Do country stereotypes influence language learning motivation ? A study among foreign language learners in Malaysia

It has been acknowledged in research literature that stereotypes about a target language country held by language learners influence the students’ motivation to learn a foreign language (L2 motivation). However, there is a lack of studies that link explorations of these stereotypes and L2 motivation in a single research project. This mixed-methods study addressed this gap in research literature. It examined relationships between endogenous stereotypes about target language countries and L2 motivation using a sample of 194 students learning foreign languages in a major public university in Malaysia. These languages included French, German, Italian, Portuguese (European and Brazilian varieties), Russian and Spanish. The data were collected through a questionnaire survey. One open-ended question sought the students’ mental images of the target language countries; two thermometer-type scales assessed the students’ general attitudes toward the target language countries and people and 16 closed-ended statements with attached Likert-type scales assessed their L2 motivation. The findings from the qualitative strand of the analysis revealed that the respondents had distinct and predominantly positive images of each of the target language countries. This allowed making a tentative proposition that the country stereotypes would have a positive relationship with the students’ L2 motivation, especially the integrative orientation. Results of the statistical analyses in the quantitative phase indicated that the relationship between the country stereotypes and L2 motivation was the strongest in the case of the integrative orientation. The article concludes with a brief discussion of implications that can be drawn from this study.


Introduction
There has been a renewed interest among educational researchers in stereotypes about a target language (TL) country, its culture and people that language learners bring into the classroom (Houghton 2010, Níkleva 2012, Drewelow 2013, Heinzmann 2014).It has been acknowledged that images and stereotypes about a TL country held by language learners serve as an important factor that motivates or demotivates the students to engage in the process of language learning (Gardner and Lambert 1972, Dlaska 2000, Castellotti and Moore 2002) and calls have been made to explore the relationships between these stereotypes and language learning motivation (Byon 2007).However, there is still a lack of such studies.This could be due to a methodological dissonance between predominantly qualitative studies on stereotypes about target language countries on the one hand and primarily quantitative assessments of motivation to learn an additional language (or L2 motivation) on the other hand (Nikitina 2015).
Stereotypes is a complex psychological construct that incorporates cognitive, attitudinal and evaluative components (Spencer-Rodgers 2001).Stereotypes are often perceived as erroneous and negative mental images.This in not necessarily so.They can be positive and accurate.The main feature of stereotypes is their rigidity and inflexibility.Researchers also recognize that stereotypes have a capacity to shape and guide an individual person's attitudes and actions toward stereotyped entities (Greenwald andBanaji 1995, Lippmann 2012).Based on these considerations, the present study defines stereotypes as "evaluative mental images and beliefs about a target language country held by language learners that have the potential to influence the learners' L2 motivation".In this article, the term 'stereotypes' is used interchangeably with the expressions 'country stereotypes' and 'mental representations'.
Recent empirical investigations have linked stereotypes with motivation.For example, Forbes and Schmader (2010) explored how stereotypes and attitudes about an academic field influence students' motivation to perform well in this particular field.Hypotheses about linkages between language learners' stereotypes about a TL country, its culture and people and their L2 motivation have been put forward in the field of foreign language education (Castellotti and Moore 2002, Dlaska 2000, Gardner and Lambert 1972) and calls have been made to empirically examine these relationships (Byon 2007).
Some investigations of L2 motivation did incorporate country stereotypes into their analysis.For example, Heinzmann (2014) included language learners' stereotypes about TL countries in her explorations of links between language attitudes and L2 motivation.However, the exploration of country stereotypes in her study involved only a fraction of the respondents and not the entire sample.In addition, the data on stereotypes had been gathered through closed-ended questions, which precluded the analysis of the language learners' endogenous country images.Nikitina (2015) used a mixed-methods approach to assess the relationships between country stereotypes held by learners of French, German and Spanish and the students' L2 motivation.She found that the students' stereotypes about the TL countries were for the most part positive.Also, the results of the Pearson's correlation test performed by the researcher indicated that there had existed a positive and statistically significant relationship between the students' country images and integrative orientation within the students' L2 motivation.However, the correlational analysis in Nikitina's study did not allow assessing whether the country stereotypes had influenced the students' L2 motivation, which necessitated further research on this issue.

Research on stereotypes and its methodology
Two main approaches to soliciting data on stereotype content, or what stereotypes about a study object consist of, are structured and unstructured approaches.The most popular among the former have been checklists, such as inventories of character traits.Participants in such studies are typically given a list of traits pertaining to a group of people and are asked to either endorse or reject the proposed traits.In contrast, unstructured approaches require that participants in a study list all words, traits or images which they feel would adequately describe a study object (Spencer-Rodgers 2001).
Studies on language learners' images of a TL country have predominantly relied on unstructured approaches (Taylor 1977, Berg 1990, Schulz and Haerle 1995, Abrams 2002, Allen 2004, Drewelow 2013).The qualitative methods are appropriate when a study is exploratory or when it aims to examine respondents' own endogenous representational structure about a study object (Spencer-Rodgers 2001).In the context of second language acquisition research, Storme and Derakhshani (2002:659) noted that employing unstructured data collection techniques would "undoubtedly elicit stereotypes".Importantly, they noted that stereotypes embed language learners' attitudes toward target language countries and cultures.As the researchers maintained, "…any activity requiring learners to report what they know about a target culture or to describe it will give us some insight into their attitudes towards it" (Storme and Derakhshani 2002:659).Hence, exploring country stereotypes could enable researchers to examine and measure language learners' attitudes.Some of the previous studies have made attempts to classify stereotypes about a TL country into positive and negative (Taylor 1977, Schulz andHaerle 1995).However, these classifications were rarely based on numerical data and they mostly relied on the researchers' own opinions.To depart from this approach, Nikitina and Furuoka (2013) adopted a method from psychology research where numerical ratings assigned by respondents to their images of a research object serve as proximal indicators of the respondents' attitudes (e.g., Spencer-Rodgers 2001).The researchers asked 116 learners of Mandarin to list their images of China and rate each image on a scale ranging from -3 (to denote a very negative image) to +3 (to mark a very positive image).However, Nikitina and Furuoka did not proceed to examine the relationships between the students' country stereotypes about China and their motivation to learn Mandarin.

Motivation in language learning: integrative and instrumental orientations
Pioneering research by Robert C. Gardner andWallace E. Lambert (1959, 1972) has empirically demonstrated an important role that motivation plays in learning an additional language (L2 motivation).The researchers introduced the influential concepts of "integrative orientation" and "instrumental orientation" within L2 motivation that relate to the main goals to be achieved in the end of the language learning process.To be more specific, the instrumental orientation involves "a desire to gain social recognition or economic advantage through knowledge of a foreign language", while the integrative orientation is characterized by a language learner's intent to emulate or to "become associated" with native speakers of the target language (Gardner and Lambert 1972:14).The proposition that the integrative orientation involves a desirability of being 'associated' with the target language speaking community and culture has attracted much criticism and some researchers expressed their doubts concerning the relevance of this concept in the context of learning a foreign language (see Nikitina, Zuraidah and Loh 2016 for a review of literature).This debate underscores the need for a broader definition of the integrative orientation.More recently, Gardner (1985Gardner ( , 2001Gardner ( , 2010) ) suggestedand the current study shares this view-that definitions of the integrative orientation should incorporate language learners' favourable interest toward a TL country, its culture, the target language itself and the community that speaks this language.Moreover, Gardner (1985) proposed that language learners' attitudes toward and beliefs about a TL country, its culture and people collectively known as "language attitudes" play an important role in the process of learning an additional language.The following subsection focuses on the construct of attitude and discusses methodological approaches adopted in studies on attitudes in L2 research.

Attitude and its measurements in psychology and L2 motivation research
The present study adopts Eagly and Chaiken's (1993:155) definition of an attitude, which approaches this construct in terms of "tendencies to evaluate an entity with some degree of favor or disfavor, ordinarily expressed in cognitive, affective, and behavioral responses".In L2 research, "language attitudes" encompass not only attitudes held by language learners toward a target language itself but also their attitudes toward-and beliefs about-speakers of this particular language and countries where this language is spoken (Dörnyei 1990, Kormos, Kiddle and Csizér 2011, Al-Hoorie 2016).Moreover, language attitudes are recognized as "emotional precursors" of a language learning behaviour (Kormos, Kiddle and Csizér 2011:3), which points to a link between these attitudes and L2 motivation.It should be noted that in the context of foreign language teaching and learning, students often do not have direct or sufficient contacts with target language countries, cultures and speakers and, therefore, their language attitudes tend to be rather vague (Dörnyei 1990, Kormos andCsizér 2007).This reality highlights the importance of employing appropriate methodologies to identify attitudes that are pertinent and endogenous to language learners.Spolsky (2000) suggested that researchers should strive to supplement structured approaches with unstructured data collecting techniques.In a similar way, Kormos and Csizér (2007) argued that questionnaire surveys that contain only closed-ended statements prepared by researchers are not sufficient for collecting and assessing language attitudes and cultural beliefs, especially in the contexts where direct contacts between the learners' own culture and a TL culture are lacking.Steele and Suozzo (1994 as cited in Storme and Derakhshani 2002) proposed to employ approaches and techniques developed in the field of psychology.
Recently, there have been signs of a methodological shift toward employing innovative data collection techniques in research on language attitudes and L2 motivation.For example, Al-Hoorie (2016) used a computerized reaction-time measure called the Implicit Association Test to gather language learners' implicit attitudes toward L2 speakers.Another way to gather language learners' endogenous attitudes is using a free-listing technique where respondents are asked to list all words, images or traits to describe the entity under study (Sutrop 2001).Storme and Derakhshani (2002) recommend to adopt this approach in studies that assess language learners' attitudes and cultural beliefs.
In addition to employing innovative methodologies, the variables that measure language attitudes have become more diverse in recent empirical investigations.For example, Al-Hoorie (2016) included the variable "implicit attitudes toward L2 speakers/community", Csizér and Dörnyei (2005) examined "vitality of the L2 community" and "cultural interest".Kormos, Kiddle and Csizér (2011) incorporated measures assessing various "language-related attitudes".These studies have provided empirical evidence that language attitudes do have a direct influence on L2 motivation.
To recap, on the one hand, quantitative studies that employed closed-ended questions or items to gather information on language attitudes yielded numerical data that were analysed by statistical methods.However, these closed-ended statements were selected by the researchers and were not based on the endogenous beliefs and attitudes held by the respondents.Therefore, these studies might have failed to capture the attitudes and beliefs that are most pertinent and salient to the language learners.On the other hand, qualitative studies that relied on open-ended questions have provided useful insights into the participants' authentic and endogenous language attitudes (Taylor 1977, Schulz and Haerle 1995, Allen 2004, Drewelow 2013).However, due to the absence of numerical data the estimations of language attitudes as either positive or negative were based on the researchers' assumptions and not on the learners' own evaluations.Clearly, some methodological advances are needed to resolve this paradoxical situation and bridge the methodological gap.The following subsection outlines a theoretical framework to support such an endeavour.

Theoretical linkages between stereotypes, attitudes and L2 motivation
Psychological functions are not separate entities that exist in isolation.As Lev Vygotsky observed, these functions have a complex systemic relationship with each other (see Kozulin 1990).Concerning the intertwined relationships among the constructs of attitude, cognition and motivation, Buck (2005 as cited in Dörnyei 2009:235) noted that "in their fully articulated forms, emotions imply cognitions imply motives imply emotions, and so on".
As discussed earlier in this article, stereotypes are recognized as an important cognitive tool that aids people to process new information (McGarty, Yzerbyt andSpears 2002, Lippmann 2012).In addition to this cognitive function stereotypes embed attitudes toward the stereotyped entity (Greenwald andBanaji 1995, Lippmann 2012).Importantly, these attitudes can influence people's behaviour and guide their actions (Greenwald andBanaji 1995, Forbes andSchmader 2010).The awareness that stereotypes contain attitudes and thus have a propensity to nudge people to behave in certain ways logically leads to a proposition that country stereotypes held by language learners-and attitudes embedded in these imageswould influence the students' motivated language learning behaviour.However, while numerous studies have explored links between language attitudes and L2 motivation, there is a lack of studies that incorporated country stereotypes into the investigations.This lacuna could be due to methodological constraints highlighted in the previous subsections.The following section proposes a method to overcome these limitations.

Aim and research questions
The aim of this mixed-methods study is to link a qualitative exploration of language learners' individually-held endogenous country stereotypes with a quantitative assessment of their L2 motivation.It addressed the following research questions: (1) What are the language learners' stereotypes about and attitudes toward the TL countries?
(2) What are the relationships between the students' country stereotypes and their L2 motivation?
(3) Which of the L2 motivation components is mostly affected by the country stereotypes?

Participants
Participants in this study were 194 (N=194) beginner learners of all European languages taught in University of Malaya at the time of data collection, and also beginner learners of Brazilian Portuguese.The target languages and the number of participants were as follows: French (n = 23), German (n = 26), Italian (n = 34), European Portuguese (n = 29), Brazilian Portuguese (n = 23), Russian (n = 28) and Spanish (n = 31).Learners of Asian languages were not included in this study due to its focus on the target language countries that are culturally and geographically distant from Malaysia.
The majority of the participants (n=183) were Malaysians; international students who took part in this survey (n=11) hailed from Bangladesh, Egypt, Kyrgyzstan, Saudi Arabia and Spain.The participants majored in various disciplines (e.g., the social sciences, law, linguistics, sport education, engineering and medicine) and took the language courses as an elective subject.Among the participants, 142 (73.2%) were women and 52 (26.8%) men, which reflected the gender ratio of students studying in the country's public institutions of higher learning.The age of respondents ranged from 20 to 26 years (M = 21.87;SD = 1.08).

Research design and research instrument
This study adopts the concurrent or parallel mixed-methods design where the mixing of the qualitative and quantitative strands is done synchronously (Teddlie and Tashakkori 2009).This research design can be expressed as QUAL + QUAN.The "+" sign indicates that the qualitative (QUAL) and quantitative (QUAN) strands of the study were carried out simultaneously; the upper case letters attest to the equal importance given to the QUAL and QUAN phases.
The research instrument contained one open-ended question, two thermometertype scales and 16 closed-ended statements (see Appendix for a sample for the questionnaire).The open-ended question sought to obtain the students' endogenous country stereotypes.In order to assess attitudes embedded in these stereotypes, the participants were instructed to give each image a mark from -2 (for a very negative image) to +2 (for a very positive image).Such marks are known as 'favourability ratings' in psychology research (Spencer-Rodgers 2001).To acknowledge the reality that language learners' attitudes toward a TL country are often vague due to the lack of cultural exposure (Dörnyei 1990, Kormos andCsizér 2007), the research instrument included two thermometer-type scales.Such scales enable measuring less well-defined attitudes held by participants (Spencer-Rodgers 2001).
The measures of L2 motivation in this study were based on the model developed by Gardner andLambert (1959, 1972) and they assessed: (1) General motivation or the effort that the students were willing to expend to learn a target language (Items # 2, 3,5,6,8,13); (2) Instrumental orientation or the language learners' perceptions of the TL utility and their intention to use the TL for pragmatic purposes (Items # 7, 9, 12, 14, 16); (3) Integrative orientation or the language learners' intention to learn the TL in order to communicate with the TL speakers and to understand their ways of life (Items # 1, 4, 10, 11, 15).
Prior to performing the statistical analysis, this study established internal consistency of the L2 motivation components.It was found to be good: Cronbach's alpha (α) for the General motivation scale was .70;Cronbach's α for the Instrumental orientation scale was .83,and Cronbach's α for the Integrative orientation scale was .75.

Data collection and analysis
Photocopied forms of the research instrument (see Appendix) were distributed to the respondents during their language classes.The students filled-in the forms in situ, which took between 15 and 20 minutes, and returned them to the researcher.Each questionnaire collected from the students was given an identification number.The qualitative data on the language learners' country stereotypes were typed ad verbatim together with the favourability ratings given to each image.
To answer research question 1, frequency and content analyses were performed.Content analysis can be carried out at several levels, such as words, themes, concepts, ideological statements, subject topic and so on (Berg 2001, Julien 2008).The unit of analysis in this study were stereotypical images of TL countries held by the students which were expressed in single words or short phrases.The frequency analysis assessed the number of times with which each image about a TL country had been mentioned.During the content analysis the country stereotypes were divided into categories.To establish reliability of the content analysis, the intrarater reliability analysis was performed (Loewen andPlonsky 2015, Mackey andGass 2015).This analysis, also known as the "intracoder reliability" check (Van den Hoonaard 2008:445), assesses consistency of the manner in which the same researcher codes the data at two different points in time.Following the methodologists' guidelines, first of all, the researcher coded all of the qualitative data on the country stereotypes.Few weeks later a portion of the data was re-coded by the same researcher without looking at the previous results.Then the intra-rater reliability rate was calculated as a simple percentage agreement between the coding done by the researcher at time one and time two.The achieved reliability rate was 96%, which is above the desirable percentage agreement of 80-90% recommended by Loewen and Plonsky (2015).
The qualitative analysis of the data was integrated with the quantitative assessment of language attitudes embedded in the country stereotypes through calculating mean valence (MV) value of each category of images.A mathematical formula for calculating MV values can be expressed as where MVj is the mean valence value of the category j; Vji is the valence ratings given by the students to the images j in this category; Fj is the number of images in the category.
In order to compare language attitudes held by the learners of various languages, composite mean valence (CompMV) of all images of a particular TL country was computed as follows: where Vji is the valence rating given to image j by student i; M is the total number of images of a particular TL country given by the students.These calculations were done using the Microsoft Excel program.
To answer research question 2, regression analysis was performed.First of all, multiple regression analysis of the data on all of the TL countries was done.The dependent variable was L2 motivation or, more exactly, its four components revealed by the exploratory factor analysis (EFA).The independent variables were the language learners' stereotypes about the TL country, their general attitudes toward the TL country and their general attitudes toward the L2 speakers.This step was followed by simple regression analysis of the data on each TL country.In the simple regression analysis, there were two variables.The dependent variable was L2 motivation while the independent variable was the country stereotypes about the TL country.The variables measuring general attitudes toward the TL country and the speaking community were excluded from the analysis.Before proceeding to the statistical tests, the underlying assumptions for the regression analysis were tested.Firstly, the diagnostic graphs were examined and, secondly, the residual normality test, the homoscedasticity test, the linearity test, the multicollinearity test and the autocorrelation test were performed.No serious violations of any of the underlying assumptions were detected.All tests and analyses were done with the aid of the SPSS 21 computer software.
To answer research question 3, a type of path analysis known as structural equation modeling (SEM) was performed.Regression analysis allows estimating relationships only between two research variables at a time, while SEM permits to simultaneously assess the interplay among all variables in a study.Due to this ability SEM is gaining popularity in studies on L2 motivation (Winke 2014).The SEM was done using AMOS 21 software.

Language learners' country stereotypes and attitudes
Findings from the qualitative analysis revealed that the students had a rich array of images of the TL countries, namely, France, Germany, Italy, Portugal, Brazil (where a variety of Portuguese is spoken), Russia and Spain.The most ubiquitous representations about each of these countries were as follows.Brazil was associated in the minds of students with "football", "beaches", "FIFA world cup", "travel destination" and "culture".The top five frequently mentioned images of France were "Eiffel Tower", "Paris", "romantic country", "love" and "fashion".The predominant images of Germany included "advanced technology", "football", "cars", "Volkswagen" and "engineering".Italy evoked the images "pizza", "spaghetti", "Venice", "pasta" and "food".As to Portugal, the most ubiquitous stereotypes were "Cristiano Ronaldo", "Portuguese dance", "football" and "Portuguese egg tart".Russia was associated with "big country", "cold country", "Matryoshka", "Russian alphabet" and "difficult language".Finally, the top five frequently mentioned images of Spain were "football", "bullfight", "wine", "movies" and "beautiful girls".
Next, the findings revealed that the students held predominantly positive attitudes toward all seven TL countries.Interestingly, it was found that all 12 categories of images of Spain had positive mean valence values that ranged from 2.000 ("Nice and beautiful country") to 0.500 ("Food and wine").As to the images of France, only one out of eleven categories had a negative mean valence value; this category was labelled "High cost of living" (MV = -1.000).Mean valence values of the rest categories ranged between 1.920 (for "Paris and Eiffel Tower") and 1.000 (for "Cold weather").In the case of Germany, out of 14 categories of images only the cluster labelled "History" had a negative mean valence value (MV = -1.200).The highest positive mean valence of 1.952 had the category "Cars and carmakers", while the cluster "Language" had the lowest positive mean valence value of 0.556.
Among 14 categories of images of Brazil, 15 categories of country stereotypes about Italy and 11 categories of representations of Portugal, only two clusters of images of each country had negative mean valence values.In the case of Russia, among 11 categories of images four were negative.To be more specific, mean valence values of the positive categories of images of Brazil ranged between 1.800 ("Beach and ocean") and 0.600 ("Country size and location").The two negative clusters referred to "Economy" (MV= -0.500) and "Crime and disarray" (MV= -1.667).Concerning Italy, the positive categories ranged from "Scenery and landscape" (MV=2.000) to "Fashion and luxury brands" (MV=1.225)while the negative clusters of images were "Economy" (MV=-1.000)and "Mafia" (MV=-2.000).As to Portugal, the category labelled "Weather" (MV=2.000)had the highest positive mean valence value while the cluster labelled "Language" (MV=1.000)had the lowest positive mean valence value.The two negative categories were labelled "Obscure country" (MV= -0.333) and "Former colonial power" (MV= -0.417).The analysis of images of Russia revealed that "Technology, aerospace industry and education" had the highest mean valence value (MV= 1.733); the lowest positive mean valence had the cluster "Russian language and alphabet" (MV = 0.476).The four negative categories contained the images referring to "Cold weather" (MV = -0.368),"Military and war" (MV= -0.800), "Soviet Union and communism" (MV= -1.000) and "Mafia" (MV = -1.500).
The finding that the language learners' stereotypes about each and every TL country were positive allowed putting forward a tentative proposition that the country stereotypes would have a positive relationship with the students' L2 motivation, especially the integrative orientation.Another tentative proposition was that the learners of Brazilian Portuguese, French and Italian could have a stronger integrative orientation compared to the learners of European Portuguese, German and Russian.This assumption was made on the basis of higher composite mean valence values of the images of these countries, which also might attest to stronger positive language attitudes held by the learners of these languages.This proposition was in line with the earlier findings concerning the relationships between language attitudes and L2 motivation (Gardner and Lambert 1959, 1972, Gardner, Masgoret and Tremblay 1999, Csizér and Dörnyei 2005, Humphreys and Spratt 2008, Csizér and Kormos 2009, Kormos, Kiddle and Csizér 2011).

Dimensions of L2 motivation
First of all, exploratory factor analysis (EFA) was carried out to establish construct validity and identify latent dimensions in the learners' L2 motivation.Suitability of the data for this statistical procedure was confirmed as recommended by Hair et al. (2006): the Kaiser-Meyer-Olkin measure of sampling adequacy was .85; the Bartlett's test of sphericity was significant (χ 2 (194) = 979.17,p < .01)and the antiimage correlations were above .7.
The EFA analysis was performed using the principal component model and the varimax rotation method with Kaiser normalization. 1 The analysis identified four dimensions in the students' L2 motivation: "Effort", "Perseverance", "Instrumental orientation" and "Integrative orientation".The newly extracted four factors explained approximately 61% of variance, which is considered satisfactory in the social sciences research (Hair et al. 2006).The subsequent statistical tests were based on this four-dimensional structure of L2 motivation.
Thus, the dependent variables in this study were "integrative orientation" (INTEG), "instrumental orientation" (INST), "effort" (EFFORT) and "perseverance" (PERSEV).The independent variables were "stereotypes" (IndMV) which derived from the ratings given by the students to their country stereotypes, "attitudes to TL country" (Att_count) and "attitudes to native speakers" (Att_peop) which derived from the students' marks on the two respective thermometer-type scales.

Relationship between country stereotypes, language attitudes and L2 motivation
The regression analysis explored the relationships between the students' country stereotypes and their L2 motivation (research question 2).Firstly, the multiple regression analysis was performed on the data for all seven TL countries.Table 1 presents the findings.To establish the nature and magnitude of the relationships between the independent and dependent variables, the standardized coefficients (β), the t-statistic and the coefficient of determination (R2 ) were compared.As the table shows, the Italian language learners' stereotypes about the TL country had a positive and statistically significant relationship with both the integrative orientation (β=.396, t-statistic=2.441,R 2 =.157; p<.05) and the instrumental orientation (β=.389, t-statistics=2.391,R 2 =.152; p<.05).This relationship was slightly stronger for the integrative orientation as reflected in the higher values of the relevant parameters.In the case of the Brazilian Portuguese learners, the country stereotypes had a positive and statistically significant relationship with the students' integrative orientation (β=.573, t-statistic=3.203,R 2 =.328; p<.01), however, no such relationship was detected for the instrumental orientation.
The country images held by the learners of Russian and Spanish had a positive and statistically significant relationship with the students' instrumental orientation; no such relationship was detected between the stereotypes and integrative orientation.The β statistic was higher for the data on the learners of Spanish (β=.428; t-statistic=2.549;R 2 =.183; p<.05) compared to the learners of Russian (β=.415; t-statistic=2.328;R 2 =.172; p<.05).In the cases of the students learning French (β= .382;t-statistic=1.895;R 2 =.146; p<.10) and European Portuguese (β=.322; t-statistic=1.767;R 2 =.104; p<.10), the country stereotypes had a positive and statistically significant relationship with the integrative orientation at the 10 percent level of statistical significance. 2

Findings of the path analysis
It was hypothesized earlier in this study that the language learners' country stereotypes would have a positive relationship with all aspects of the L2 motivation, particularly the integrative orientation.This tentative proposition was based on the findings of the earlier studies (Gardner and Lambert 1959, 1972, Gardner, Masgoret and Tremblay 1999, Csizér and Dörnyei 2005, Csizér and Kormos 2009, Kormos, Kiddle and Csizér 2011).This assumption was tested using a type of path analysis known as structural equation modeling (SEM).Before assessing the full structural model, as the SEM requires, a measurement model was built and evaluated as recommended by Hair et al. (2006).The findings of the confirmatory factor analysis (CFA) indicated the need for some modifications.Therefore, in order to fulfil the validity requirements five questionnaire items with lower loadings on their factors were removed.To achieve a better goodness-of-fit for the full model, the variables measuring general attitudes were removed.Figure 2 present the full SEM model.The respecified model shown in Figure 2 had good validity parameters (χ2 /df = 1.373; χ2 = .058,p > .05;GFI = .956;IFI = .976;CFI = .976;RMSEA = .044).The results of the SEM analysis indicated that the language learners' country stereotypes had a positive, considerable and statistically significant relationship with all four components of the students' L2 motivation.The stereotypes were positively and statistically significantly related to the effort (β=0.62,p<.01) and perseverance (β=.58, p<.01) aspects.Also, they had a positive and statistically significant relationship with the instrumental (β=.66, p-statistic <.01) and integrative (β=.98, p <.01) orientations.In a nutshell, the findings from the SEM analysis provided empirical support to the tentative proposition concerning the nature of the relationship between the country stereotypes and L2 motivation.

Discussion and conclusions
The need for using various measures and techniques to gather data on language attitudes and L2 motivation has been highlighted in many earlier studies (e.g., Spolsky 2000, Storme and Derakhshani 2002, Kormos and Csizér 2007) and, most recently, by Al-Hoorie (2016).The current study employed a concurrent mixedmethods approach to examine the relationship between endogenous country stereotypes held by students learning foreign languages in a large public university in Malaysia and these students' L2 motivation.The data in the qualitative strand of the study were gathered through one open-ended question and the numerical marks provided by the respondents.This approach allowed obtaining the language learners' endogenous images of and attitudes toward the TL countries; quantifying these data enabled the analysis of the relationship between the students' language attitudes and their L2 motivation.
The qualitative findings in the current study revealed that the students viewed the TL countries in a very positive light, which was reflected in the positive composite mean valences of the images of each country.This finding allowed making a tentative proposition that the respondents' integrative orientation would be prominent, which was supported by the findings from statistical analyses in the quantitative strand.
To be more specific, initially, the results of the regression analysis indicated that the country stereotypes had a statistically significant relationship with only two components of L2 motivation, namely, the instrumental orientation and the integrative orientation.However, the findings from a more sophisticated statistical procedure, the SEM, allowed concluding that the relationship between all four aspects of L2 motivation and country stereotypes held by the students was positive, considerable and statistically significant.This finding provided empirical support to a proposition advanced in earlier qualitative studies (e.g., Dlaska 2000, Castellotti and Moore 2002, Byon 2007) that language learners' stereotypes and cultural beliefs about a TL country would have a positive association with the students' L2 motivation.
This finding also aligns with the results reported in earlier quantitative empirical investigations.It should be noted, however, that the previous studies have rarely explicitly examined the relationship between language learners' endogenous country stereotypes and their L2 motivation.Nevertheless, some similarities in the results can be identified and discussed.Some researchers (e.g., Gardner, Masgoret and Tremblay 1999, Csizér and Dörnyei 2005, Csizér and Kormos 2009, Kormos, Kiddle and Csizér 2011) discovered that positive language attitudes were causally linked to L2 motivation, particularly to the integrative orientation.Al-Hoorie (2016:423) reported that implicit attitudes toward L2 speakers held by L2 learners in his study were associated with the "strength of correlations among attitudinal and motivational variables".In a similar way, in the current study the language learners' country stereotypes-and the attitudes embedded in them-had a positive and statistically significant relationship with the students' L2 motivation.Importantly, this relationship was the strongest in the case of the integrative orientation (β=.98), followed by the instrumental orientation (β=.66), effort (β=.62) and perseverance (β=.58).These results support the findings of an earlier and smaller scale study by Nikitina (2015).
This study has some limitations.Among them is that the data were collected only from beginner language learners who had chosen to learn a particular language as an elective course.Such learners tend to be more enthusiastic and they have higher levels of L2 motivation (Dlaska 2000).Including students at more advanced levels of the language programs might have yielded different findings concerning the country stereotypes and the patterns in L2 motivation.Also, the participants were mostly Malaysian students learning either a European language or Brazilian Portuguese, which means that considerable geographical and cultural distances existed between the respondents' own culture and the target language countries and cultures.These factors might have had an impact on the findings regarding the stereotype content.Future studies might want to explore country stereotypes and L2 motivation among Malaysian learners of Asian languages.
Despite these limitations, the findings reported in this study have some implications for the theory of L2 motivation and for pedagogical practice.From a theoretical perspective, this study confirmed validity and usefulness of the integrative orientation concept.The findings revealed that though the students' stereotypes about the TL countries had significant and positive relationships with each and every of the four aspects of L2 motivation, the link was the strongest in the case of the integrative orientation.As to the pedagogical implications, Gardner (2010) maintained that integratively-motivated learners are likely to be more successful in learning a new language and in achieving better linguistic and nonlinguistic outcomes of a language program.In view of the current study's findings, language educators might want to collect country images from each new group of language learners in the beginning of semester.This information could be valuable for developing the cultural component of the language program that suits the students' learning needs.
In addition, exploring country stereotypes held by language learners would enable foreign language educators to practice advanced pedagogies based on constructivist principles that link new learning experiences with the students' prior knowledge.A variety of enjoyable learning activities can be developed to expand the language learners' cultural knowledge and raise their awareness of the influence and power that stereotypes have in shaping our perceptions of other countries, nations and cultures.These activities could include humorous quizzes based on the students' erroneous images of a target language country (e.g., "True or False?-It always snows in Russia.").Furthermore, group-based projects may encourage more profound explorations of the target cultures (Abrams 2002, Allen 2004, Byon 2007, Nikitina 2016, 2017).Such projects might focus on history, art, high culture, important 'keywords' or cultural notions, all of which are the areas that tend to be lacking in the students' images of TL countries (Nikitina 2017).Importantly, such activities would help to nurture the language learners' positive interest in the target cultures, maintain their motivational levels and deepen their cultural knowledge.
Another promising but less explored pedagogical approach could be juxtaposing country stereotypes held by language learners with self-stereotypes predominant in the target culture (Níkleva 2012).Alternatively, language learners can be introduced to popular stereotypical images of their own country held by people in other countries and encouraged to critically review these stereotypes (Berg 1990).To conclude, recognizing that endogenous country stereotypes do have some influence on language learners' L2 motivation not only enhances the understanding of psychological processes involved in learning an additional language but also encourages foreign language educators to develop a culture-rich foreign language curriculum.Studying this language is important because it will enable me to better understand the way of life in Germany.
1 2 3 4 5 5 After I complete Level 1 of the German language I will continue learning this language at Level 2.

Figure 1 .
Figure 1.Full model of the relationship between the students' stereotypes and their L2 motivation 3 Note.All coefficients indicated in the Figure are standardized and statistically significant.
tick (√) your answer to each of the following items.The scale is: 1= Strongly disagree (SD); 2= Disagree (D); 3= Neither disagree nor agree (N); German language can be useful for my further studies, such as at the Master's or PhD level

Table 1 .
Summary of multiple regression analysis on stereotypes, attitudes and L2 motivation Notes.β is standardized regression coefficient; t is t-statistic; R 2 is the coefficient of determination; ** indicates statistical significance at the 1% level; * indicates statistical significance at the 5% level