Student Definitions of Intercultural Competence (IC)- Are They Context-Specific?

Higher education institutions increasingly seek to promote students’ intercultural competence (IC), yet its conceptualization remains a challenge. The first aim of this study was to explore how a purposive sample of n=77 domestic, undergraduate students at a public university in Germany define IC. The second aim was to assess to what extent such definitions are context-dependent by comparing IC definitions provided by domestic students (with predominantly German nationality and little international experience/exposure) with those provided by n=130 international, undergraduate students at a private university in Germany (with predominantly non-German nationality and more extensive international experience/exposure). A qualitative content analysis showed that domestic students defined IC mostly in terms of attitudes followed by external outcomes while the opposite was true for the international students. Differences in IC definitions between the two samples are discussed with regard to various contextual factors. Our results suggest that such contextual factors should be considered when designing measures to promote IC development in university students.


Introduction
Intercultural competence (IC) has become essential for successful participation and navigation in professional and social life in today's ever more diverse societies and workplaces.Many higher education institutions seek to find ways in which such competence can be taught.In Europe, IC is increasingly considered an interdisciplinary soft-skill that students across programs should acquire and develop during their studies-a demand reinforced by both the labor market and the Bologna process (Erll & Gymnich, 2015;Hiller, 2010).Due to rising student and staff mobility, university campuses (especially in popular study-abroad destinations such as the USA, Canada, the UK, Australia, or Germany) have become more multicultural.This holds a vast potential for intercultural learning and students' IC development although many scholars have concluded that this potential is yet to be fully realized (Bennett, Volet, & Fozdar, 2013;Bosse, 2010;Deardorff, 2006;Harrison & Peacock, 2009;Ippolito, 2007;Leask, 2009;Otten, 2003).Despite the fact that many higher education institutions seem to consider IC development as desirable, there is evidence that they often seem to lack a clear understanding of what such competence means and how it can be realized (Deardorff, 2006;Ippolito, 2007;Otten, 2003).Research has consistently documented that there is only minimal interaction between domestic and international students implying the need for actively introducing intercultural learning opportunities into students' curricular and/or extracurricular life (Campbell, 2012;Halualani, 2008;Harrison & Peacock, 2009;Summers & Volet, 2008).However, to make such opportunities appealing to students, we argue that it is important to understand how they conceptualize IC and its relevance for their own lives.
Since the emergence of the concept in the 1950s, scholars across academic disciplines have suggested a plethora of IC models (cf.e.g.Bolten, 2007;Spitzberg & Changnon, 2009).A major concern about the available models is their potential ethnocentricity as most of them have been developed in Western, Anglo-Saxon contexts (Spitzberg & Changnon, 2009;Wang & Kulich, 2015), often considering the construct of IC 'in a vacuum devoid of context' (Deardorff, 2009, p. 267).In their review, Spitzberg and Changnon (2009) conclude that it is difficult to ascertain at present 'the extent to which such contexts may bias or shift emphasis' (p.43) in understanding IC.Likewise, Moosmuller and Schonhuth (2009) have discussed the German discourse on IC which emphasizes 'context boundedness' (p.211) of the concept.This context boundedness, however, remains to be explored with studies focusing on emic conceptualizations of IC in various contexts to assess to what extent context can shift emphasis.In recent years, scholars have discussed culture-specific, emic notions of IC in the African, Arab, Indian, Latin American, Chinese, and German context to complement, not replace, mainstream models -yet, all of these contributions are conceptual and lack empirical studies to explore how people in different contexts define IC (Chen & An, 2009;Luo, 2013;Manian & Naidu, 2009;Medina-López-Portillo & Sinnigen, 2009;Moosmuller & Schonhuth, 2009;Nwosu, 2009;Zaharna, 2009).A notable exception is a study by Arasaratnam and Doerfel (2005) who took a qualitative approach to understanding IC by gathering vernacular descriptions in interviews with 37 American and international student and non-student volunteers affiliated with an American university.Using semantic network analysis, they identified eleven unique dimensions of IC which represent elements common across cultures in their study (Arasaratnam & Doerfel, 2005).
Building upon the idea of taking a qualitative, descriptive approach to how people understand IC, this study seeks to make an empirical contribution while shifting the focus from national or regional cultural contexts to the university context where we argue for exploring (1) how certain groups, in this case students, conceptualize IC and (2) if their emphases in conceptualizing IC depend on their immediate context.Put differently, we are interested in the extent with which student definitions of IC depend on the specific context in which these students live and learn.Underlying this research is the assumption that universities will be better able to facilitate students' IC development if they understand students' experience and reality and the ways in which IC is relevant to them.To this end, our aim is to collect and compare data from two different student samples, a domestic sample of German students studying at a state university in Northern Germany (this study) and a culturally diverse sample of international students studying at an international, private university in the same city in Germany (Odag, Wallin, & Kedzior, 2015).

Theoretical Models of IC
Over the past decades, IC models have evolved from list models focusing on personality traits, to more structural models acknowledging the multi-dimensional nature of IC and its affective (motivation, attitudes), cognitive (knowledge) and behavioral (skills) aspects (Bolten, 2007;Spitzberg & Changnon, 2009).More recently, scholars have become interested in the process or interplay of these different aspects in intercultural situations as well as in internal (self-reflection, frame of reference shift) and external (effective and appropriate interaction) outcomes of IC (Bolten, 2007;Deardorff, 2006;Spitzberg & Changnon, 2009).Extensive reviews of contemporary models of IC already exist with the majority of these models having been derived theoretically (Spitzberg & Changnon, 2009).Against this background, the current study aims to empirically test (1) to which extent university students' subjective understanding of IC align with IC models and (2) whether there are similarities and differences in definitions between the two different samples of students who live in the same city and yet study in two different contexts.
To be able to address aim 1, we anchor our study in the empirically-derived theoretical framework by Deardorff (2006), the so-called Pyramid Model of Intercultural Competence, which was chosen for three main reasons.First, although the Pyramid Model was derived in the USA, it synthesized the diverse approaches and models of IC.The model is thus widely accepted beyond the U.S. context in Europe, including Germany (Bolten, 2007;Schumann, 2012;Van de Vijver & Leung, 2009).Second, the Pyramid Model was empirically derived from definitions of predominantly-US based researchers and administrators in the field.It therefore provides a base for our own empirical study allowing us to assess students' understanding of IC and to compare these student definitions to the expert definitions in Deardorff's study (2006).Finally, the Pyramid Model was also used as a framework for our earlier study examining definitions of an international student sample (Odag et al., 2015) and, thus, provides a shared theoretical backbone for the comparison between the (existing) international sample and the (new) domestic sample of students.Deardorff (2006) took an empirical approach to the question of defining IC in the university context and established consensus on definitions (and assessments) of IC in a three-round Delphi process with 23 leading scholars in the intercultural field as well as administrators of 24 higher education institutions in the USA.Her research yielded agreement with the broader definition of IC as 'the ability to communicate effectively and appropriately in intercultural situations based on one's intercultural knowledge, skills, and attitudes' (Deardorff, 2006, p. 248).This definition captures three elements found in most contemporary definitions or models that conceptualize IC as (1) a multidimensional competence (comprising cognitive, affective, and behavioral dimensions), (2) a culture-general competence and (3) a competence that facilitates effective and appropriate (intercultural) interaction (Deardorff, 2006;Spitzberg & Changnon, 2009).Culture-general in this context means that IC is not specific to any particular culture, but enables the individual to understand cultural differences, their effect on interpersonal behavior, and how to manage such differences to be effective and appropriate across various intercultural situations (Graf, 2004).Effectiveness refers to 'the achievement of valued objectives or rewards' while appropriateness pertains to 'avoiding the violation of valued rules or expectancies' (Spitzberg, 1989, p. 250).In addition to this general definition of IC, scholars in the Delphi study agreed on a list of 22 specific items related to IC which were subsequently also confirmed by the administrators (Deardorff, 2006).This list of items was used to develop the Pyramid Model of Intercultural Competence (Deardorff, 2006) depicted in Figure 1.Deardorff (2006, p. 254) The Pyramid Model (Figure 1) can be entered at various levels with lower level components enhancing the upper level components.In other words, attitudes such as respect, openness, curiosity, and discovery form the foundation motivating individuals to acquire relevant knowledge, including culture-specific information, but also cultural selfawareness and a deep understanding of culture and its influence on human perception, cognition, and behavior.Besides knowledge, the individual might be motivated to develop interaction skills such as the ability to listen, to observe, and to interpret.All of these in turn support achieving desired internal outcomes, such as flexibility, empathy, and the ability to shift the frame of reference to be able to take the perspective of the interaction partner.Eventually, these outcomes promote the external outcome of appropriate and effective communication and behavior (Deardorff, 2006, p. 255).Taking these components and their interrelationships as a point of departure, Deardorff (2006) paid special attention to potential processes and movements between the various parts of the model, emphasizing that IC development is a continuous process in which 'one may never achieve ultimate intercultural competence' (Deardorff, 2006, p. 257).Therefore, Deardorff underlined that IC development is a lifelong learning process with no ultimate finishing line.
Since one aim of this study is to explore how students in a German public university context define IC, we have also reviewed the German scholarly discourse on IC.While Deardorff's Pyramid Model of Intercultural Competence is widely supported, three additional aspects surrounding IC were identified: First, IC might be conceptualized not as one larger competence in its own right, but rather as individual, social, professional, and strategic competence in intercultural situations (Bolten, 2007).Second, some German scholars shift emphasis from elements of IC to how IC manifests itself in interactions between people from different cultures, as captured in the German term 'Interkulturen' which pertains to dynamic processes of creating a third culture from synergy potentials of different interaction partners (Erll & Gymnich, 2015;Schumann, 2012;Thomas, 2011).Finally, discourse evolves around defining underlying core concepts of IC and in particular the notion of culture, shifting from the more traditional 'container model' of clearly distinguishable national or regional cultural groups to more fluid and dynamic notions of culture assuming that various subcultures exist in any human collective, thus allowing for different layers of cultures and identities (Rathje, 2006).Yet, all of these aspects relate to the scholarly discourse and thus might have little to no influence on how students, who are not familiar with these writings, define IC.
Building upon Deardorff's (2006) Pyramid Model, we seek to assess the extent to which students' subjective understanding of IC reflects the various dimensions of attitudes, knowledge, skills, internal and external outcomes.We also aim to explore to which extent the components of IC noted by students depend on their immediate context as well as factors such as prior intercultural experience.To address the latter, we build upon previous study by Odag et al. (2015) which examined a culturally diverse sample of predominantly international, undergraduate students at a private university in Germany.Though students' definitions of IC were largely in line with what experts in Deardorff's study (2006) agreed upon, Odag et al. (2015) found differences in how frequently elements of the different dimensions were mentioned and expanded their coding scheme (based on the Pyramid Model) by including additional, inductively derived categories.While international students most frequently referred to external outcomes (including communication and interaction), attitudes (such as tolerance), and knowledge, they mentioned intrapersonal outcomes and skills only infrequently (Odag et al., 2015).In addition, some elements in students' definitions were not captured in Deardorff's (2006) Pyramid Model, so that Odag et al. (2015) added categories such as intercultural harmony, integration, and collaboration/cooperation to the external outcomes dimensions.These inductive categories appear to be fairly specific to the students' living situation on a highly diverse campus.They might, however, also have been associated with around 38% of the sample belonging to collectivistic cultures, which tend to place more emphasis on harmony.This emphasis on harmony in collectivistic cultures has also been found by Holmes (2008) who conducted interviews with 14 Chinese international students and 10 domestic students in New Zealand and concluded that Chinese students put more emphasis on maintaining harmony in interpersonal interactions, via facework, silence, and being other-oriented.
In another study, Krajewski (2011) found differences between how students and staff members of an Australian university conceptualized IC.However, this study could only offer limited information on students' definitions of IC because it merely assessed students' agreement with statements derived from staff members rather than asking them for their own opinions.Root and Ngampornchai (2013) looked at students' understanding of intercultural competence by analyzing reflective papers of 18 students after their return from a study abroad program.They found that students discussed changes in all dimensions, i.e. knowledge, attitudes, and skills and reported being more knowledgeable, patient, flexible, and open-minded upon their return.However, the authors point to an overemphasis on affective and cognitive changes and a lack of discussing more specific skills related to IC.
Building on this research, our study seeks to address the influence of context on students' understanding of IC more systematically by comparing two purposefully selected samples of students who live and study in the same city in Northern Germany (Bremen), but attend two distinct universities in terms of organizational structure (size, language of instruction) and cultural composition of students and staff.In order to distinguish between both universities, we refer to both samples either as 'the domestic students' or 'the international students'.The international student sample in the study by Odag et al. (2015) stems from a small, private, international university (Jacobs University Bremen) with English as the official language of instruction and approximately 1,300 students from over 100 countries, who live on a residential campus.In contrast, the domestic student sample in the current study is drawn from a large, public, state university (University of Bremen) with German as the official language of instruction and approximately 20,000, predominantly German students, who live in various areas of the city.

Methods
The current study is descriptive in nature, exploring how students in a particular context define IC and comparing their definitions to those obtained in a previous study in another context.

Sampling and Participants
Domestic student sample.In line with Odag et al. (2015), all participants were recruited from two second-semester, undergraduate study courses (via email) using a purposive sampling strategy.Following the written informed consent, pen-and-paper questionnaires were completed during class time by 86 undergraduate students enrolled in the Bachelor of Psychology program at the University of Bremen in July 2015.Nine participants (10%) did not define IC and were thus excluded from the study, resulting in a final sample of n=77.This sample represents 49% (n=156) of the second-semester undergraduate psychology students enrolled at the University of Bremen in 2015.We decided to target psychology students for several reasons.First, the population of psychology students at the University of Bremen is similar in size to the total population of students at Jacobs University.Second, in contrast to the heterogeneous sample of international students in different study majors at Jacobs University, we aimed to obtain a homogenous sample of students in terms of their academic interests and nationality (German).
International student sample.The international student sample in Odag et al. (2015) was recruited at Jacobs University Bremen.The study was conducted one semester after all students completed an intercultural training during orientation period and had studied and lived on the international campus.The qualitative definitions of IC were obtained using a self-developed pen-and-paper questionnaire from n=130 second-semester undergraduate students representing 47% of all second-semester students at Jacobs University Bremen in 2013 (n=277) (Odag et al., 2015).
The data were collected during the second semester of studies because another aim of the study was to evaluate the effectiveness of the intercultural training in terms of academic and social integration one semester after the training (Kedzior et al., 2015).

Data Collection
Domestic student sample.Data were collected by means of a self-devised pen-and-paper questionnaire comprising one open-ended question asking for students' own definition of IC (question 1) and ten additional demographic questions (see Supplementary Materials S1 for the full questionnaire).All eleven questions were adapted from the Englishlanguage questionnaire used in Odag et al. (2015), translated to German, back-translated to English (by the current authors), and administered in German.The German questionnaire was pre-tested with three naïve participants to ensure that they comprehended the questions.Once this was confirmed, the questionnaire was administered during class time in two undergraduate courses (methods and statistics in psychology) in July 2015 at the University of Bremen.
International student sample.In the study by Odag et al. (2015), the self-devised pen-and-paper questionnaire described above was integrated into the questionnaire of another study which aimed to quantitatively assess the effectiveness of an intercultural peer-training at Jacobs University (Kedzior et al., 2015).The original English questionnaire (in Odag et al., 2015) was developed based on a qualitative study assessing the effectiveness of the intercultural training at Jacobs University (Binder, Schreier, Kuhnen, & Kedzior, 2013).The questionnaire was pretested with three students to ensure clarity of the questions and administered during class time in four large undergraduate courses (across different majors) (Odag et al., 2015).

Data Analysis
Domestic student sample.Data from 77 participants in the domestic student sample were analyzed using qualitative content analysis according to guidelines by Schreier (2012).The coding frame developed in Odag et al. (2015) was used and adapted to the data from the domestic sample (see Supplementary Materials S2 for full coding frame, including category labels, definitions, examples and decision rules).As shown in Figure 2, the coding frame is divided into several dimensions based on the IC model by Deardorff (2006) in Figure 1, capturing the general distinction between attitudes, knowledge, skills, as well as internal and external outcomes.Based on work by Stier (2006), the skills dimension was separated into intrapersonal and interpersonal skills to distinguish cognitive and emotional from interactive skills.The dimensions were then filled with more specific subcategories which were largely derived deductively from the literature, while some emerged inductively from the data (Odag et al., 2015).To achieve a better fit between the data from the domestic student sample and the coding frame from the Odag et al. (2015) study, two subcategories were added to the external outcomes dimension ('non-discrimination', 'helping others integrate'), both of which emerged inductively from the domestic student data.

Figure 2. Final subcategories of the coding frame used in the current content analysis
After dividing each definition into segments to separate units of meaning, the coding frame was applied to all text segments by coder one (AL) and subsequently to a subset of approximately 25% of all text segments by coder two (NB) to establish inter-rater agreement.Each text segment was assigned to one code only to satisfy the criterion of mutual exclusiveness (Schreier, 2012).Across dimensions, a high score of 91% could be obtained with inter-rater agreement ranging from 87-100% for the different dimensions (see Supplementary Materials S3).Unclear cases were discussed by both coders until consensus was established.
International student sample.Due to the changes to the original coding frame from Odag et al. (2015) described above and to allow for comparison with the international student sample, we recoded the data from the international student sample in Odag et al. (2015), paying special attention to the two new categories ('non-discrimination', 'helping others integrate').Two text segments were re-coded and assigned to the subcategory 'helping others integrate'.

Results
The qualitative codes were summarized in terms of descriptive statistics (quantitative frequency of responses).Such frequencies allowed us to visualize the relative importance of each IC dimension in student definitions.Furthermore, we compared both samples based on their intercultural experience/exposure and IC dimensions using univariate chisquare tests.

Participants-Domestic Students
Demographic characteristics of the domestic student sample are shown in Table 1.The participants were mostly female and young (on average 24) undergraduate university students.The age range (18-45 years) suggests that the sample included some mature participants.The majority of students were enrolled in their second semester of undergraduate Bachelor of Psychology degree.
The current sample was homogeneous in terms of nationality and intercultural exposure.Most participants reported a German nationality, having attended local (not international) high schools with German as the language of instruction, and resided in Germany within the last two years.The majority had no study abroad experience and also did not attend any workshops or courses related to intercultural competence.

IC Definitions-Domestic Students
Domestic students defined IC with reference to all dimensions from the Pyramid Model, including Attitudes (64%), External Outcomes (53%), Knowledge (42%), Internal Outcomes (29%), Intrapersonal Skills (6%), and Interpersonal Skills (4%; Table 2).Observation (n=1; 33%) Note.Percentage scores within each dimension and among the six dimensions exceed 100% because multiple responses from the same participants were coded into more than one dimension and/or more than one subcategory within each dimension.
The six dimensions of IC included the following subcategories according to students in our sample (Table 2).Within the Attitudes dimension the majority of text segments referred to tolerance/acceptance and openness and to a lesser extent to curiosity/discovery and respect.The External Outcomes dimension included text segments that pertained mainly to effective/appropriate interaction, effective/appropriate communication and, to a lesser extent, collaboration/cooperation. The Knowledge dimension consisted of mainly intercultural awareness and culture-specific knowledge.Less than 30% of the sample mentioned Internal Outcomes (including empathy and general adaptability/adjustment), Intrapersonal Skills (culture detection and judgement inhibition), and Interpersonal Skills (interactive learning and observation) in their definitions of IC.
In addition, we assessed the complexity of IC definitions by counting how many different dimensions were included in each student definition.The majority of definitions referred to one or two dimensions only (38% and 43% respectively), while only 18% of all definitions captured four dimensions and only 1% referred to all five dimensions.

IC Definitions-Domestic vs. International Students
The second aim of this study was to compare IC definitions provided by domestic students (current study) and international students in the study by Odag et al. (2015).Both studies utilized the same methods of sampling, data collection and analysis except for the language of data collection (German in the current study and English in Odag et al., 2015).According to results shown in Table 3, the domestic students most frequently mentioned Attitudes followed by External Outcomes while the international students mentioned the same dimensions but in the opposite order (External Outcomes followed by Attitudes).Both samples referred to Knowledge, Internal Outcomes, Intrapersonal Skills, and Interpersonal Skills in the same descending order of frequency.
A closer inspection of text segments revealed some commonalities but also differences between both samples in terms of subcategories of the six IC dimensions.Both samples most frequently referred to the same subcategories of External Outcomes (interaction and communication), Internal Outcomes (adaptability, empathy), and Interpersonal Skills (interactive learning, observation).Both samples only partially referred to the same subcategories of Attitudes (tolerance/acceptance), Knowledge (intercultural awareness), and Intrapersonal Skills (culture detection).The other subcategories of Attitudes, Knowledge, and Intrapersonal Skills differed between both samples.In general, the international students' definitions reflected elements crucial to their everyday life at an international university with frequent intercultural interactions, including respect (Attitudes), understanding others' worldviews (Knowledge), and .07Note. a The sample of international students from the study by Odag et al. (2015).b Pearson's chi-square with continuity correction for 2×2 comparisons.Cramer's V is a measure of an effect size on a scale from 0 to 1 and the same interpretation as that used for Pearson correlation coefficients (<0.3 is a small effect, <0.5 medium effect, >0.5 large effect).

Discussion
According to the current study, domestic students defined IC mostly in terms of Attitudes (mainly tolerance/acceptance and openness) and External Outcomes (mainly interaction and communication), followed by Knowledge, and to a much lesser extent Internal Outcomes and Intra-or Interpersonal Skills.These student definitions confirm all of the dimensions and most of the elements of IC in Deardorff's (2006) Pyramid Model.However, the comparison of responses from domestic students in this study vs. international students in Odag, Wallin, and Kedzior (2015), apart from a number of similarities also revealed differences in conceptualizations of IC, pointing to the assumed contextspecificity of IC.

Domestic Student Definitions and the Pyramid Model of Intercultural Competence
IC definitions of the German domestic undergraduate students who took part in the current study are consistent with Deardorff's (2006) Pyramid Model in that they refer to affective, cognitive, and behavioral dimensions as well as to internal and external outcomes.Likewise, the distinction between intrapersonal and interpersonal skills based on work by Stier (2006) is also supported based on data from the domestic students in the current study.Thus, it can be concluded that students in our domestic sample -as well as the international students in the study by Odag et al. (2015) -described IC using many of the aspects that scholars as well as higher education administrators in the study by Deardorff (2006) agreed upon.From this perspective, our data support the general applicability of Deardorff's (2006) Pyramid Model of Intercultural Competence to students in higher education, providing further evidence that this model is useful for designing IC development measures for university students.
However, student definitions in the current study -as well as in the study by Odag et al. (2015) -did not include all dimensions equally.More than 80% of all definitions in the current study referred to one or two dimensions only and mostly focused on attitudes and external outcomes, followed by knowledge and internal outcomes, while the skill component was largely neglected.Only 4-6% of all text segments pertained to skills in the domestic sample.This result is in line with previous qualitative studies with international students that showed an overemphasis on affective and cognitive aspects as well as outcomes while the behavioral aspect was seldom mentioned (Odag et al., 2015;Root & Ngampornchai, 2013).These findings can be interpreted in different ways.On the one hand, one is tempted to assume that students do not consider behavioral skills to be crucial for IC.On the other hand, these results might indicate a limited understanding of what it takes to transfer attitudes (such as tolerance and openness) and knowledge into concrete action in intercultural situations.This point has been discussed by Schaetti, Ramsey and Watanabe (2009) who argue that IC comprises culture-specific knowledge, culture-general knowledge, and an intercultural practice which refers to the moment-to-moment choice to apply the knowledge to a specific situation.Establishing such a practice requires 'a commitment to self-awareness, self-reflection, and self-monitoring' (Schaetti et al., 2009, p. 129) in a continuous process of learning.In light of research documenting students' tendency to have only limited interaction with students from other cultural backgrounds (Campbell, 2012;Halualani, 2008;Harrison & Peacock, 2009;Summers & Volet, 2008), most of them might not perceive the need to establish an intercultural practice nor feel a motivation to do so.In conclusion, the differences in how frequently student definitions include the various dimensions of IC suggest that some elements of Deardorff's (2006) Pyramid Model might be more salient to them while others tend to be neglected.Thus, the Pyramid Model can serve as a framework for universities when designing measures to promote IC among students, but they should also assess which elements of it are more or less salient to students, and how to support students in developing a more complex understanding of IC.

Comparing IC Definitions of Domestic and International Students
The comparison of IC definitions provided by international (Odag et al., 2015) and domestic students demonstrates that domestic students most frequently refer to attitudes, whereas international students focus most on external outcomes.In line with our assumption that students' subjective understanding of IC might be context-dependent, we argue that this difference could be explained by contextual factors.The international sample in the study by Odag et al. (2015) attended an international, private university and lived and studied on a small campus with other students from over 100 different countries.They were thus bound to socially interact and study with people from other cultural backgrounds on daily bases.In this context, frequent references to external outcomes such as effective/appropriate interaction and communication are hardly surprising.In contrast, our domestic student sample consisted mainly of domestic (German) students belonging to the majority culture of the university, city, and even country, and had presumably limited contact with people from other cultures in daily life, especially as data were collected in July 2015 prior to the high influx of refugees to Germany.This might explain why these students conceptualized IC mainly in terms of attitudes such as tolerance/acceptance and openness and less in terms of special interactional skills.In their theoretical understanding of IC, being open to people from different cultures and tolerating or accepting cultural differences appeared to be the most relevant and salient aspects of IC -which resembles the idea behind the German term Willkommenskultur that has mainly to do with accepting and being tolerant to refugees.This assumption is much in line with the idea that attitudes constitute the fundamental basis upon which knowledge and skills can be built, in turn facilitating desired internal and external outcomes as put forward in Deardorff's (2006) Pyramid Model.
When interpreting the differences in IC definitions between the international and the domestic students, one needs to bear in mind that the two samples differed from each other on a number of contextual dimensions.One such contextual factor was the country of residence in the last two years and previous study-abroad experience which indicated significantly higher exposure to other cultures in the international than the domestic samples.Based on their practical experience with foreign cultures, the international students in the study by Odag et al. (2015) might have placed more emphasis on external outcomes as well as intrapersonal skills (such as their ability to adapt or adjust to the new environment and to show empathy) because they have experienced the need for such skills first-hand.Further support for this explanation comes from Covert (2011) who obtained rich qualitative data from photographs, diary entries, and interviews with seven U.S. undergraduate students who spent a semester at different universities in Chile.Her analysis showed that students' IC development was a learning process, with students reaching higher levels of IC as they accumulated more intercultural experience and had more social contact in the host country.In addition, their understanding of IC evolved around language skills, communication styles, and interpersonal communication and behavior, and thereby focused on the skills and internal as well as external outcomes dimensions of Deardorff's (2006) Pyramid Model.One could argue that these aspects become more salient for students who experience a foreign culture from within.In contrast, students who study at home and have limited intercultural contact refer predominantly to the foundational level of the pyramid (attitudes) when defining IC.
A second contextual factor that distinguished the two samples from each other was the fact that only 14% of the domestic students had previously participated in intercultural workshops or courses, whereas all the international students in Odag et al. (2015) took part in a mandatory intercultural workshop during their orientation week.This might constitute yet another factor influencing the salience of different dimensions and aspects of IC among the two samples of students.
A third contextual difference between the two comparison groups was study major, constituting yet another potential factor of influence on understanding IC.While Odag et al. (2015) explored definitions of a diverse sample of students from different majors, the current study included psychology students only.One could argue that due to their study field, psychology students are particularly attuned to values and attitudes and therefore make most reference to the Attitudes dimension.Taken together, the current study cannot draw confident conclusions about which of these contextual factors (country of residence, study abroad experience, IC training, study major) alone or in combination with others influenced the differences in IC definitions.Individual influences of such contextual factors could be examined in more detail using quantitative or mixed-methods designs.Future studies could investigate if demographic and contextual factors can predict the preferences for different IC dimensions using regression analysis (for example, is respect chosen predominantly based on gender, age, and to a lesser extent intercultural experience).Alternatively, it would be interesting to multivariately explore clusters of demographic and contextual factors together with IC dimensions using cluster analysis (for example, would younger female students with little intercultural experience be more likely to define IC in terms of respect than older male students with extensive intercultural experience).Such quantitative analyses could reveal what predictors are most important for specific IC dimensions and could allow to classify students as more or less interculturally competent depending on the context.
Further evidence for the influence of students' context on salience of different aspects of IC is that we had to create two new categories-non-discrimination and helping others integrate-the latter of which mainly comprised references to the integration of migrants or refugees.The refugee situation in Europe and Germany has become a hot topic in the media and politics shortly after data collection but might have already influenced the context in which some domestic students thought about IC in the current study.By contrast, as this topic had not been extensively discussed in the media at the time of data collection in Odag et al. (2015), these two categories had not emerged in the IC definitions in the international student sample.International students might have been more concerned with aspects of IC that helped them come to terms with the practicalities of living and studying in a multicultural environment (Odag et al., 2015).Besides mentioning effective/appropriate communication, interaction, and respect as the most important aspects of IC, they also placed more emphasis on how other people view the world (understanding other's worldviews).Domestic students, by contrast, focused more on culture-specific knowledge when defining IC -i.e.factors that might help them when interacting with specific cultural groups, such as immigrants or refugees.Overall, domestic students seemed to consider IC as a theoretical construct.This aspect highlights once more that the emphasis students put on different aspects of IC might depend on their immediate context and daily reality, thus supporting the assumption by Spitzberg and Changnon (2009) as well as Moosmuller and Schonhuth (2009) that the concept of IC is context-bound.
In sum, our analysis of IC definitions by domestic students in the current study as well as the comparison with international students in the study by Odag et al. (2015) supports the general structure of the Pyramid Model and the majority of its specific dimensions (Deardorff, 2006).Yet, we also demonstrated that the salience and relative importance of the different dimensions and aspects of IC depend, to a large extent, on students' unique contexts of living and learning, supporting our assumption that understanding of IC is fluid and subject to change across contexts.We are convinced that this context-specificity has to be taken into account when conceptualizing IC in the higher education context.If universities seek to promote their students' IC development by curricular and extracurricular measures (such as courses or training on intercultural topics, buddy-programs, mentoring), they need to be aware of the specific context and day-to-day reality of their students to tailor those measures to their needs and address those aspects of IC relevant to them.

Limitations
There were a number of limitations in the current study.First, our findings have been derived from a purposive sample.Therefore, it is unclear how representative they are of the opinions of other students in higher education.Our study offers a preliminary insight into how domestic students in a particular context define IC, providing a basis for larger and representative investigations.Second, collecting data by means of short written answers to an open question in a survey certainly limits the depth in which students can express their understanding of IC.The use of more in-depth methods, such as semi-structured interviews or focus groups would allow to explore the topic further.Third, the current study is descriptive in nature.We utilized a qualitative method of data collection and summarized the data quantitatively (in terms of simple frequency of responses and univariate chi-square tests).Future studies might build on our findings to derive hypotheses about the influence of demographic and contextual factors on IC and test these using mixed-methods or quantitative approaches.Fourth, it is likely that other latent contextual factors exist that might have affected the opinions of the students.Such factors may be related to the current global events and the use of social media.Therefore, further studies should carefully examine how other contextual factors affect IC in students.Fifth, our classification of both samples as either culturally homogeneous with little intercultural experience (the domestic students) or heterogeneous with extensive intercultural experience (the international students) is rather simplistic.For consistency reasons we referred to the sample from the public university as 'the domestic students' (although 5% reported other than German nationality) and the sample from the international university as 'the international students' (although 42% reported German nationality).This classification reflects the different contexts of both samples rather than nationality with 'domestic students' describing the students at a public, German-speaking university with a predominantly German student body and 'international students' describing students at a private, English-speaking university with an international student body.Therefore, it cannot be assumed that all students in the domestic sample had limited contacts with other cultures.For example, their parents or close friends might have come from other cultures and they might have lived elsewhere than Germany earlier than in the last two years.Similarly, the international student sample was diverse in terms of individualistic and collectivistic cultures (including 38% of students from collectivistic cultures) and multinational students who might have never lived in their passport culture.Therefore, other studies should obtain more demographic information from the samples to better assess and understand student biographies and their influence on IC definitions.Furthermore, a systematic comparison of IC definitions from students with collectivistic vs. individualistic cultural backgrounds remains to be conducted.Finally, our study does not explicitly define the concept of culture.Instead, we focus on how students understand and define the concept of IC.Future studies may attempt to also seek the definition of culture since the understanding of this concept may also affect how students define IC.

Implications
While acknowledging the above-mentioned limitations, this study makes two important contributions.First, it provides further empirical evidence for the multi-dimensional construct of IC as derived from the study by Deardorff (2006).Similar to the students in the study by Odag et al. (2015), the students in the current study referred to the dimensions of attitudes, knowledge, and skills, as well as internal and external outcomes, mentioning most of the more specific elements included in Deardorff's (2006) Pyramid Model.The prevalence of references to attitudes in this study also supports the notion that attitudes constitute an important foundation of IC as suggested by Deardorff (2006).Second, we were able to demonstrate that context might influence IC definitions according to two samples of students from two different universities in the same city.This highlights the importance for higher education to take into account the realities of students (as well as staff) in their efforts to promote IC.IC development in higher education can only be successful if the day-to-day realities and contextual factors are taken into account in internationalization policies and strategy papers.Learning more about how students in various contexts understand IC could help increasingly multicultural universities to prepare the students to live and work in ever more diverse societies and workplaces.

Table 1 .
Demographic characteristics of 77 domestic (undergraduate) students

Table 2 .
Dimensions of intercultural competence according to 77 domestic students in the current study

Table 4 .
Dimensions of IC in domestic and international students depending on intercultural experience/exposure