The Major in Cultural Context : Choosing Liberal Arts in the Marshall Islands

Choosing a major is part of liberal arts (LA) education in American-accredited colleges across the world. In global second-language (L2) contexts, the choice of major is shaped by local cultural factors. This study of 192 undergraduates at an English-medium-of-instruction (EMI) college in the Republic of the Marshall Islands (RMI) used a survey, content and Appraisal analyses to explore what the LA major means to RMI youth. Results showed they were positive about LA, but little engaged with it outside the classroom. This probably reflected the institution’s traditional concept of LA and outdated western teaching approaches, and a failure to incorporate elements of an authentic local culture of teaching and learning. Appraisal data indicated participants associated positive, congruent desire, interest and affection for the LA major, but also low utility and worth with LA class content, revealing a need to convey the utility of the LA skill set for employment. Finally, LA majors were intrinsically, whereas education, business and nursing majors were pragmatically motivated, reflecting the colonial heritage. Overall, results foreground the colonial character of current teaching practice, and the need to use authentic teaching and learning modalities, to support RMI students’ pragmatic needs, particularly given their emigration prospects.


Introduction
Marshall Islanders have inhabited two chains of volcanic seamounts and coral atolls in the southern Pacific Ocean, for over four thousand years.During this time, they have developed complex traditions, livelihoods and social hierarchies, with a rich cultural heritage including stories, music, dance and crafts (Keating, 2000).European colonisation, beginning with the Spanish in the early 1500s, resulted in the extensive deracination of locals within their ancestral islands (Hezel, 1995;2001).In the 20 th and 21 st centuries, the Marshallese have faced environmental, economic and linguistic change (Lee, Balick, Ling, Sohl, Brosi, & Raynor, 2001) 2009), assessment of their own abilities (Arcidiacono, Hotz, & Kang, 2010), and comparisons with peers' abilities (Wiswall & Zafar, 2011).LA has roots in classical philosophy, medieval rhetoric and Enlightenment civic ideals.It delivers broadly-applicable practical competencies including writing, speaking, critical thinking, ethics, civics, and intercultural sensitivity, rather than specialist, discipline-specific theoretical and methodological knowledge (Bogue & Hall, 2003).The LA major is viewed as less valuable in contexts where specialisation and professionalisation are culturally normal (Farrell & Vander Werf, 2007).Globalisation has increased the demand for a return on the cost of education in terms of employment (R. K. Koshal & M. Koshal, 1999).American-accredited colleges worldwide have responded by reframing LA competencies to support professional, business and science needs (Pascarella & Terenzini, 2005).These competencies are highly desirable in both western and global work environments (BHEF & ACE, 2003).Yet most research into major choice, including the choice of LA, has occurred in western contexts and focused on western students.Western-style college education is now a global phenomenon (Brecht & Ingold, 2000).But teaching and learning modalities vary across cultures."We have recently become more sensitive to the understanding that knowledge does not exist in objective, decontextualised form but is intimately linked to specific contexts, people and issues.This understanding is particularly relevant for indigenous people whose systems of knowledge have been subordinated by the forces of colonization" (Abdullah & Stringer, 1999).Perceiving what LA majors in non-western cultures are learning has been likened to "grappling with Godzilla" (Stone & Friedman, 2002).This article explores what Marshallese college students report about their choice of the LA major.
Western models of intellectual achievement, such as social cognition and self-efficacy, foreground an instrumental narrative of individual beliefs, and their instantiation in learning actions, with resulting outcomes (Eccles & Wigfield, 2002;Bandura, 1997).In western contexts, students' choice of major is often connected to pragmatically motivated expectancies regarding the employment value of the major (Eliot & Dweck, 2005).The role of intrinsic pleasure in learning, based on personal interest in the subject, plays a secondary role (Wentzel & Wigfield, 2009).The paradigmatic distinction between instrumental motivations for achieving professional goals, and intrinsic positive attitudes towards the content, underwrites current research into the motivations students feel, that keep them invested over the long term in the difficulties of college-level learning (Schunk & Zimmerman, 2007).However, these theories were developed within western cultures where school-based education is an established element of childhood, family and social life, and a bridging element for youth entry into employment.These models are not useful in understanding other global and cultural contexts.The function of western education in the Pacific has often been to reinforce colonialism (Falgout, 1992;Pene, Taufe'ulungaki, & Benson, 2002).In the Pacific, western education has tended to limit learner autonomy, entrenching identity loss and leading to poor performance (Burnett, 2005).School-based education in these contexts has been associated with the loss of language, and of authentic, local ways of knowing (Burnett, 2003).
Cognition cannot be separated from its cultural contexts, as it is local practices that give rise to understanding (D'Andrade, 1995).Pacific models of teaching and learning integrate field-specific knowledge with practical skill (Williamson & Dalal, 2007).They valorise oral over written communication, integrate ethical with epistemological insights, and understand the community more than the individual as the unit of production of knowledge (McLaughlin & Whatman, 2008).Learning styles involve inductees listening in on the edge of actual, ongoing learning situations, which they then enter and participate in, in the manner of "firsthand learning" in native American communities (Rogoff, Paradise, Arauz, Correa-Chavez, & Angellilo, 2003).Indigenous Pacific models of learning share elements of situated learning, where theory is understood procedurally via its instantiation in the activities of daily life, and where outcomes resemble the long-term process of moving from novice towards expertise, rather than acquiring a body of declarative knowledge (Lave & Wenger, 1991).They emphasise authenticity, where learning takes place in real contexts beyond the classroom, learning activities require negotiating real circumstances, and learning outcomes require shared responsibilities in selecting and achieving learning activities (Richardson, 2001).These practices do not fit easily into classroom-based, disembodied and decontextualised abstract information acquisition, separated from the activities and events of daily life and the shared wisdom of the community.
In some Pacific contexts, the classroom and the associated style of teaching and learning, are recent implants (McConaghy, 2000).In RMI, the Education Act of 1991 led to the hiring of foreign native-speaker English teachers, enhancing the schools-building begun in the 1960s (Ministry of Education, 2005).This diminished the importance of traditional ways of teaching and learning, while at the same time the benefits of EMI, classroom-based learning remained uncertain (Thaman, 2003).Thus, while 12 years of schooling are compulsory, RMI school truancy rates remain above 20% (Hickman, 2017).Rapid change impacts youth wellbeing (Lowe, 2003).Coming of age in societies undergoing extensive change creates psychosocial stress for Pacific island youth (Mayeda, Okamoto, & Mark, 2005).Marshallese youth face an elevated risk of alcohol, drugs and suicide, creating barriers to college-level academic performance (Masten & Coatsworth, 1998;Booth, 1999).EMI college education in the Marshall Islands is supported by universal access to PELL grants, granted under the 1986 Compact (Ministry of Education, 2005).But while US aid has funded the required infrastructure and financially enabled student attendance, it has also inculcated passive learning habits, undermining Marshallese students' ability to control their own learning, and thus their employment prospects (Chutaro & Heine, 2003).RMI youth function in a complex learning situation.Indigenous culture shapes what and how they learn (Oberle, 2009).Their own language is the code through which the local repertoire of social roles and relations is acquired and enacted (Keating, 1998).But L2 education requires culturally-specific interactions with foreign teachers, and non-native linguistic and cultural performances (Waller, 2001).An L2 is a medium through which people recontextualise their former values, construct and legitimise other ideals, and test alternative practices (Woolard & Schieffelin, 1994).EMI education tends to valorise global over authentic local cultures and insights, assimilating L2 students to westernised models of self and society (Mauranen & Ranta, 2009).Most L2 youth do not want to assimilate to a specific geographic or national native-speaker identity, for example British or American, preferring to blend elements of local and international cultures into transnational identities (Levitt, 2001).Identity formation is harder when youth must construct viable versions of self congruent with both traditional and competing values and practices (McDade 2001).It is in the context of these tensions, that Marshallese college students choose a major.
This study explored the reasons RMI youth give for choosing the LA major.Clearly, LA and the western models of teaching and learning that have shaped it, are not well-suited to the RMI context, yet large numbers of RMI students select this major.As institutional regulations require all college students to select a major before they can be registered, this study probed what the LA major meant to RMI youth.Factors explored include student concepts of LA teachers, LA learning, their self-image as students, LA course content, classroom experience and course offerings.Research questions explored in this study were: What reasons do Marshallese youth report for choosing the LA major?How do these differ from those given for choosing nursing, education and business?How do these groups compare, for their pragmatic and intrinsic motivations?How can this understanding be used to help Marshallese youth use a classroom-based culture of teaching and learning to their advantage?

Method
Both quantitative and qualitative data were collected, from students at an RMI college where selecting a major was required in order to register.

Participants
In all, 192 students participated, of whom 108 = 56.25%were LA, 35 = 18.23% were Education, 27 = 14.06% were nursing, 19 = 9.90% were business majors and 3 = 1.56% did not identify their major.All had either tested out of the English-language placement test, or gone through required preparatory sub-degree coursework.Thus, all had attained at least an upper-intermediate proficiency level equivalent to CEFR B1 or IELTS 4.5.All were freshers or sophomores registered in an LA course.

Instrument
Quantitative and qualitative data were collected through a 54-item, four-part survey.A 4-point Likert scale was used, to prevent social desirability bias and clearly differentiate negative from positive responses (Leung 2011).Likert scale values were 1 "strongly disagree", 2 "disagree", 3 "agree" and 4 "strongly agree".Part 1 contained 2 clusters of 4 items focused on campus facilities, Part 2 four clusters of 4 items each on LA course teachers, and Part 3 three clusters of 4 items on LA student learning.Part 4 contained 3 clusters of six subjective terms on LA students' subjective responses.Data was collected mid-spring semester, after students had become used to their selected courses but prior to any exam stress, to ensure that questions elicited responses to question content.One qualitative question prompt eliciting written response was also included on the survey.

Quantitative Data
Quantitative data was elicited through the survey.Qualitative data elicited on Part 4 of the survey used subjective terms selected from Appraisal analysis, a semantically delicate form of sentiment analysis derived from systemic functional linguistics (Martin & White, 2005).Two affect terms identified emotions which have been identified in previous research as fundamental to intrinsic motivation to learn; "interesting" (Dis/satisfaction-interest) and "enjoyable" (Dis/satisfaction-pleasure) (Pekrun, Goetz, Titz, & Perry, 2010).These tested personal satisfaction as a learning motivation.Being directly affective terms, they invited participants to define the degree of their congruent response, meaning the direct connection between the self and the emotion.Responses to prompts framed in this way are the most reliable indicators of personal feelings, as they "bear a natural relation to the meanings they have evolved to express" (Halliday, 1985, p. xviii).Two further terms elicited indirect response, with prompts framed as judgments; "easy" (Social esteem-tenacity), testing personal satisfaction, and "useful" (Social esteem-capacity), testing pragmatic motivations for learning.Finally, two more terms elicited Appreciation, "challenging" (Composition-complexity) and "worthwhile" (Valuation-worth), both testing pragmatic motivations.Participants rated LA course offerings they might take in future, the content of courses they had already taken, and the classroom experience of courses they were currently taking, using these 6 terms, of which half tested for intrinsic and half for pragmatic motivations.

Qualitative Data
Qualitative data was collected through two prompts eliciting written comment: "What is your major, and why did you chose it?"and "Is there anything you want to add?" Written responses were aggregated into a corpus, and analysed in two ways; content analysis and Appraisal analysis (Halliday & Matthiessen, 2004).Examples were selected to illustrate significant content themes and Appraisal subcategories.Examples have not been edited for grammar, due to the potential impact on both the content and the sentiment they convey.Content analysis reformulates corpus textual data as a set of thematic frequency scores (Smith, 2000).The corpus is surveyed to generate a coding frame (Weber, 1985).Frequencies are then counted at the level of the clause (Carley, 1990).Clauses are coded for a single theme, with cross-clausal transitivity contributing to frequency scores (Neuendorf, 2002).As most responses to the two prompts were brief, containing 1-3 sentences, N-gram dichotomous word-stem values were not counted (Hopkins & King, 2010).Constrained themes focus on relevant content, which is more useful than coding a corpus for its complete ideational meanings (Feldman, 1994).In this study, only themes with a direct bearing on the question were included in frequency scores, generating more conservative but more reliable frequencies (Bazerman & Prior, 2004).Six frequently-realised themes were identified, as in Table 1.Two researchers independently coded the corpus.Inter-coder reliability scores were calculated using ReCal2 (Freelon, 2010), defined by Cohen's κ (percent-overall and free margin) (Lombard, Snyder-Duch, & Bracken, 2004).
Corpus data was also analysed using Appraisal analysis.Sentiment analysis quantifies the subjective lexicogrammar realised in texts within multiple semantic subcategories.Fundamental emotions are experienced by all people, and encoded in all languages, with specific emotions eliciting signature brain states (Wurm & Vakoch, 1996).All languages offer resources to express emotion directly ("I like it"), or indirectly by dispersing subjective attitudes throughout the elements of a sentence ("I choose it because I'm sure I'm capable of doing certain classes and it is what I want to do to graduate college with the AS degree").Appraisal analysis organises these resources into 3 sets, 9 categories and 14 semantic subcategories, as in Figure 1.The Affect set, comprising the lexis through which we realise emotion, underlies the Judgment and Appreciation sets.The latter taxonomise realisations which rework emotion indirectly as subjective statements about persons, events and objects outside the self (Martin & White, 2005).Judgments realise social norms and expectations (social esteem), and evaluations of people and behaviours (social sanction) (Whitelaw, Garg & Argamon, 2005).For example, "LA.It's easy" realises a judgment of tenacity, or how much effort is required for the task adjudicated.Judgments use the writer's past experiences as a background against which to compare the experiences under review (Oatley, 2006).Appreciations comprise the words and phrases used to describe and characterise objects and events.In the example "it is very general: it allows me to explore many courses/subjects & decide what I really want to study", "general" realises a quality.
"Appraisal theories of emotion have gained widespread acceptance in the field of emotion research" (Kuppens, Van Mechelen et al., 2007).The Appraisal system network is "not arbitrarily posited" (Bednarek, 2009).This classification system has gained validity from the increasing convergence of linguistic and psychological models of subjectivity (Oatley, Keltner, & Jenkins, 2006).Automated sentiment-tagging techniques are well-established in corpus linguistics (Polanyi & Zaenen, 2006).Emotional lexis is reliably identified by software (Scherer, Schoor & Johnstone, 2001).These softwares are sophisticated through lexical concordancing (Bednarek, 2009).Concordances constructed from supervised classification tasks taxonomise thousands of linguistic instances, using statistical machine-learning and natural language processing techniques (Wiebe, Wilson, Bruce, Bell & Martin, 2004).Computational linguistic analysis is thus "robust, result

Results
Content analysis was independently undertaken by two researchers, and inter-coder reliability scores then calculated using Cohen's κ, for which the percent-overall value is weighted to compensate for the possibility of agreement occuring by chance, and the free margin value reflects the degree of disagreement (Lombard, Snyder-Duch, & Bracken, 2002).Appraisal analysis was then undertaken by two researchers and inter-rater reliability calculated using the same metric (Lombard, Snyder-Duch, & Bracken, 2002).Combining content with Appraisal analysis offers insight, as it identifies both thematic content and subjective stance (Krippendorf, 2004).

Quantitative Data
Averages were calculated for survey items.These values were treated as Likert-like.The interval character of Likert scale data is artificially-defined, and does not reflect attributes within the data set.But participants do not assume that intervals are discrete, when answering Likert-type questions (Goldstein & Hersen, 1984).Therefore, "parametric statistics are robust with respect to violations of these assumptions" (Norman, 2010).Ordinal values may thus be treated as continuous when question items involve subjective factors and when data is taken from homogeneous participant groups (Lubke & Muthen, 2004).Treating this data as Likert-type facilitates the use of parametric measures (Ary, Jacobs, & Sorensen, 2010).
The average for all items was 3.06, slightly above "agree", as in Figure 2. Part 1, with 8 items about campus facilities, was rated averagely at 3.08.Part 2, with 16 items about LA teachers, was rated averagely at 3.06.Parts 3 and 4 had diverse results.Part 3 focused on LA students.The lowest average was found for the 4 items about student learning, at 2.84, but students rated their participation in learning activities averagely at 3.07, and learning activities above average at 3.56, between "agree" and "strongly agree".They rated their class preparation between "disagree" and "strongly disagree" at 1.88.That is, students disagreed that they (a) read the textbook in advance, (b) spent time preparing for upcoming classes, (c) checked with a classmate about what had been done if they missed a class, or (d) reviewed past work to prepare for the next class.It is notable that students rated themselves the lowest of all parts and items in the Survey.Part 4 focused on students' subjective attitudes.Here, the 3 clusters of 6 items checking students' subjective responses to course offerings, course content and classroom experience, were rated higher than average (3.24).Of these, classroom experience was rated highest at 3.61 between "agree" and "strongly agree", course content above average at 3.19, but course offerings below average at 2.91, below "agree".
Figure 2. Survey averages by part and cluster Data for subjective items shows that the congruent, directly-realised Affect subcategories "interesting" and "enjoyable" were rated above average, as in Figure 3.
Figure 3. Subjectivity item averages By comparison, "easy" and "useful" were rated below, where other terms were rated above the 3.06 average, "agree".Congruent Affect ("interesting", "enjoyable", 3.45) and Appreciation terms ("challenging", "worthwhile", 3.33) exceeded the overall average by a third of a scale degree or more, where Judgment terms (2.74) were rated lower than average, at less than "agree".Terms testing pragmatic motivations were rated averagely at 3.05, where those testing intrinsic motivations were rated above average at 3.29.These results may be explored in terms of the student experience, as in Figure 4. Students assessed LA courses as not useful, prospectively for courses they might take, and retrospectively regarding course content they had taken, but rated courses they were currently enrolled in as averagely useful.The value for "useful" was overall slightly above average, though lower than for all other subjective attitudes about classroom experiences.This identifies a large subjective difference between LA courses viewed in the imagination or memory, and those being currently experienced.Courses taken previously were recalled as below average for utility, easiness and worth of which two are pragmatic motivations, but above average for enjoyment, challenge and interest, of which two are intrinsic motivations.Recalled course values fell into above-average (interesting, enjoyable and worthwhile, of which two are intrinsic), average (easy and worthwhile, one intrinsic and one pragmatic) and below-average (useful, pragmatic) groups, and were overall below average between "agree" and "disagree" (2.56).Future possible courses were viewed as challenging (pragmatic), with values for all other subjective items were overall below average between "agree" and "disagree" (2.76).Classroom experience values were most similar to each other, rated at above-average between "agree" and "strongly agree" (3.56, σ=0.22).Thus, participants reported their current class experience as more than averagely enjoyable, worthwhile, challenging and interesting, and averagely easy.

Qualitative Data
Corpus data was analysed using content analysis, as in Figure 5.This data identified specific reasons why students chose their major.The survey prompts elicited a corpus of 4,197 words, containing 832 clauses (5.04, σ=0.13), 667=80.17%with thematically relevant content.165=19.83%were excluded as irrelevant (for example "put better food in food store", "there is nothing I want to add").Within the corpus overall, ethical ("I've always wanted to help sick people", "to help all the upcoming students") and pragmatic ("I'm interested in the accounting field", "I wanted a more stabled start") reasons accounted for about half of all reasons given across all majors (48.47%) (see Appendix 1).Easiness, an intrinsic motivation, comprised 9.69% across all majors, but 17.59% among LA majors.
About one third of LA majors were undecided on a career at the time of choosing ("Still undecided on my field of study", "I chose LA major since I was undecided by the time I enrolled", "what I really want for my career in still in question").About 1/5 said their first choice was unavailable ("I choose this major, because I didn't know what to choose after I discovered that there are no IT course").Desired but unavailable majors included anthropology, marine science, law, IT and engineering.Intrinsic ("easy") and pragmatic motivations each accounted for about 15% ("Liberal Arts is my major, because it's the easiest and quickest way out of here", "LA.Is easy").However, most pragmatic motivations involved transferring to another institution for further study ("if I get succeeded in this major and further my studies overseas", "Liberal Arts, because out of all the majors offered the credits for Liberal Arts are more likely to get transferred"), and speed of graduation ("I chose LA major indicate of graduate early", "I choosed LA because it is has a free classes, and I am graduate and get my first degree early") rather than career plans.Other reasons included disliking other options ("I choose this major because I don't like education, business, and nursing", "I'm not interested in education and Nursing.Plus I suck at math so NO to Business").
Figure 5. Reasons for major choice in the whole corpus, and in liberal arts Appraisal analysis was used to identify frequently-realised subjective attitudes within the 2,180 words comprising the LA major portion of the corpus, for the first prompt.Of 194 attitudes realised, 135=69/59% were positive and 59=30.41%negative.Of a possible 24 semantic subcategories, 18 were infrequently realised at <5%, where 6 comprised 2/3 of all attitudes realised, as in Table 2. Three affect subcategories, desire, interest and affection, comprised 71=55.04% of frequently-realised attitudes, highlighting the significant role of emotion in selecting the LA major.Impact is the nearest subcategory to affect outside the affect set (Martin & White, 2005).For example, in "I choose this major because my goal is undecided", the word "undecided" applies grammatically to "goal", but must be understood as the author's feeling.Such cases comprise reactions of impact.This suggests that emotional states may account for 93=72.09% of frequently-realised attitudes, and play a significant role in selecting LA.Emotions including desire (" I want to learn many things from it", I want to study more about things that I haven't studied."),interest ("I want to learn different courses that it offer which I find will be something interesting to me to learn", "I was really interested by learning new things that are knowledgeable, and artistic"), and liking ("I like major because courses we had learned", "I like it … them etc…") reflect intrinsic motivations to learn.Pragmatic motivations were seen in judgments of capacity ("I choose it because I'm sure I'm capable of doing certain classes", "if I transfer I can [take] any major I want in other college"), and appreciations of worth ("this would help me explore first and then decide later," "it helps us to find what we fit in for future career"), which together comprised 36=27.91% of frequently-realised attitudes.

Reliability
Inter-coder reliability scores for content analysis, and inter-rater reliability scores for Appraisal analysis were calculated, as in Table 3.As all scores were greater than 0.80, results are reliable and not attributable to chance.

Discussion
Five issues were apparent in these results.First, participants' strongly positive response to LA, and strongly negative reviews of themselves as students must be read as a product of a system which does not embrace local, authentic models of teaching and learning.This distressing consequence of western models of teaching and learning could be alleviated, if greater sensitivity was shown to context and culture.It also suggests that the survey's clearly positive overall ratings should be received sceptically, as participants' responses to some degree reflect their acceptance of LA's self-representation.At the institution studied, LA course offerings conventionally included literature, history, world religions and popular culture, with an outmoded emphasis on textual analysis and information acquisition.Learning objectives were poorly-connected to student abilities or local culture, content levelling was idiosyncratic and inapplicable to the point of fictionality, and syllabi retained archaisms such as gendered language.The LA philosophy was minimally articulated, in a desultory manner rather than reflecting any genuine engagement with it.Departmental teaching and outreach practices lacked any mechanisms to convey the value of LA competencies to employment.The scant curricular mapping that had been undertaken was cosmetic in character, and did not reach into classrooms, or shape activities, content delivery, project definition or the learning experience.Few Marshallese students have the international experience required to evaluate their situation.Few have any means of determining the utility of LA competencies for the US employment context they are destined to become part of.Few have any previous experience to draw on in connecting LA outcomes to their own interests and ambitions.Their positive overall ratings, then, reflect the failures of institutional leadership and departmental management, as much as their own experiences.This is the style of educational leadership frequently critiqued in Pacific contexts as colonial.
Second, survey results revealed that students did little out-of-class learning.Most respondents disagreed or strongly disagreed that they read, reviewed or prepared for upcoming sessions.Again, this reflects the low degree of institutional engagement with local culture, which could be instantiated in curricular design and classroom practice.While students rated their subjective experience of the LA classroom highly, their self-review begs the question, what do these majors learn?The data reflects their view that LA is interesting and fun, but not worthwhile.The high ratings overall may reflect student comfort with LA classrooms that replicate the passive learning habits of their previous schooling.That is, the American curricular content and pedagogical architecture intended to achieve western accreditation in fact constitutes a kind of camouflage, concealing the lack of learning, but more significantly, the lack of effective teaching (Sauer, 2014).At this institution, a lack of "professional integrity" at the top of the administrative order had diminished the "trust and confidence of faculty" (Chronicle of Higher Education, 2008).This, along with poor teacher qualifications led to a revolving door problem, short-term stays and annual hiring cycles, and eventually the loss of accreditation (CMI, 2006).In western contexts, LA is known for re-inventing itself, with reform focused on student learning outcomes clearly connected to utility and employment prospects (Schneider, 2004).LA should be able to embrace the situated, authentic and hands-on learning that would support Marshallese in acquiring specific competencies.This study has foregrounded the need for LA in global contexts to reflect authentic, local culture, rather than imported, colonial assumptions and practices.
Third, unlike their western counterparts, Marshallese LA majors are intrinsically rather than pragmatically motivated.This result identifies a significant discrepancy between LA students' motivations, and their future employment needs and the purpose of their college education.LA majors' intrinsic motivations cannot be understood to reflect Marshallese culture or learning styles, as Nursing, Education and Business majors were clearly pragmatically motivated.Given the challenges facing Marshallese people, LA majors need to be prepared to perform in demanding new employment contexts.Schools-based, pre-college outreach programs could be used to increase college preparedness, and prep-year and fresher intervention strategies, goal-setting and achievement metrics could be used to support at-risk students (Cabrera et al, 2006).Authentic course offerings and learning outcomes could be developed (Knight, 2001).Strategies for communicating LA's pragmatic competencies could be incorporated into advertising, course materials, student advising and career counselling (Zins, 2004).
Fourth, the qualitative data showed that many LA majors are undecided about their intended career ("I chose this major because I don't really know what I want to do yet but I'm pretty sure I'll find out what I want to become soon").Their subjective indecision was mirrored in their view of LA courses as informal, relaxed, and indistinct ("LA.Is easy," "because I want to graduate early," "I chose this major because it is very general," "it could be anything", "this would help me explore first and then decide later").Cultural attitudes vary in how much they tolerate social loafing, the tendency for individual learners to expend minimal effort within a group (Clark & Baker, 2011).However, mechanisms such as task accountability, structuring tasks to ensure role indispensability, and evaluating component, individual and group performance have been found to enhance individual investment in learning (Aggarwal & O'Brien, 2008).In collectivist cultures, teachers may also appeal to the group's self-understanding as inter-dependent, and individuals' personal sense of satisfaction, to boost individual learning performance (Teng & Luo, 2015).
Finally, the characterisation of LA courses as impractical requires interrogation.Marshallese LA majors' ratings were not self-consistent.Participants prospectively and retrospectively rated LA courses as less than averagely useful, yet their classroom experience received the strongest subjective endorsement overall, as useful.This supports a view of Marshallese as experiential learners, and suggests LA majors are intuitively but not explicitly aware of the utility of competencies, such as writing, speaking, critical thinking, citizenship, and intercultural sensitivity.These majors could be helped to appreciate the pragmatic value of LA competencies, if courses were reframed using co-operative, or project-based learning (Smith, Sheppard, Johnson &Johnson, 2005).Outcomes and their utility could be defined and showcased at the departmental level (King, Brown, Lindsay & Vanhecke, 2007).This approach might also improve ratings for course offerings viewed prospectively as challenging, which may indicate the difficulty of imagining what kind of thing is meant by course descriptions.Overall, the institutional failure to understand its students' views and motivations, and adopt learner-centred and authentic approaches must be seen as a form of continued colonialism in education.
This study has reached four conclusions.First, results highlighted the need for reframing LA teaching using authentic, functional and dynamic approaches, to counteract currently low learner investment in learning.To support students in appreciating LA competencies, more contemporary and employment-relevant objectives and outcomes should be identified for LA courses.Community-based mechanisms could be found, which enabled students to experience that utility, connecting it with their own future employment prospects.Second, results highlighted the likelihood that current LA teaching and learning modalities reproduce passive-learning high-school norms.To support active learning, courses could be reframed to emphasise student-centred approaches, and service-, experiential-, and project-based learning.Together, these conclusions characterise the current teaching and learning culture as colonial in character.Third, results highlighted the power of intrinsic motivations for Marshallese LA majors, which offers teachers many opportunities in achieving more authentic and relevant learning objectives and outcomes.A comprehensive curriculum mapping exercise could link content, classroom activities, projects, objectives and outcomes, providing culturally relevant and institutionally realistic vertical coherence.At the same time, pragmatic competencies could be worked in, as this motivation is culturally strongly attested.Fourth, the low self-esteem of LA majors likely reflects the current colonial-style, teacher-centred LA teaching culture.Marshallese students would likely be more engaged in acquiring LAs competencies, if LA courses used authentic local content and activities, and delivered content via modalities more closely resembling the pragmatically-oriented nursing, education and business programs, which offer practicums, field trips, and community projects.
This study has limitations, which further study could address.A larger sample and explicit comparisons with other majors would ensure greater accuracy.Survey items could be added probing student-teacher relations, student attentional processes during input and activities, and student and teacher perceptions of the difficulty levels and employment value of input and activities.Comparative studies of different majors could detail motivations, concepts of teaching and learning, and self-esteem.This initial study has highlighted the need for substantial changes in the teaching culture and curriculum.Longer-term studies could explore where and how such interventions were effective.
The people of the Marshall Islands have a long history of colonial maltreatment, and face quite extraordinary challenges, as they move forward into their uncertain future.The LA major has proven its ability to instil confidence, model lifelong learning, and build resilience.Equally, colonial education offers to entrench identity loss and under-performance, which must impoverish all.Rarely have the Liberal Arts had such an opportunity to demonstrate their worth.

Figure 1 .
Figure 1.The attitude system

Figure 4 .
Figure 4. Subjective attitudes to course offerings, course content and class experience

Table 1 .
Content-coded reasons for choice of major

Table 2 .
Frequently-realised attitudes by LA majors

Table 3 .
Inter-coder and inter-rater reliability scores