA?’ or ‘What is the range of performance on post-SA proficiency measures?’ Quantitative methods are particularly suited to exploring relationships between phenomena, e.g., ‘What is the relationship between preprogram proficiency and development of intercultural sensitivity?’ or to questions testing hypotheses, for example, ‘Does strategy training improve student outcomes on SA?’ Research exploring relationships is referred to as ‘associational,’ ‘correlational,’ or ‘survey’ (Creswell 2013). The goal of associational research is to determine whether a relationship exists among factors and, if so, to measure the strength of the relationship. Researchers undertaking associational studies are not interested in controlling every variable possible to determine causation, but instead they focus on correlation, or co-occurrence of variables.

In contrast to correlational studies, the goal of experimental studies is to establish causation as clearly as possible. In experimental studies, researchers deliberately manipulate variables in order to determine the effect of one variable on another. A true experiment must contain both a control group (receiving no treatment) and a treatment group (exposed to the intervention) to which participants must be randomly assigned (Creswell 2013). However, in SA research, true experiments can be very difficult to arrange. Random assignment of participants to one group or another is nearly impossible due to the expense involved in SA. Furthermore, student responses to being assigned without choosing SA or formal instruction at a home institution (AH) experiences can lead to motivational challenges. Huebner (1995) is the only study in this report involving an experimental design, utilizing randomly assigned control (AH) and experimental (SA) groups participating in 9-week programs comparable in classroom hours, pedagogical approach, and so on. Five out of 23 of Huebner's participants dropped out during the study, with motivation being a key factor, leaving only ten abroad and eight AH. Small numbers like these make it difficult to draw conclusions that can be applied beyond the group of participants. While Huebner's SA learners out-gained AH controls on virtually every measure of language gain, these differences were not statistically significant. Huebner writes, ‘With small numbers of students, tests of statistical significance have very low power.’ Therefore, ‘how the results . . . can be generalized to the possible performance of future students is the central question, which sadly cannot adequately be answered by the inferential statistical methods employed’ (pp. 183–184). Due to the challenges of random assignment and creation of comparable control groups of adequate size to draw inferences, most attempts at experimental research in SA are, therefore, best categorized as either ‘pre-experimental’ or ‘quasi-experimental.’

Pre-experimental research is so named because the design involves a treatment given to one group as in an experimental study but does not include a control group (Creswell 2013). The one-group pretest and posttest design that is so common to SA research thus qualifies as ‘pre-experimental.’ Examples of this method include Martinsen's (2010) study of the development of learners’ spoken Spanish following a 6-week SA program or Dewaele, Comanaru & Faraco's (2015) study of the effect of SA on the anxiety levels and willingness to communicate of students. Pre-experimental designs are typically thought to be useful for generation of hypotheses that can be tested later with more rigorous methods, but they are often the only option SA researchers have and seldom lead to more carefully controlled follow-up studies.

Quasi-experimental design involves a treatment group and a control group, but the groups are not randomly assigned (Creswell 2013). SA research comparing the linguistic skills of self-selecting students on SA (the treatment) and control students enrolled in formal traditional instruction AH or in domestic IM programs thus represents ‘quasi-experimental’ work. Examples of such studies include Dewey's (2004) comparison of gains in reading comprehension of learners of Japanese in SA and IM settings. Similarly, Cubillos, Chieffo & Fan (2008) compare listening comprehension results for American learners of Spanish in Spain and Costa Rica with students enrolled in a course of similar length at home, and Juan-Garau, Salazar-Noguera & Prieto-Arranz (2014) contrast the overall L2 English competence, motivation, and beliefs of comparable AH and SA learners. Taking a different quasi-experimental approach, other studies compare SA students receiving a specific treatment such as homestay placement (Vande Berg, Connor-Linton & Paige 2009), work placement (Mitchell, Tracy-Ventura & McManus 2015), or pre-departure orientation (Vande Berg et al. 2009) to control students not receiving the treatments. These studies improve on pre-experimental designs in that they can give a stronger indication that the treatment (SA, homestay placement, etc.) is effective. As Larsen-Freeman & Long (2014) note, ‘Quasi-experiments exist as compromises for those interested in studying human behavior in naturally occurring settings in which complete experimental control is difficult, if not impossible’ (p. 22).

There are considerations in quantitative research other than selection or assignment of participants. One such consideration is timing of testing. Mackey & Gass (2005) describe the following common timing pattern: pretest/posttest, in which some aspect of a learner's language is tested both before and after SA to determine whether SA resulted in any changes. Much of the SA research employs this method, either within a single group or between SA and AH groups (comparing changes from pre to posttesting over time) for each. The Martinsen (2010) pre-experimental study referenced above follows a pretest/posttest approach.

Pretest/posttest is a form of a repeated measures testing design. Repeated measures (also referred to as within-subject) involves testing the same individuals more than once—either testing the same person on multiple occasions (time-series) or administering multiple measures to the same people. Repeated measures can be used in both pre-experimental and quasi-experimental designs. McManus, Mitchell & Tracy-Ventura's (2014) study, comparing the language gains of students in various placement types, involved both multiple measures of linguistic abilities and multiple occasions over a 23-month period. Similarly, Trenchs-Parera & Juan-Garau (2014) administered a questionnaire on learners’ motivations and beliefs related to language learning four times over the course of learners’ university experience: (1) upon entry to university, (2) after two terms of formal instruction at university and prior to SA, (3) immediately following SA, and (4) 15 months after SA, following 80 hours of additional hours of formal L2 instruction. This longitudinal/time-series study indicated that the initial formal instruction ‘triggered positive changes in self-confidence and intended effort, while SA heightened the development of positive motivational stands, including the reduction of anxiety and the attribution of importance to listening abilities’ (p. 259). These changes held constant following the 15-month post-SA period. The only additional significant change from entry to university to final administration was a drop in the perceived importance of autonomous learning through self-access materials. In short, this study indicates how motivation and beliefs can change in different ways following formal instruction and SA experiences.

Factorial design is a measurement approach in which more than one independent variable and any number of moderating variables (variables that can affect the relationship between independent and dependent variables) are considered. Dewey et al.’s (2014) exploration of factors that influence L2 use during SA employed a factorial design. The researchers examined a variety of variables, including preprogram proficiency, gender, age, and social networks (independent variables), to determine their potential effects on the students’ reported use of the L2 (dependent variable) during SA.

One finding that is apparent in nearly every quantitative study is large amounts of variation (numerically evident in high ranges and standard deviations). Because SA experiences are so varied and complex, quantitative research alone cannot adequately examine them. The research has also consistently shown that individual differences play a prominent role in language learning—or lack thereof—during SA. Quantitative literature has attempted to explicate variables that may affect the differential outcomes associated with SA, but this approach does not always succeed in isolating and explicating variables and can reduce learners’ identities and experiences to an array of variables such as proficiency, motivation, personality, and time on task—measured psychometrically and represented as numbers. Such quantitative studies can give only an incomplete picture of SA, providing rudimentary and at best numerical attention to learners’ motives or dispositions, or to the nature of their experience while abroad. As Kinginger (2009: 68) states, ‘In research attempting a maximum level of generalization to whole populations, a recurrent finding is of striking individual differences whose explanation offers only tantalizing glimpses into the nature of SA experiences.’ Coleman (2013a: 25) agrees, adding, ‘individual trajectories are in fact the essence of recent SA research, in which the focus has shifted from quantitative to qualitative, from product to process, from a search for generalizability to a recognition of complexity and variation.’

Thus, while quantitative studies of SA can answer a number of questions about trends across learners and relationships among factors, qualitative accounts of the SA experience are uniquely positioned to describe the ‘nature of the SA experience.’ In the next section, we provide an overview of qualitative research, its goals, and its methods.

2.2 Qualitative research

A broad definition of qualitative research is ‘any research that uses data that do not indicate ordinal values’ (Nkwi, Nyamongo & Ryan 2001). Qualitative research is the ‘method of choice when researchers seek to understand processes, events, and relationships in the context of the cultural and social situation’ (Sullivan & Ebrahim 1995: 196).

The advantage of qualitative approaches is that they allow scholars to examine SA in all its complexity, exploring multifarious recursive interactions between students’ motivations, identities (Kinginger 2004; Isabelli-García 2006; Kinginger 2008; Allen 2010b), their reception by members of the host culture (Pellegrino 2005; Isabelli-García 2006), their attitudes toward the host culture (Isabelli-García 2006; Kinginger 2010), and their evaluations of the SA experience (Menard-Warwick & Palmer 2012). Certain methods of qualitative research also allow researchers to examine the place of SA within the greater ‘ideological environment’ (Kinginger 2004: 222) of FL learning (Gore 2005; Trentman & Diao 2015).

Qualitative research can be conducted in dozens of ways, some methods with long histories behind them. Among the most prominent in SA research are case study, grounded theory, narrative analysis, discourse analysis, and conversation analysis. Each tradition differs in terms of research purposes, types of data collected, and, to some extent, methods of data collection and analysis. Van Lier (2005) offers a useful paradigm for understanding differences in qualitative data analysis. He suggests that qualitative methods differ in the amount of ‘intervention’ specified in the research design, as well as in the degree to which the method is concerned with the individual or with groups. By intervention, we will understand the extent to which researchers work with existing data (including observation of naturally occurring interactions) or elicit data via interviews or journals collected specifically for the purposes of research. Ethnography, some forms of conversation analysis, and discourse analysis involve little intervention, focusing instead on collecting or analyzing existing data. Methods like grounded theory, case study, and biography usually elicit data. In terms of the second dimension, ethnography tends to be more concerned with groups, while biography and narrative analysis are frequently, though not exclusively, focused on individuals; and methods like grounded theory and case study can focus either on groups or on individuals.

To van Lier's (2005) two dimensions, we add a third—the extent to which data are analyzed linguistically or analyzed for content. In the former, researchers use the data obtained from learners to examine evidence of linguistic or pragmatic development, as in discourse and conversation analysis. In the latter, researchers are concerned not with how the message is encoded, but rather with what the individuals are saying, as in much case study research. In fact, when researchers are primarily interested in the message rather than the medium, the language of interviews and journals is frequently, though not always, in the native language (L1), so as to allow learners to adequately express their thoughts and emotions.

When texts are analyzed linguistically, rather than for content, researchers use methods such as discourse analysis or conversation analysis to analyze samples of natural language. Such studies seek to document language socialization, the nature of learner interactions, and linguistic development over time.

Documenting language socialization and learner interactions involves analysis of conversations between learners and their native-speaking interlocutors, examining how relationships between learners and their hosts are constructed in conversation. Some focus on the discourse patterns that learners and their hosts use in their interactions. Wilkinson's (2002) study, for example, showed how learners and their hosts alike rely heavily on classroom roles and discourse structures to manage their interactions. Other studies document the ways in which the hosts’ perception of foreigners, and their beliefs about foreigners, affect the input and corrective feedback that they provide (Siegal 1996; Cook 2006; Iino 2006; Kinginger et al. 2016; Lee et al. 2017).

The second line of research employs recordings of naturally occurring conversations to trace learners’ language development. Such studies have focused on the development of pragmatic features of language such as requests (Shively 2008, 2013) and openings and closings of service encounters (Shively 2008). Other studies have examined the development of linguistic accuracy (DeKeyser 2010) or acquisition of particular discourse features, like the German particle also (Schirm 2015). Thus, close qualitative analysis of learners’ interactions can document the processes by which they learn particular features of the language and their progress over time.

Another area of qualitative research involves what van Lier (1994) terms action research. In a 1994 article, van Lier defined action research as ‘small-scale intervention in the functioning of the real world and a close examination of the effects of such intervention’ (p. 32). In response to qualitative (and quantitative) studies that show that learners’ expectations of SA are often not met, and that gaining access to communities of practice in the target culture is often difficult, many researchers have recommended training and other interventions that may help learners to make more of their SA experience. Research has documented the value of such interventions as requiring learners to participate in extracurricular activities and service opportunities (Engle & Engle 1999; Kinginger 2011), to conduct ethnographic research projects while in-country (Jackson 2006a; Kinginger 2011; Lee 2012), and to keep blogs or other journals during their sojourn (Allen 2013). Jackson (2006a) and Kinginger (2011) both highlight the importance of teaching learners the skills of observation and introspection. Belnap et al. (2016) emphasize training in self-regulation as well as the importance of providing regular conversation partners for SA students, particularly when homestays are not an option. Findings suggest that these interventions promote social interaction during SA as well as increased intercultural development.

Over the past two decades, qualitative research in SA has become popular, particularly as a means of understanding the complexities of SA contexts. The vast majority of the research seeks to understand the individual variation noted in variables-based research, and to examine the roles played by social and environmental factors as well as by individual differences. Such research has focused on various language learning behaviors (Pellegrino 2005; Isabelli-García 2006; Magnan & Back 2007) and motivations for learning the L2 and studying abroad. Qualitative research to date has clearly indicated that language learning in SA is, as Kinginger (2011) states, ‘a dialogic and situated affair whose success depends not only on the attributes and intentions of the student but also the ways in which the student is received within his or her host community’ (p. 60).

Though qualitative research methods have painted a richer picture of SA participants and the communities in which they interact, some scholars have noted limitations to their application to SA research. Kinginger (2009) notes a tendency to limit research to the perspectives of the participants, without reference to the host community. For example, much of the research on gender-related incidents examines only student reactions to incidents deemed ‘harassment’ without an attempt to understand the host perceptions and the typical gender-related practices of the community. Thus, scholars have suggested that researchers expand their inquiry to consider all of the stakeholders in a SA program, including the teachers, administrators, host families, and others with whom the learners interact, as in Trentman & Diao (2015). Another limitation includes the relative lack of longitudinal studies examining the effects of SA on learners’ language or identities after SA. In response, some SA researchers have also turned attention to the long-term effects of SA. In particular, they have become concerned with the effects of SA on the change in learner identity (see Benson et al. 2013) and have probed learners’ views of themselves a year or two after the experience. However, this type of research is relatively new, and few have had the time to follow a group of learners for more than a year or so after the SA experience.

2.3 Mixed-methods approach

A fruitful and growing approach to SLA research is the mixed-methods approach, in which researchers combine in-depth qualitative analysis of students’ experiences with quantitative documentation of learning outcomes. Among the studies that successfully blend these two lines of inquiry are Isabelli-García (2006, 2010), Kinginger (2008), Briggs (2016), Jackson (2016), and Tracy-Ventura et al. (2016). These studies evaluate measures of linguistic gain against accounts of learners’ experiences in the host culture, contributing to a fine-grained analysis of individual students’ behavior as an intervening variable between aptitude and initial proficiency and language learning success. The earlier studies used small samples, whereas Briggs (2016) documents the vocabulary strategies of 241 adult SA learners in two learning contexts and Tracy-Ventura et al. (2016) investigate the personality changes of 58 British undergraduate students who spent their third year abroad in a French- or Spanish-speaking country. And Jackson (2016) researched the language and intercultural learning of 149 Chinese students who participated in a semester-long exchange program in an English-speaking country.

Other mixed-methods projects include the LANGSNAP research project that documents a range of language domains, connecting progress in L2 to a range of individual, social, and contextual variables. The project tracked a cohort of 56 students majoring in French or Spanish before, during, and after spending their year abroad in France, Spain, or Mexico (McManus, Mitchell & Tracy-Ventura 2014; Mitchell 2015). Klapper & Rees (2012) link cases of ‘quantified’ individual progression with qualitative data. The study involved 57 students in total, with an in-depth focus on 12 who either failed to show progress or who showed substantial progress on various measures of linguistic gain.

Qualitative research is an important component of research on SA phenomena, providing a rich picture of individual variables and of the complexity of language learning in varied SA contexts. Combining quantifiable measures of linguistic or intercultural gains with qualitative descriptions of learners’ experiences, beliefs, and behaviors lends validity to the research. Moreover, there is a need to expand the focus of the research to include a broader range of perspectives.

3. Linguistic constructs and gains: Documenting learning outcomes

DeKeyser (2014) aptly points out that the field of SA is still far from ‘go[ing] beyond descriptive adequacy to explanatory adequacy, and ultimately . . . predict what we are going to find in the next study’ (p. 319). The SA experience—a complex social ecosystem with an endless number of independent variables—is difficult to control for.

Research carried out on SA has blossomed from a field that focused on isolated linguistic constructs to multiple considerations of the language learner. Research studies focus on a variety of topics, ranging from linguistic performance on a single token such as lexical, morphological, or phonological features to nonlinguistic domains such as fluctuating identity formation that occurs abroad and development of intercultural competency. As is to be expected, although SA participation continues to be viewed as an experience that ‘allows for intensive, regular, contextualized L2-use opportunities in situ. . . . A closer look . . . however, indicates contradictory findings’ (Yang 2016: 67).

The aim of this section is not to carry out a review of all SA research in all its transformations. Instead, we focus here on holistic linguistic performance gains and areas where contradictions and omissions are present. These are limited to include empirical studies based on the more frequently researched aspects of SLA in SA: the four modalities of language use and the development of particular aspects of communicative competence. This itself is a challenging task as recent research focuses on multiple aspects of SA, measuring multiple learning outcomes.

3.1 Modalities of language use

3.1.1 Reading

Few scholars have examined the influence of SA on reading development. Churchill (2006) and DuFon (2006) note that the relative lack of studies on reading reflects a bias both in-program design and in expectations toward the development of speaking and listening skills. Moreover, the unique affordances of SA allow learners to interact with native speakers in a way that classroom learning experiences do not, thus predisposing learners to focus on speaking and listening during their time abroad. Students typically have many opportunities to read the L2 even in their home country, but they have fewer opportunities to speak. Thus, many learners see less value in reading during SA. Allen & Herron's (2003) study confirmed this bias, finding that students prefer to obtain information by conversing with others rather than by reading. Nevertheless, a number of studies have demonstrated that time spent reading abroad leads to measurable gains in reading comprehension (Brecht, Davidson & Ginsberg 1995; Lapkin, Hart & Swain 1995; Kinginger 2008), as measured by various standardized tests of reading comprehension.

While the studies cited above have looked at the development of reading comprehension, others have examined more closely the development of particular skills in reading abilities. For example, Fraser (2002) examined the ability of learners to match anaphora and cataphora to their referents, while Dewey (2004) reported that American learners of Japanese became stronger in vocabulary recognition and text comprehension, along with developing greater confidence in their reading ability. Briggs (2016) also found that learners made significant gains pre- to post-SA on receptive and productive vocabulary knowledge—knowledge necessary for reading comprehension and writing proficiency development respectively.

Some contradictory evidence does exist, however. For example, Dewey's (2004) quasi-experimental study comparing the reading gains of learners on SA with learners in an intensive IM environment setting did not find differences between the SA and the IM group on objective measures of reading ability. The SA group did, however, report greater confidence in their reading ability than did the IM students. Dewey also found differences in the types of strategies used by SA students and IM students to comprehend written text. The SA group monitored their understanding of what they were reading more than the IM group, while the IM group more often responded effectively to the text. One might also consider Huebner's experimental (1995) study contradictory, since differences between SA and AH learners were not significant in this study. Huebner noted possible differences in motivation between the two groups, with the SA group embracing the introduction of various Japanese writing systems early in their instruction because they had a greater instrumental need to read than did their counterparts at home. Huebner suggests a possible connection between this motivation and the higher (but not significantly different) scores for SA than AH learners.

Whereas much of the focus on reading comprehension in SA has been on learning outcomes, Kline (1998) advocated for a process-oriented approach. In an ethnographic study of the reading habits of learners on an SA program in France, Kline concluded that reading behavior was influenced by a number of factors, including gender, group affiliations, and host family preferences. Kline, however, did not establish a relationship between learners’ reading behaviors and their reading ability.

Reading, thus, represents an understudied domain in the research on SA. Moreover, much of the focus has been on discretely measurable skills, such as reading comprehension, recall, and vocabulary development. There is very little research on student motivations to read, their values and reading preferences, or on their development over time (Kline 1998). Moreover, the development of literacy skills while abroad involves more than development of linguistic and pragmatic skills (such as word decoding or identifying textual genres); indeed, learning to read and write requires negotiating new academic cultures and adapting to new forms of literacy (Kinginger 2009), especially, as Taillefer (2005) documents, very different academic literature cultures among European countries and different reading strategies among students from different nationalities.

3.1.2 Writing

Development of writing skills during SA is even less studied than development of reading skills, though some studies indicate that learners report improvement in their writing skills (Meara 1994; Adams 2006). Very few studies have used objective measures to document the development of writing in such areas as proficiency, fluency, or quality of essays, and those studies have yielded contradictory results. Freed, So & Lazar (2003) compared the written fluency of learners on SA with that of learners AH. They found that native speakers (NSs) did not judge the writing of SA students to be any more fluent than that of their AH counterparts, though objective measures demonstrated that the essays of SA learners had grown in length and lexical density. On the other hand, Sasaki (2004, 2007, 2009) found that SA did have a positive effect on participants’ writing, as compared to the writing of students AH. In particular, she found that SA participants developed different strategies for writing and became more motivated to improve their L2 writing than did their AH counterparts. A later follow-up study (Sasaki 2011) revealed that, in addition to improved writing skills, students who studied abroad formed ‘imagined communities’ with L2 speakers that motivated them to improve their writing in the L2.

Another set of studies related to the SLA project, studying the long-term effects of SA on language learning, suggests that SA may have a positive impact on L2 writing skills, even if learners do not receive formal instruction in L2 writing. Pérez-Vidal & Juan-Garau (2009) compared the development of complexity, accuracy, and fluency among SA students following either a 6-month period of instruction AH or a 3-month SA program. They found that SA participants, as opposed to AH participants, made significant progress in fluency measures and lexical diversity. Barquin (2012, cited in Pérez-Vidal & Barquin 2014), likewise, found improvement after SA in fluency, lexical richness, accuracy, syntactic complexity and variety, and cohesion. A later study by Pérez-Vidal & Barquin (2014) considered the development of writing skills among a group of 72 learners over a period of 30 months, which included an SA period preceded and followed by formal instruction AH. The authors found that a 3-month period of SA ‘led to significant gains in [. . .] writing in the domains of fluency, accuracy, and lexical complexity, and that no such gains were found in periods of [f]ormal [i]nstruction either before of after the SA context’ (p. 232). Moreover, they found that the gains made during SA were maintained 15 months after the SA program had ended.

The dearth of literature on the development of writing in SA may reflect a bias toward development of aural/oral skills—at least for SA in FL environments. Much more has been written about the development of writing in L2 situations. Nevertheless, the scant literature on the subject does suggest that SA can lead to some improvements in writing. If nothing else, SA researchers may wish to examine the role of writing as a support skill, since at least one of Kinginger's (2008) participants attributed much of his academic progress in France to the activity of writing.

3.1.3 Listening comprehension

As with literacy skills, the literature on listening comprehension in SA is scant. Meara (1994) and Cubillos & Ilvento (2012) found that learners feel more confident about their listening skills following SA (with Kaplan's [1989] study representing something of an exception: students felt that they had made the most gains in receptive skills, listening and reading) than they do about their literacy skills. This is striking since Ikeguchi (1996) finds that literacy skills appear more durable after SA than oral/aural skills. The existing research largely confirms their self-assessments: according to the literature, students on SA make significant gains in listening comprehension, particularly when compared to ‘at home learners’ (Dyson 1988; Ginsberg, Robin & Wheeling 1992; Llanes & Muñoz 2009). Other studies using a pre- and postprogram design have also found that learners on SA make significant gains in listening comprehension (Allen & Herron 2003; Kinginger 2008; Davidson 2010), though Davidson (2010) and Cubillos & Ilvento (2012) suggest that longer-term SA is necessary for greater gains. This finding, however, contradicts that of Llanes & Muñoz (2009), who report that students on a 3–4-week SA program made significant gains in listening comprehension. A number of other studies using pre- and postprogram designs have found only modest gains in listening comprehension (Huebner 1995; Tanaka & Ellis 2003). Davidson further found that preprogram listening comprehension was a strong predictor for the development of oral proficiency, while Cubillos et al. (2008) note a potential threshold effect; that is, learners must have reached a certain threshold of listening comprehension in order to make any progress in their comprehension during SA.

Very few studies have compared the gains of students on SA to those in other settings. Huebner's (1995) study represents an exception. He compared gains on the Educational Testing Service's Japanese Listening Proficiency Test for students on a 9-week SA program in Japan with learners in an IM program of comparable length and pedagogical approach. In his study, he found that the SA group performed slightly better on the listening test than did the IM group, but the number of participants was small and the numbers were not significant. Cubillos et al. (2008) likewise compared performance on a standardized listening comprehension test by groups of American learners of Spanish on 5-week SA programs in Spain and Costa Rica and by students enrolled in a course of similar length in their home country. The authors found no significant differences between the two groups, though the authors assert that the SA learners differed from the AH learners in their use of strategies. In particular, the SA group tended to use more ‘sophisticated, social, and top-down strategies,’ while the AH learners tended to use less-productive bottom-up strategies.

Whereas most of the above studies have focused on listening comprehension broadly defined, Taguchi (2008, 2011) has examined particular skills associated with listening, including speed of comprehension, lexical access, accuracy of comprehension, and pragmatic comprehension—that is, the ability to understand the intended meanings of particular types of utterances. Her 2008 study examining the lexical access speed and the speed and accuracy with which learners comprehended indirect refusals and opinions noted gains in all areas for SA participants, though their ability to understand indirect opinions in their L2 was not as great as their ability to comprehend indirect refusals. In her 2011 study, Taguchi compared the pragmatic competence of SA learners and AH learners and found that SA learners comprehended certain implicature types more accurately than did their AH counterparts.

Though the primary focus on listening comprehension has been on documenting gains in listening, other studies have attempted to document the types of listening activities in which learners engage (Kaplan 1989; Freed, Segalowitz & Dewey 2004; Kemp 2010; Llanes & Muñoz 2013), usually in the context of examining overall language contact, though Kemp (2010) focuses on a listening log. Kemp's purpose is to document the ways in which a listening log helped learners to identify comprehension problems, to reflect on their motivation for listening, and the strategies they employed to improve their comprehension abilities.

Kinginger (2009) notes a potential problem of ecological validity with the majority of studies listed above, in that they have assessed listening comprehension via multiple-choice tests, which are biased toward academic language and functions and may not reflect the variety of listening in which students engage during SA. Moreover, she notes that the studies have largely ignored the quality of students’ engagement in L2-mediated activities as well as learners’ identities and histories that might affect their listening comprehension. Kemp (2010) documents the use of a listening log among Erasmus students studying abroad in the UK. She notes that the log helped researchers to better understand learner motivations for listening, and appeared to help learners reflect on the sources of any comprehension difficulties and to develop strategies to aid their comprehension.

3.1.4 Speaking

The most well-researched outcome of SA is speaking gains. Studies within this modality frequently compare outcomes with the control group of AH students, but overall, research to date has not been able to offer a concise holistic conclusion of whether SA is beneficial for oral skills development. Granted, pre- and postprogram speaking measures either show that the student remains at the same level or improves. While most measurement of speaking gains in the SA literature has involved proficiency, the field has also become curious about how specific aspects of oral proficiency change and, therefore, as a natural progression, we have ‘deconstructed’ oral proficiency into more detailed researchable areas. These studies have shown that even if holistic measures of oral proficiency do not show growth over time, atomistic aspects such as fluency (speech rate, length of turns, hesitancy, and self-repetitions), pronunciation, spoken grammatical accuracy, and pragmatics often do (see Section 3.2 for some of these findings).

Global oral proficiency is often measured during SA using instruments such as the ACTFL Oral Proficiency Interview (OPI) (East 2016) that can be face-to-face, telephonic or computer-mediated and is intended to recreate authentic interaction between the student and the test rater. The validity of the OPI has been called into question (Bordón & Liskin-Gasparro 2014; Lantolf 2014; East 2016) on several facets, the most notable being the ‘artificiality’ of the interaction between the interlocutor and student (East 2016: 41). A simulated OPI—an alternate method with similar issues—involves the use of prompts either recorded or written in a booklet to elicit speech samples similar to those elicited through the OPI (Vande Berg et al. 2009; Hernández 2010). Semi-guided oral interviews, where students interact with each other, as interviewer or interviewee, are also used (Valls-Ferrer & Mora 2014). Given concerns about the artificiality of such measures as these, some have used recorded natural conversations to understand spoken politeness (Shively 2011), accent (Diao 2017), and food-related discourse patterns (DuFon 2006).

Studies that provide helpful conclusions regarding speaking development abroad move away from using the SA context as a categorical label (e.g., Wood 2007; Trenchs-Parera 2009; Serrano, Tragant & Llanes 2012; Jochum 2014; Prokhorov & Therkelsen 2015) and more toward a research methodology that measures multiple independent variables of the social ecosystem that SA involves (e.g., Martinsen 2010; Baker-Smemoe et al. 2014; Di Silvio, Donovan & Malone 2014; Valls-Ferrer & Mora 2014). Speaking assessment gains is the dependent variable of the dynamic system the students find themselves in, many times taking into account factors such as preprogram proficiency, differing learner characteristics, and the quantity and quality of out-of-class language contact. In general, speaking proficiency development is as varied as the multiplicity of factors that are found in SA and should be assessed taking those factors into consideration.

3.2 Communicative competence

The term communicative competence in SLA has undergone much discussion. Although other versions of communicative competence exist (e.g., Hymes 1972; Bachman 1990), we use Canale's (1983) model that consists of the following competencies: grammatical (vocabulary, word formation, sentence formation, pronunciation, spelling, and semantics), sociolinguistic, strategic, and discourse. Of the elements listed in these competencies, we focus here on the SA research trend of studying accuracy development in grammatical and sociolinguistic competence. Moreover, we review those studies that are situated within methods that take into account the social aspect of language use.

SA research encompassing accuracy, defined as ‘the degree of deviancy from a particular norm’ (Housen & Kuiken 2009: 3), is normally included within studies of complexity, accuracy, and fluency and ‘cover a wide spectrum of methods in applied linguistics, ranging from holistic and subjective ratings to objective quantitative measures of L2 production. The latter are clearly the preferred method in L2 studies’ (Housen, Kuiken & Vedder 2012: 8). These authors also note that, although early research in the area of L2 accuracy was more general in nature, a recent trend ‘calls for finer-grained analyses . . . and hence a return to measures targeting more specific subdomains of language and more distinct linguistic features, as a complement to the use of more global measures’ (p. 8). They add that in SLA research, the number of accuracy measures is ‘daunting and partly reflects the lack of consensus on how [they] should be defined as constructs’ (p. 8). Research focusing on accuracy development in SA has typically focused on the number and type of errors in two areas. The first is the SA learners’ development and approximation to NS phonological/phonetic norms, the second is grammatical development toward the prescriptive norm, and the third area discussed in this subsection is development toward target-like interlanguage norms during SA.

3.2.1 Grammatical competence–phonological

As highlighted by Müller (2016), research on the acquisition of L2 pronunciation focuses on differences in pronunciation gains, similar to other aspects of linguistic development abroad, and has found mixed results mostly attributable to individual differences. Studies typically focus on the approximation of participants to the native production of salient phonological features of regional language varieties (e.g., O'Brien 2003; Knouse 2012; Reynolds-Case 2013; Avello & Lara 2014) as well as the learners’ perceptual competence (e.g., Mora 2014; Romanelli, Menegotto & Smyth 2015). In general, such research ‘has shown that learners can adopt the regionally specific features of the language but typically do not use them to the same extent as native speakers’ (George 2014: 98).

Measuring phonological features (usually acoustic analyses realized using Praat [Boersma & Weenink 2013] and spectrographic analyses [e.g., Lord 2010; Stevens 2011]) either treats the produced variable as an error that is not in line with the target language norm or as an alternation of a target language norm (Salgado Robles 2014: 238). In fact, a trend in SA pronunciation research conjectures that L2 pronunciation errors do not always impede comprehensibility. That is, these errors could be due to a ‘trade-off effect’ (Serrano, Llanes & Tragant 2016), as learners may focus on other aspects of their oral production, such as proficiency aspects that deal with optimal comprehensibility with native interlocutors outside of the classroom.

Treatment options measuring pronunciation typically involve NSs' rating recordings of participants’ read-aloud tasks using pre-established rubrics, oral narratives based on a series of pictures, and elicited imitation tasks analyzing how successful learners are at replicating native speech norms. However, this nonspontaneous speech, although convenient for data collection, is not true to that used in face-to-face, conversational interactions, as noted in Hardison (2014), nor do they take into account whether errors or variations impede comprehensibility while abroad. More spontaneous conversations simulating the OPI prove to elicit more natural language (e.g., Fernández-García & Martínez-Arbelaiz 2014; Martinsen, Alvord & Tanner 2014). They tend to be less artificial than read-aloud tasks and can capture changes during SA in pronunciation that might be seen in more authentic speech.

The extent to which conclusions can be drawn and compared are limited since the research methods do not ensure that the procedure includes a measure of prior knowledge and maintenance of the features under consideration (George 2014), measuring language contact information abroad (Bongiovanni et al. 2015), or learner perceptions of token items (Pope 2016). Müller (2016: 113) reminds her readers that individual learning factors need to be included in the analyses of future studies, which explains why, to date, it has been ‘difficult to derive common trends from learner cohorts.’

3.2.2 Grammatical competence–word and sentence formation

Grammatical accuracy is typically measured in one of two ways. One method measures accuracy of specific grammatical features in controlled oral production. And the other is via tools that measure accuracy using written surveys or questionnaires. Results from studies employing quasi-experimental designs using an AH control group are just as valuable as within-group pre-experimental studies. The former seek to investigate the effect of SA by comparing with AH learning, whereas the latter tend to provide a descriptive notion of how the social nature of language use comes into play in SA, taking into account learner variation factors.

The focus of grammatical accuracy typically centers on items in the L2 that do not appear in the L1, since dissimilarities may make these items more challenging, causing them to merit more extensive exposure through SA. Generally, development in grammatical accuracy is in line with other aspects of language domains. That is, although some students show development (e.g., Duperron 2006; Marques-Pascual 2011; Juan-Garau et al. 2014; Salgado Robles 2014; Grey et al. 2015), not all do (e.g., DeKeyser 2010; Geeslin et al. 2010; Isabelli-García 2010; Wang 2010), and quasi-experimental work often indicates no difference in development between SA and AH (e.g., Martínez-Arbelaiz & Pereira Rodríguez 2008; Arnett 2013).

Methodologies that include NSs as a comparison group tend to have results that are more true to the norm of the target language group (Serrano 2010; Kanwit, Geeslin & Fafulas 2015). It is arguable whether or not learners’ development (or lack thereof) is explained by different durations abroad. Nevertheless, grammatical accuracy is documented either via measures that elicit oral production (e.g., Juan-Garau 2014) or written production (Juan-Garau et al. 2014).

Foster & Wigglesworth (2007) call for research that includes reliable accuracy tools that incorporate a system to measure the level of seriousness of an error. Written data collection measures frequently employ grammaticality judgment tests or written narration of picture stories. Oral data collection measures also include narration of picture stories or semi-guided interviews. Both measures are given at a preprogram and postprogram interval as the norm, at times collecting data from intermediate time points, a repeated measures approach. These multiple collection points provide a nuanced look at the restructuring of acquired knowledge that occurs during a longer stay abroad. In both situations, those methods that base the data collection process on ‘a more communicative assessment task instead of a strictly grammar-driven assessment provided a more detailed and appropriate measure of students’ knowledge’ (Arnett 2013: 710).

3.2.3 Sociolinguistic competence–interlanguage pragmatics

Research in the field of sociolinguistic competence during SA tends to focus on pragmatics—the study of those aspects of language that cannot be considered in isolation from its use, language in its situational context, including the knowledge and beliefs of the speaker and the relationship and interaction between speaker and listener. It is generally known SA has a significant impact on the acquisition of sociolinguistic competence (Rasouli Khorshidi 2013; Ren 2013; Félix-Brasdefer & Hasler-Barker 2015). Within SA, research has been generally carried out in interlanguage pragmatics (ILP) with intercultural pragmatics and cross-cultural pragmatics receiving less focus.

ILP research in SA focuses on L2 learners’ production of speech acts, such as requests (Schauer 2009; Bataller 2010; Shively 2011; Bella 2012; Nguyen & Basturkmen 2013; Li 2014; Al-Gahtani & Roever 2015), and the extent that request strategies change to target-like norms during SA. Other areas vary widely to include the development of address terms (Schauer 2009; Hassall 2013) and appropriateness (Diao 2014; Taguchi 2015), use of honorifics (Brown 2013), apologetic behavior (Warga & Schölmberger 2007), and refusals and opinions (Taguchi 2008), among others. The quantity of interaction in SA has been shown to play an important role (Taguchi 2008; Bella 2012; Vilar-Beltrán 2014). Nonetheless, development in the production of requests varies during SA (Schauer 2009; Bataller 2010; Iwasaki 2010), not unlike the other aspects discussed in the previous pages of this review. Furthermore, on the receptive side, learners can show development on measures of speed but not accuracy (Taguchi 2008), indicating the importance of assessing development in multiple ways.

Ren (2013: 722) states that ‘it is generally agreed that there is no ideal method for all investigations in ILP’; however, the data collection tool commonly used is the Discourse Completion Task (DCT). A DCT is most often a written survey with situational prompts. That differs from an oral DCT (oral response to a written or oral prompt) and from a role-play, which involves the learner interacting orally with another person in real-time, usually involving multiple turns. The use of DCTs has been shown to be problematic (Schauer 2009; Golato & Golato 2013). Golato (2003), as cited in Félix-Brasdefer & Hasler-Barker (2015: 78) explain this is because this instrument ‘elicits single-turn responses, has a low degree of content validity, and is simulated and non-interactive’. However, Félix-Brasdefer & Hasler-Barker (2015: 78) acknowledge that ‘this type of instrument can be reliable for measuring off-line pragmatic knowledge (what learners know) when they have a limited amount of time in which to respond.’ Using natural data, such as audio-recorded interactions (Shively 2011), is recommended to complement experimental data. Other experimental tools used to measure ILP include the Language Awareness Interview (Kinginger 2008) and role-play scenarios (Nguyen & Basturkmen 2013) that are designed to elicit more naturally occurring discourse and to ‘observe how context factors such as power, distance, and imposition affect the speaker's choice of pragmatic strategies’ (Nguyen & Basturkmen 2013: 20–21). As pointed out by Stokoe (2013), because the use of language in social situations and in role plays may differ, comparing experimental data at various points in time abroad with natural data is encouraged. However, the challenge with natural data is its collection and comparability. While natural data is highly desirable, another alternative to the written DCT that has been advocated recently (see, e.g., Bardovi-Harlig & Shin 2014) is to move away from the written DCT and use oral DCTs to improve the validity of pragmatics data collection.

4. The language learner: Evolving views of the learner

This section will trace the recent evolution of research on the learner, from investigations of variables related to individual differences (anxiety, motivation, or attitude) to more recent views of learner identity and agency and student perspectives about language learning.

Just as scholars in SA research have broadened their view of the learning context in SA, so, too, have researchers changed their conceptualizations of the language learner. This shift in perspective can broadly be summarized as moving from a view of learners as homogenous, yet metaphysically independent, processors of linguistic input to a conceptualization of the learner as a ‘whole person’ (Coleman 2013a). This transformation, which largely parallels trends in mainstream SLA, has occurred in stages. The initial stage comprised a cognitive view of the learner as an independent processor of linguistic information. Early on, however, scholars adopted a social psychological model of the language learner, examining individual difference variables and their effects on the SA experience. Block (2003) noted that by the end of the twentieth century, a ‘social turn’ in language acquisition occurred, which greatly expanded our conceptualizations of language learning and learners. At this time, the research began to focus on learners as ‘individuals, with intentions, agency, affect, and . . . histories’ (Pavlenko & Lantolf 2000: 157). At about the same time, scholars also began to view learners as apprentices or ‘legitimate peripheral participants’ (Lave & Wenger 1991) trying to become fully fledged members of a particular community. Recently, following Coleman's (2013a) injunction, researchers are beginning to recognize learners on SA as ‘whole persons’ rather than as simply learners. In this section, we examine the literature that reflects each conceptualization of the language learner. Early outcomes-based literature on SA treated language learners as homogenous processors of linguistic information. Paralleling the undifferentiated view of SA as a ‘context’ of language learning, scholars adopting this view tended to compare students on SA programs to students in other learning environments, especially AH learners and sometimes IM learners. Just as the context was assumed to be largely the same for each learner, and broadly comparable, so, too, were the learners in each program assumed to be similar enough as to invite aggregate data allowing scholars to draw conclusions about the effectiveness of each environment. This view of the language learner persists even in more recent studies of SA. For example, recent research has compared the learning outcomes for students on two types of short-term SA programs and those in domestic residential IM program (Martinsen et al. 2010), while Llanes & Muñoz (2013) have studied the difference in writing gains for SA and AH students, and Taguchi (2011) compared the pragmatic comprehension of classroom learners and SA learners as well.

Over time the persistent variability in learning outcomes led scholars to examine individual differences among learners, usually casting variables such as age, gender, aptitude, and motivation in terms of ‘predictors’ of language gains (Brecht et al. 1995; Davidson 2010). Thus, researchers sought to uncover the source of variation in traits that could be measured and isolated, taking a correlational approach. This so-called ‘social psychological’ view of the language learner is still employed in SA research (e.g., Pérez-Vidal 2014; Leonard & Shea 2017).

Studies on predictors of language gain provided some insight into possible variations in learning outcomes, but still took a relatively unnuanced view of language learners. In large measure, language learners were still seen as individual language processors with different abilities within different settings. Kinginger (2009: 37) notes that this search for ‘causal variables’ can lead to a view of language learners as balls in a pinball machine, in which the learner is a ‘hapless and unconscious object, identical to all its companion objects, run[ning] a chaotic and meaningless course through an immutable social structure,’ in which, in Schumann & Schumann's (1977: 248) words, ‘knobs represent the social, psychological, cognitive and personality variables.’

Despite its continued use in SA research, the social psychological approach to the language learner has been criticized on various grounds. Kinginger (2009: 156) points out that the broad array of affective and personality factors treated as ‘causal variables’ are not only dynamic, but ‘even volatile, responding to specific events and practical or emotional circumstances’ that can be ‘reshaped and even stamped out . . . at least temporarily.’ In fact, Pavlenko & Lantolf (2000) argued that the meanings of fixed, apparently ‘biological’ factors such as age, gender, and race are largely dependent on culture, and that these factors may be socially constituted.

Critics of the social psychological approach to learners and learning argued for a view of learners as agents who ‘actively engage in constructing the terms and conditions of their own learning’ (Lantolf & Pavlenko 2001: 145) and who make choices about how they will participate—or not participate––in communities and practices they encounter. Scholars recognized that agency plays a central role in the construction of learner identities, degrees of learner autonomy, motivation, and strategic learner behavior (Toohey & Norton 2003; Ushioda 2007; Bown 2009; Gao 2010). Moreover, researchers adopted a more nuanced view of SA contexts, asserting that SA language learners are received in various ways. Though some may be welcomed and given ample opportunities for involvement, others may find the host community less welcoming, as a result of which, opportunities for participation are limited (Wilkinson 1998a). Though quantitative studies of language learners are less suited for exploring much of the complexity of language learning in SA contexts, researchers are attempting to deal with some of that complexity through use of factorial research design, especially for the generation of hypotheses. Scholars have turned their attention to studying factors that might explain learners’ behaviors. Whereas in previous research, L2 use was seen as a predictor of proficiency gain (Spada 1986; Yager 1998), more recently, Dewey et al. (2014) have examined the factors that contribute to L2 use, including personality, gender, preprogram proficiency, and size of social networks.

As views of the learner and of language learning have changed, many scholars have turned to qualitative research as means of comprehending emic (or insider perspectives), rather than the etic (researcher-focused perspectives), characteristic of the variables-based approach to research. In the mid-1990s, a number of qualitative studies of SA language learning emerged, including Polanyi (1995), Pellegrino (1998, 2005), Wilkinson (1998a, 1998b, 2002), and Talburt & Stewart (1999). These scholars explicitly examined learners’ perspectives on their SA experiences. In the ensuing years, an explosion in the number of qualitative studies has taken place, with many focusing on learners’ competing goals relative to language use and learning (Pellegrino 2005), as well as learners’ responses to unmet expectations and additional affective challenges (Allen 2010a; Isabelli-García 2010; van Maele, Vassilicos & Borghetti 2016). These studies present learners as individuals, with intentions and goals and complex histories that often affect their motivation, their language learning behaviors, and their access to interactions (for a thorough review of literature related to identity and SA, see Kinginger 2013a).

In examining language learners as individuals, qualitative research has also sought to understand the dynamic nature of learners’ motivations for language learning in SA and beyond. Research demonstrates that motives for studying a language are as complex as are the learners themselves; motives can be tied up with learners’ personal histories (Kinginger 2004), their views of the target culture (Isabelli-García 2006), their perceptions of the SA experience (Allen 2010a), as well as with the discourses surrounding particular languages and cultures (Kinginger 2004; Trentman & Diao 2015). Additional research has suggested that imagination plays an important role in sustaining motivation and motivated behavior; in particular, learners envision their ‘ideal selves’ (Dörnyei 2009) interacting in ‘imagined communities’ of L2 speakers (Kinginger 2004; Anderson 2006; Trentman 2013). Isabelli-García (2006) demonstrated the recursive relationship between motivation and participation in social networks, concluding that learners’ motivations ‘must be understood with reference to social relations of power that create the possibilities for language learners to speak’ (p. 254).

While the studies mentioned above hint at the role played by contacts in SA networks, especially teachers and host families, they tend to privilege the perspectives of the learners, who are often not equipped to fully understand the culturally bound meanings associated with particular practices and behaviors. Thus Kinginger (2013b) called for scholars to give more attention to the local meanings attached to particular practices and to the host community's perspectives (Stephenson 1999; Knight & Schmidt-Rinehart 2002; Schmidt-Rinehart & Knight 2004). Recent research, emerging from a language socialization view of language learning, emphasizes local practices and perspectives (Cook 2006; DuFon 2006; Kinginger 2015; McMeekin 2017).

Language socialization conceptualizes language learning as a process of socialization and not merely of acquisition. Learners are not construed as independent processors of linguistic forms, but rather as apprentices, striving to learn not only the practices of the host community but also the meanings attached to those practices by local members. Thus language learning involves more than the accumulation of competence by individuals; instead, it represents one aspect of becoming a person in society (Ochs 2002). Research examining the processes of language socialization in SA has provided important insights into how learners are received by their host cultures and how hosts socialize (or not) learners into particular practices (Cook 2006; DuFon 2006; Kinginger 2015; McMeekin 2017).

While trends in SA research employ a broader and more nuanced view of language learners, we must keep in mind that ‘language learner’ may constitute only a very small part of the identity of SA participants (Kinginger 2009; Coleman 2015). In fact, Kinginger (2013a) questions the extent to which participants are language learners at all. Even on language-focused SA programs, some learners adapt different stances toward language learning, resulting from their own goals and as a result of interactions with significant others. Indeed, Kinginger (2013a) argues that because participants do not privilege their identities as language learners, SA program administrators must be proactive in structuring the environment and in motivating students to learn the target language while abroad (Kinginger 2010, 2013a). As a result, a number of scholars have examined the effectiveness of various program interventions to promote interaction and foster self-regulation (Jackson 2006a: 132–185; Vande Berg et al. 2009; Kinginger 2010; Allen & Dupuy 2012; Di Silvio et al. 2014; Belnap et al. 2016).

An exclusive focus on SA language learning may obscure the most significant elements of the sojourn, including romance, self-discovery, and relationship building (Coleman 2013a). Recent SA research has moved beyond language learning, examining the effects of SA on learners’ identities, including religious identity (Poag & Sperandio 2015), sexual identity (Bryant & Soria 2015), global engagement (Shadowen, Chieffo & Guerra 2015), and vocational identity (Kronholz & Osborn 2016). Whereas most of the previously cited research has focused on the impact of already-formed identities on SA experiences, scholars have begun to examine the impact of experiences of SA on the development of learners’ identities during and after SA (Benson et al. 2013). In such studies, SA is viewed as a potentially ‘critical experience’ (Benson et al. 2013: 33), in which L2 identity development is likely to be observable in students’ narratives of their experiences. This research agenda offers a broadened view of SA. SA is no longer seen as simply an environment for language learning and use; instead scholars recognize that SA can trigger identity processes and the building of new ways of learning and using languages.

This reversed direction of research highlights the complexity of SA, which can be seen as a social ecosystem in which all parts of the system are in constant flux. In such an ecosystem, any change can have consequences for the participants in the system and for the system in general. That is, ‘[u]nderstanding the study abroad phenomenon requires researchers to take into account the whole person and the whole context’ (Coleman 2013a: 36).

Coleman (2013a) further notes, however, that, while SA has traditionally been seen as a subfield of applied linguistics, the broader truth is that the subfields of applied linguistics can be applied to SA research. Indeed an overview of recent research demonstrates that SA research is quite varied, addressing subfields of applied linguistics, such as individual differences, sociolinguistics, IM, and literacy. Scholars examine it from a variety of perspectives, fewer and fewer of which might seem to be readily related to language and language learning.

5. Pedagogical implications: Curricula for SA language education

5.1 Five categories of SA curriculum

Curricula, syllabi, pedagogical or teaching and learning approaches, specific course materials, readings or textbooks, course assignments, and formative and summative assessments receive little attention in most L2 SA research. This might be surprising given the potential of such to influence the linguistic, communicative, and intercultural learning outcomes that this research measures or explores. Perhaps because most research studies on SA are primarily interested in out-of-class opportunities for language use, studies measuring students’ L2 proficiencies usually provide only minimal description of the curricular program. They provide specific information related to the language research scope and design rather than a commitment to explore pedagogy and programming. Research or scholarly reports that solely or primarily focus on curriculum and pedagogy for SA, or the articulation of SA with domestic curricula, are rare.

More research on curricular issues is needed because, as Engle & Engle (1999: 43) maintain, ‘curriculum and extracurriculum are intimately linked.’ If instruction is given by a foreign instructor or offered by a foreign partner institution, the cultural norms of the host location will surely already be present in the classroom too—for either positive or negative effect (Brecht & Robinson 1995; Polanyi 1995; Bacon 2002; Kinginger 2004; Pellegrino 2005; Churchill 2006). Nonetheless, the general issue for the field has been not knowing what students are doing in the target language while on SA (Clyne & Rizvi 1998; Adams 2006; Kinginger 2009 2010; cf. Ranta & Meckelborg 2013). Thoughtful curriculum design ensures that learners are engaging constructively with the SA IM environment. Plews, Breckenridge & Cambre (2010: 18) criticize the curriculum of mere exposure for leaving potential learning ‘to chance and individual initiative,’ and many scholars have called for more effective program structure or curricular interventions (e.g., Bertocchini & Costanzo 1996; Brierley & Coleman 1997; Engle & Engle 1999, 2002; Wilkinson 2000, 2001; Byram, Nichols & Stevens 2001; Paige et al. 2002; Vande Berg 2007; Jackson 2008, 2010; Brewer & Cunningham 2009; Vande Berg et al. 2009; Vande Berg, Paige & Lou 2012; Bown, Dewey & Belnap 2015). After all, students do not acquire L2 skills by their mere presence in that unfiltered environment; rather, they learn ‘through individually constructing and through their membership in various cultural communities, co-constructing the world through the very act of perceiving it’ (Vande Berg et al. 2012: xv). Brewer & Cunningham (2009) identify factors pertinent to curriculum and pedagogy that impede such processing perception and successful SA: when participants remain too much in the company of co-nationals; when excursions are touristic rather than academically engaging; when students’ language skills are insufficient for effective communication; when students lack knowledge of the host culture; when they are unwilling to suspend judgment or take risks or are unprepared to work through disorienting experiences; when programs do not give students opportunities to reflect on and make sense of experiences for themselves, or make new sense of themselves.

A laissez-faire attitude to students’ out-of-class experience of IM in the L2 environment along with either importing domestic L2 class instruction into the host environment or sourcing classes offered by a foreign partner institution might exemplify one curriculum approach to SA. Beyond this, there are four ways in which SA researchers and program designers have attended to the experience of the SA environment also as curriculum and pedagogy. Thus, in the second curriculum category, there is the intentional use of the immersion environment as an extracurricular addition to a standard classroom course (Tang & Choi 2004; Brubaker 2007; Cohen & Shively 2007; Di Silvio et al. 2015). Third, there is the explicit integration of course content and IM environment (Tschirner 2007; Plews 2013). Fourth, SA is understood as a process involving the three stages of moving from one place to another and back again (Lewis & Niesenbaum 2005; Jackson 2006a, 2006b, 2008, 2010). Finally, SA is largely or fully articulated with the domestic degree as a central component and means of achieving the overall curriculum goals (Brockington & Wiedenhoeft 2009; Bathurst & La Brack 2012).

5.2 Intentional addition

By way of intentional addition to a standard course, some SA programs add features to help learners take advantage of the immersion environment. A variety of interventions are discussed in the literature, including homestay, cultural excursions, and tandem partnerships. Di Silvio et al.’s (2015) study of homestay interaction provides a rare example of an intervention in this area. Host family members received one hour of training—including understanding their role, typical conversation contexts, and talking prompts—to help encourage meaningful interaction with American students of Mandarin, Russian, and Spanish. Brubaker's (2007) investigation of a short-term SA program for upper-level German students highlights the need for curricular interventions to help learners engage with the culture. She notes that ‘while culture-specific learning was certainly addressed in and out of class, the concept and process of culture learning was not explicitly presented or discussed, as is the case with most programs’ (p. 121). As a result, the students in her study did not possess a theoretical frame or develop the vocabulary necessary to express their thoughts, feelings, experiences, and observations, and most exhibited a passive attitude toward learning about culture. Thus Brubaker suggests exercises such as an iceberg inventory, cultural and personal differentiation, and other activities described by Paige et al.’s (2002) Maximizing study abroad to help students focus explicitly on the L2 culture (see also Goldoni 2007). Such exercises are designed to assist intercultural and language learning but do not necessarily require immediate L2 use; however, Cohen et al. (2005) report statistically significant language gains in students from assigned weekly readings from the Paige et al. (2002) guide.

In SA programs designed specifically for pre- and in-service teachers, immersive cultural experiences include ‘visiting schools’ (Badia 1994: 137) or ‘teacher contact with schools in the host country’ (Roskvist et al. 2013: 7). For example, Tang & Choi (2004) describe an English IM program for Hong Kong pre-service teachers in Canada and Australia that includes fieldwork in schools. Participants were required to conduct ‘classroom observation, curriculum and lesson planning as well as individual or team teaching with the support of mentor teachers in the schools they are placed’ (p. 53), though this specialist experience was not formally linked to the IM language class or to the ongoing course of study AH. Nonetheless, Tang & Choi claim the international experience led to increased language competence and personal growth. Plews et al. (2010, 2014) investigate a residence abroad program for in-service Mexican teachers of English that required them to assist Spanish classes at school and university in Canada. The researchers find that homestay for teachers specifically with host-country teachers or education administrators strengthened professional and disciplinary knowledge and potential access to professional networks.

Engle & Engle (1999) take a more proactive stance with the extracurricular IM context by recommending mandatory activities that aim to integrate students into the local culture and encourage reflection on the home culture. In addition to the regular academic course and exploring local culture independently or with fellow students, their program in France requires American students to participate in a weekly conversation exchange with a tandem partner (see also, e.g., Badstübner & Ecke 2009; Wilkinson 2012; see Bown et al. 2015, for hired conversation partners), pursue a personal interest or hobby by joining a local club or organization (see also, e.g., Fraser 2002; Schmidt-Rinehart & Knight 2004), and contribute community service (see also, e.g., Kiely & Nielson 2003; Lewis & Niesenbaum 2005; Parker & Dautoff 2007). Engle & Engle (1999) claim that these interventions are enriching since they help students break out of the co-national group passively consuming the SA destination and instead engage with it personally. The success of these activities depends on the foreign tandem partners’ self-interest in connecting with the SA students, the students’ linguistic preparation (especially for pursuing their personal interest or hobby), student mentoring, and the fact that the students had selected the program precisely because it offered such mandatory components (for a recent review of the program, see Engle & Engle 2012). Vande Berg et al. (2009) found that the cultural mentoring in their program positively affected language proficiency and intercultural development.

Another intervention includes journaling, which facilitates reflections on homestay experiences and informal interactions with local people and culture. Plews & Misfeldt (2016) indicate that most student journals discussed in SA research are introduced as part of the research study design to gather data on intercultural awareness and not as a pedagogical intervention (in the target language) intended to enhance students’ language acquisition or linguistic awareness. In fact, Paige et al. (2002), who describe procedures for effective use of journaling, view this activity primarily as a tool for culture learning rather than language learning. However, in a small number of studies, journals are used as pedagogical interventions to facilitate language learning. For example, Bown et al. (2015) describe the use of ‘weekly speaking journals’ in which learners reflected on their interactions in the previous week and planned ways to improve their interactions in the coming week. Nevertheless, in this study, students wrote their journals in their L1. In a program described by Rollmann (2007), students wrote in their L2 to record their daily experiences and list new vocabulary. However, they were allowed to choose between the L1 or the L2 to record their impressions of the culture and to discuss problems.

Journals can also be used to help learners reflect on cultural differences, especially if, as Paige et al. (2002) suggest, they are not viewed as mere records of itinerary. Instead, the authors propose structuring journal assignments according to four categories: impressions (of people, places, things, ideas, etc.), descriptive (a picture of an event, etc.), narrative (stories of experiences), and expressive (evaluations). Ingram (2005) and Raschio (2001) describe programs in which students kept journals to record and interpret differences between their home country and the host country. In Bridges’ (2007) study, journals allowed pre-service teachers studying abroad in Australia to gauge progress in intercultural and interpersonal development, confidence in English, and awareness of personal language proficiency, and to raise critical awareness of language use, pedagogy and classroom practice, and both Hong Kong and Australian culture.

Blogging represents a contemporary form of journaling. Lee (2011) analyzes the use of personal blogs, a class blog, and project blogs by Americans in Spain to increase engagement with the target culture and interactions with Spanish speakers out of class. She concludes that the interaction with Spaniards and the collaboration required by the blogs increased students’ learner autonomy, sense of belonging, and ability to compare and exchange ideas about the two cultures, which led to intercultural competence development.

5.3 Explicit integration

The third category concerns programming that formally combines language and culture course content and assignments with engagement with the IM environment and community. Particular teaching and learning approaches are central to this endeavor, as is the intention to avoid remaining ‘touristic’ (Lewis & Niesenbaum 2005: 258). Tschirner's (2007) study of an intensive IM for pre- and in-service American teachers of German included curriculum or program structures such as knowledge about the development of oral proficiency, intensive language instruction, as well as course projects designed to increase interaction with locals and homestays and/or tandem partners. Thus, in addition to three hours of daily intensive language practice through task- and genre-based instruction, participants in this program had to work for a further three hours daily on small-group cultural research projects for in-course presentations.

Similar to Tschirner's project-based course component, Plews (2013) used task-based language teaching in an intensive advanced German language and culture IM course to raise students’ sense of personal, emotional, and creative ownership of the target language. All class tasks (e.g., drawing family trees, visiting a Turkish restaurant, mapping the local Turkish-German community, investigating the rules of romantic dating, designing local flyers) and the themes of the capstone task (individual ethnographic studies) were related to a contemporary novel read in the course. Misfeldt (2013) used the reading of a novel and drama pedagogy in an upper-intermediate course in the same program to encourage students to engage with Germans and German culture, and to build vocabulary through personal experience, discussing challenging topics with each other and their hosts, and access different kinds of intercultural knowledge and self-knowledge.

Ethnography is emerging as an especially popular and effective approach for integrating the experience of IM into the SA curriculum (Roberts et al. 2001). In preparing ethnographic assignments, students are taught techniques for participant observation, making field notes, keeping a reflective journal, attentive interviewing, and processing and analyzing data. The presentation of the ethnographic study is fundamental to completing the process. Jackson (2006a, 2006b, 2008, 2010) describes ethnographic research as a central component of a short-term SA. She chooses this approach since ‘ethnographic projects require sojourners to move beyond the safety of their group and venture into unfamiliar territory,’ ultimately using unscripted language for purposeful communication (2006b: 93). Participants in Thompson's (2002) investigation of American teachers of Spanish studying in Spain gained considerable conversational interaction with L1 Spanish speakers by conducting ethnographic interviews on topics of their own choice. These interviews helped learners refresh their linguistic knowledge, acquire cultural knowledge, and improve their attitudes toward Spaniards.

Brockington & Wiedenhoeft (2009) report on two projects that enrich students’ language, offer meaningful interaction with L2 speakers, involve learners in their own cultural learning process, increase their knowledge of other cultures, and enable self-reflection. The two projects include the Earlham College ethnographic project, undertaken as part of a field studies course or internship abroad, and Kalamazoo College's ethnographic-like project, the Integrative Cultural Research Project (ICRP), reworked from the former program. Whereas the Earlham College project tries to shape students’ worldview by having them critically compare their local placement with home, the ICRP engages with local communities in order to acquire an ‘understanding [of] how and where the activity at the project site is located in a larger cultural context’ (p. 126). Here, the L2 is used to increase the participants’ firsthand experience and understanding of the everyday lives of local people not solely for the participants’ sake, but for the sake of the hosts. At stake for the participants is achieving meaningful interaction with local people and improving analytical skills. In this project, students learn and use the ethical transformative method of DIVE: ‘Describe in value neutral terms. Interpret what is happening within the local context. Verify your interpretation with a local person. Evaluate how well it seems to be working within the local context’ (p. 126). Brockington & Wiedenhoeft underline that description and reflection are key to the success of ICRP since they help students come to terms with the external observations and challenges and internal feelings of dissonance presented by the SA context and thereby make sense of their surroundings. But it is unclear whether the description and reflection are written in the L2.

5.4 Three-stage process

The fourth way of attending to the immersion environment as curriculum and pedagogy conceives the L2 SA experience as part of a learning process or cycle of (two or) three stages that begins with guided preparation before departure, optimally includes specified engagement beyond the classroom during the sojourn, and preferably continues with a reflective and interpretive project drawing on SA experiences after returning to the home campus. This thinking underlies the culture- and language-learning strategies in Paige et al.’s (2002) Maximizing study abroad, in which activities for intercultural learning and each language skill are subdivided for pre-departure, in-country, and post-study-abroad. This category concerns a smaller group of reported programs for ‘typical study abroad programs’ and ‘do not include pre- or post-sojourn elements or specially designed components during the study abroad’ (Jackson 2006b: 94). In these examples, the pre-departure phase involves more than a general travel orientation offered by the home institution, ranging from linking to course content already discussed on campus (Lewis & Niesenbaum 2005) to a preceding course with directed readings and discussion of expectations in addition to previewing the itinerary and other practical information (Ingram 2005; Wolf 2007; Schwieter & Kunert 2012), web-based courses (Rollmann 2007), chat or other online exchanges (Godwin-Jones 2016), or an intercultural communication course and the development of research skills required for ethnographic work (Jackson 2006a, 2006b, 2008, 2009, 2010; Lou & Bosley 2012). The short-term German program described by Rollmann (2007) comprises three courses that begin online 12 weeks prior to the SA component. These courses are domestic courses revised to focus on practical vocabulary building for visiting Germany, the local history and culture of the destination city, and conversation practice (to enable students to describe themselves and where they come from) and grammar review, and are completed overseas in combination with a language class for international students and optional excursions offered by the partner institution. Dewaele et al. (2015) prove that a 2–4-week on-site pre-sessional course had a positive impact on participants’ affect and willingness to communicate by preparing them for French social and academic life.

The in- and post-sojourn components of the three-stage SA programs often require journals or the collection of material and data for ethnographic study that are then reviewed and analyzed for the purpose of term papers, class presentations, or ongoing work on the student's sense of linguistic and (inter-)cultural self (Jackson 2006a, 2006b; Wolf 2007). However, even when a contextualized pre-departure and in-program curriculum is provided, post-sojourn activities can still be lacking or concern informal choices such as enrolling in further language courses, recalling the SA experience for outreach presentations or independent study credits, attending another SA program, and opting for a major in languages (e.g., Ingram 2005; Rollmann 2007; Wolf 2007). Wilson et al. (2016) highlight the need to go beyond the typical institutional, optional reintegration session to realize the full potential of SA as experiential learning. They suggest requiring a critically reflective project that is researched while abroad, developed AH, and presented to fellow returnees, newly arrived international students, and non-SA students on the home campus. Digital technology offers possibilities for extending the participants’ reflections and insights for the benefit of future participants (Rodríguez 2010; Hampton 2015; Godwin-Jones 2016). Also, Hampton (2016) describes a novel UK Routes into Languages project in which British undergraduate students of French volunteer to collect realia with linguistic and cultural value during their year abroad to enhance language education at UK secondary schools after their return home. Her data show that the social engagement required when acting as cultural intermediaries enhanced the participants’ experience.

One of the most extensively analyzed examples of the three-stage SA curriculum process is provided by Jackson's (2006a, 2006b, 2008, 2009, 2010) work on a 5-week ‘Special English Stream’ program, which was designed to enhance the linguistic, communicative, cultural, intellectual, and interpersonal knowledge and skills of Hong Kong English majors. The program goals were based on Byram's (1997) model of intercultural communicative competence and addressed ethnocentrism, cultural world views, native-speaker communication, and the critical evaluation of both host and home cultures. The initial stage included a literature seminar with readings related to the cultural itinerary and a linguistics seminar in which students developed ethnographic research techniques and were required to conduct a ‘small scale ethnographic project to explore their own cultural world’ (Jackson 2006a: 139). The next stage included homestay, courses on English literature, language, and current affairs, a cultural program, informal interaction with other international students, and a field journal to record impressions of intercultural encounters and communication. The subsequent stage involved meetings with the researcher-director to analyze critical incidents described in the students’ diary entries and ethnographic reports. Jackson (2006a, 2006b) finds that language gains varied among the students but that those with positive attitudes and experiences invested more in building relationships and initiated communication more often, and thereby became more confident in speaking English. However, Jackson's (2009) continued examination cautions that SA is no guarantee of intercultural competence and L2 acquisition, that the two do not necessarily develop in parallel, and that some students might gain an advanced level in the L2 but remain ethnocentric. Nonetheless, she maintains that ‘well-planned pre-sojourn preparation, adequate socio-emotional support during the program, and post-sojourn debriefings can prompt and sustain deeper levels of language and intercultural learning’ (Jackson 2009: 14).

5.5 Full articulation

In the preceding category, courses or tasks within the domestic curriculum serve the SA program's linguistic and intercultural learning goals. In the fifth category of understanding the IM environment as curriculum and pedagogy, SA is fully articulated with and integrated into the domestic curriculum, as the central learning opportunity directly serving the overall curriculum goals. Examples include the aforementioned Earlham and Kalamazoo programs, whose ethnographic studies are discussed by Brockington & Wiedenhoeft (2009), as well as the SA experience in the School of International Studies at the University of the Pacific described by Bathurst & La Brack (2012). Streitwieser (2009) proposes that research-oriented undergraduate SA shape the 4-year North American degree in order to graduate culturally and globally responsible future professionals (see also Brockington & Wiedenhoeft 2009: 131). The characteristics of such programming would include pre-departure preparation aimed at training students to develop research proposals based on their personal interests and with culturally appropriate research designs for collecting and analyzing data abroad to be approved by the home institution, on-site adviser guidance in relation to being sensitive to the cultural and interactional norms of the foreign context, the ability to conduct research in an L2, and accountability through the thoughtful presentation and use of the research upon return to the home campus (see Streitwieser 2009: 402–407). Such an extensive integration of SA into the curriculum requires broad institutional structures, including an internationalized curriculum per se, scholarships reserved for SA, a variety of types of SA, and subject courses taught in the target language (see Wanner 2009). For these reasons, Wanner (2009) advocates a ‘realistic’ rather than a ‘maximal model,’ hoping that short-term programs will motivate students to incorporate language in their curriculum choices and undertake long-term SA. One example of the realistic model of L2 SA articulated with the domestic curriculum is the integration of Spanish and biology for cultural and environmental conservation in Costa Rica reported by Gorka & Niesenbaum (2001).

6. Conclusions

The future of SA research of students participating in programs in non-English-speaking countries may parallel the decline of FL study in the US (Goldberg, Looney & Lusin 2015). That is, the trend that globalization is leading to the dominance of English-language proficiency goals in educational policies due to its status as the lingua franca may be reducing the urgency of language acquisition of non-English languages. Despite the decline of FL study in the US, US-based mobile learners are the ones that are most represented in non-English-speaking countries.

This trend may, therefore, explain why the great majority of SA projects involve US-based learners of commonly taught languages. Despite this trend, UNESCO has reported new trends in global student mobility. For example,

Australia and Japan, traditional destinations in East Asia and the Pacific, are rivaled by newcomers China, Malaysia, the Republic of Korea, Singapore and New Zealand, which hosted 7% of the global share of mobile students in 2013. In the Arab States, Egypt, Saudi Arabia and the United Arab Emirates are making efforts to recruit students from abroad. These three countries hosted 4% of the global share of mobile students. (UNESCO Institute of Statistics 2014)

The current research on various aspects of the SA process in emerging education hubs includes China (e.g., Han 2008; Mu 2008; Du 2013; Taguchi, Li & Xiao 2013), Malaysia (e.g., Baharun, Awang & Padlee 2011; Knight & Morshidi 2011; Zamberi Ahmad & Buchanan 2017), Republic of Korea (e.g., Kim 2006; Beausoleil 2008; Brown 2013; Jon, Lee & Byun 2014), Singapore (e.g., Olds 2007; Collins et al. 2014), Egypt (e.g., Dewey et al. 2013; Trentman 2013; Mercer 2015), Saudi Arabia, and the UAE (e.g., Paris, Nyaupane & Teye 2014; Zamberi Ahmad & Buchanan 2017). The majority of the studies, however, focus on the reasons that international students choose to study in these regions. It is recommended that more researchers try to focus on SA experiences in these destinations.

The nascent field of student mobility research has, up to now, focused primarily on the student in the SA setting. Some methodologies use SA as a categorical label and others, more recently, as a social ecosystem (de Bot, Lowie & Verspoor 2007). SA used as a categorical label treats the experience abroad usually in opposition to AH or domestic IM. That is, it is used as the main factor to explain L2 acquisition. Taguchi (2016) states that ‘by design, these [study abroad as a black box/categorical label] studies are not concerned with what actually goes on in a SA context. Features of SA—exposure, social contact, and cultural experiences—are all consolidated into this physical context . . . without any internal components available for inspection’ (p. 5). Nonetheless, the notion of individual differences has resided in statistical analyses and results but has not been addressed more broadly. On the other hand, SA is treated as a social ecosystem that is dynamic, where interacting factors shape L2 acquisition. Researchers in this category take interacting, internal, dynamic subsystems into account to explain the L2 learning process. The complex, dynamic systems theory encapsulates this approach and ‘allows us to merge the social and the cognitive aspects of SLA and shows how their interaction can lead to development’ (de Bot et al. 2007: 19). The dynamic systems approach takes into account the effect of internal resources—‘the capacity to learn, time to learn, internal informational resources such as conceptual knowledge, and motivational resources’ (de Bot et al. 2007: 11)—to explain the L2 learning process as well as studying the role of input via situation practice and via extended, dense social networks (e.g., Isabelli-García 2006; Kurata 2011; Dewey, Bown & Egget 2012; Dewey et al. 2013; Shiri 2015). Using qualitative, quantitative, and mixed methods, research is confirming that these students are complex agents. They show ‘that learner beliefs are dynamic . . . implying that language learner beliefs are bi-directionally related to the perceptions of study abroad sojourns’ (Kaypak & Ortaçtepe 2014: 364).

King & Raghuram (2013) now call for a need to ‘decenter the student as the object of study and instead recognise the multiple players’ (p. 134) invested in and gaining from student mobility. They suggest the need for detailed ethnographic research ‘to document their complex lives in the academic, social, cultural, and economic realms’ (p. 135). This approach could include biographical research (see Roberts [2002] for more information), be multi-sited (such as the LANGSNAP project [2013]), and follow the students back home to continue post-SA research.

King & Raghuram (2013) also invite researchers to delve into the available survey data and other datasets and carry out statistical analyses, improve theorization of student mobility (e.g., Coleman 2013b, 2015; Xiao & Wray 2016), and include gendered analysis to explain the dynamics of SA experiences (e.g., Trentman 2015; King & Sondhi 2016). Jensen & Howard (2014: 32) highlight that ‘while SA has proven to be highly beneficial in certain areas, such as fluency, sociolinguistic competence and lexical development, it appears to have a less positive effect on grammatical development compared to the classroom setting.’ However, learners may evidence development if they spend longer amounts of time abroad, and it is imperative that studies include the effects of longer lengths of time on grammatical development (e.g., Nishida & Isabelli 2005; Davidson 2010; Serrano 2010; Llanes 2012; Pérez-Vidal et al. 2012; Serrano, Tragant & Llanes 2012; Jensen & Howard 2014) and differences in speed of processing (Taguchi 2015, 2016). Other suggestions are that future studies should try to collect their data at several points in time (Serrano 2010; Stokoe 2013).

In addition, SA programs vary widely in their framework, which may explain the inconclusive findings in SLA research in SA (Llanes 2011; DeKeyser 2014; Sanz 2014). For example, Grey et al. (2015: 138) note the ways a program can differ, including ‘the duration of the program, students’ proficiency in the target language upon entry, the type and context of course work, the type of student housing, and the nature and extent of opportunities for guided/structured cultural interaction, reflection, and experiential learning, among other areas (see Engle & Engle 1999).’ The publication of the Boren Report (Mason, Powers & Donnelly 2015)—focusing on non-English cognate language acquisition in SA contexts—‘acknowledged that curricular and affective variables constitute key elements, and researchers have admitted that existing on foreign soil does not necessarily guarantee active engagement with speakers of the language and members of the local community’ (Miano, Bernhardt & Brates 2016: 289).

Regardless, research focusing on linguistic constructs and gains in SA has recently become much more complex. Spoelman & Verspoor (2010: 547) observe that ‘we now know from recent studies that many aspects of language development are nonlinear and . . . multivariate and dynamic’ and that each subcomponent may interact with other subcomponents and exhibit its own developmental dynamics (Larsen-Freeman 2006; Norris & Ortega 2009). In order to meet the challenges of investigating the intricate interaction between elements, Larsen-Freeman (2009), therefore, calls for ‘more longitudinal and nonlinear research, in which difference and variation occupy a central role, and for a broader conceptual framework, such as that offered by dynamic or complex systems theory (Larsen-Freeman & Cameron 2008; Verspoor, de Bot & Lowie 2011)’ (Housen, Kuiken & Vedder 2012: 8).

A variety of SA program frameworks apply both to the out-of-classroom IM on SA as well as to the in-class and course approaches, assignments, goals, and expectations of so-called ‘island’ programs or the host culture education system. Unlike research trends, the student does not necessarily distinguish these two domains when processing language and culture, but rather draws on both and integrates the potential learning of each. Thus, to neglect one or the other in the SA research would be like reading only the even pages of a book. Somehow one might put together much of the story without the odd pages, but there will always be important missing elements that contribute to the story. While careful study of a particular even or odd page may be necessary at times, the whole story cannot be conceptualized without both. Although it may be possible to understand a key element of SA by focusing more on the classroom or more on the IM experience, both contribute to the story and are therefore necessary to understand the entire experience and outcomes.

As the preceding sections have shown, SA can be highly productive for language learning. However, there are many aspects of language learning in this setting that are not yet fully understood, and research in this area should keep pace with the broader field of applied linguistics, invoking new theories and agendas and applying new methodologies.

Questions arising

1. 1. How can the SA research community better collaborate, share data and collectively come to a clearer understanding of commonalities and variation in SA experiences?

2. 2. Given that research has largely ignored the perspectives of members of the host culture, how can scholars design research to give voice to the hosts?

3. 3. What role does literacy play in learners’ SA experiences? How does SA influence the development of literacy and other abilities not typically valued in SA?

4. 4. How do learners’ various identities shape their experiences on SA?

5. 5. Are there innovative means of assessment that might be more appropriate for capturing SA linguistic gains than currently used global assessments (such as the OPI)?

6. 6. Are there innovative means of assessment that might be more appropriate for specific research purposes or settings other than currently used methods of assessing language gains, intercultural competence development, and other outcomes?

7. 7. What role does learner self-regulation play in shaping an individual's experience abroad?

8. 8. What influence will continually evolving technologies have on the SA experience in the future?

9. 9. How will SA research knowledge change if/once SA curricula, teaching approaches, and learning materials or activities are taken more into account by researchers?

10. 10. How can we better explore the relationship between in-class teaching and learning while on SA and out-of-class experience and engagement?