Language Learning Motivation in China: Results of a Large-Scale Stratified Survey

Abstract

This article reports on the findings of a large-scale cross-sectional survey of the motivational disposition of English language learners in secondary schools and universities in China. The total sample involved over 10,000 students and was stratified according to geographical region and teaching contexts, selecting participants both from urban and rural locations. The investigation aims to present a balanced overview of the general level of L2 motivation in China through the lens of the L2 Motivational Self System (Dörnyei 2005, 2009). The wealth of data obtained in the survey provide solid empirical description of the main features of language learning in China, and the detailed information presented in the study can also serve as a baseline for future research conducted to investigate temporal, social, and geographical variation and evolution.

INTRODUCTION

Over the past decade, China has increasingly joined the global community of economy and culture, and established a high-profile presence in various areas of the international scene. This has foregrounded the significance of knowing English in the country and, indeed, by 2005 approximately 176.7 million Chinese people were engaged in some kind of English learning ( Graddol 2006 ), which constitutes the largest English learner group in the world. In spite of this boom, there have been only fragmented reports of the motivational basis of language learning from this significant country, and most of the studies reported in the literature examined limited samples by means of instruments that usually did not represent the latest theoretical insights in the field. The aim of the current investigation is to obtain a systematic and comprehensive overview of the motivational characteristics of learning English in China. For this purpose, we surveyed a large sample ( N > 10,000) and we used stratified sampling, with the categories of the sampling frame including (i) two learning contexts: university and secondary school; (ii) three geographical regions: eastern, central, and western; (iii) within secondary schools two types: urban and rural; (iv) within universities two types: key and ordinary; and (v) within universities two strands: English and non-English majors. The wealth of data obtained allow us to offer a balanced and empirically based overview that describes general trends and interesting contrasts within the English learning motivation of Chinese students.

Important aspects of the language learning context in China

In the era of economic globalization, proficiency in English has been seen in China as a definite asset of considerable value both at an individual and a societal level. Accordingly, English is taught as a core subject of the national curriculum across the whole country from primary school until the second year in university. However, the uniform educational targets do not automatically result in an equally uniform learning environment: with a population of over a billion, China displays considerable regional differences. In the past 30 years, the eastern regions of the country have experienced faster economic growth, better developmental opportunities, and greater returns than the central and western regions ( Chen 1987 ), and the gap in regional development is still increasing ( Tu 1995 ). Economic disparities among regions, in turn, result in uneven allocations of education resources in these areas ( Chen and Wu 2011 ), with English language provision being no exception to this trend ( Wu 2001 ). Consequently and unsurprisingly, students in more developed parts of the country have been found to have better English proficiency than their peers elsewhere ( Yan and Horwitz 2008 ).

Within the regional differences, we find a further developmental contrast: the urban/rural disparity. Due to the scarcity in natural and social resources, priority has typically been given to urban areas over rural regions, and thus the gap between urban and rural areas in China has been widening ( Yu 2006 ). This has had inevitable consequences for English language education because urban areas have normally been allocated better teaching staff, teaching equipment, and curricular resources. As a recognition of the unevenness in provision, a large number of students in rural areas are sent by their parents to secondary schools in cities to receive better education in spite of the often rather high costs involved.

In addition to the regional and the urban/rural disparities, we find a third major source of educational discrepancy within the Chinese educational system, related to the existence of two tiers in higher education: key universities , which were promoted by the government initiative of ‘Project 211’ and ‘Project 985’ ( MOE 2009 ), and ordinary universities ( Chen and Wu 2011 ). Inevitably, students’ English proficiency levels vary not only across geographical regions but across the two different types of universities.

A brief overview of L2 motivation research in China

L2 motivation research in China began in the 1980s under the influence of Gardner and Lambert’s (1972) social psychological perspective (e.g. Gui 1985 ; Zhang 1986 ). In the 1990s, a more critical approach emerged with several conceptual papers analysing L2 learners’ individual differences and defining the nature and role of motivation within this domain (cf. e.g. Liu 2002 ; Wang and Zhang 2005 ). Subsequent empirical research examined the main types of L2 motivation and the motivational characteristics of Chinese learners of English (e.g. Wu et al. 1993 ; Hua 1998 ; Gao et al. 2003a , b ; Liu 2007 ); the correlation between L2 motivation and other learner factors such as self-identity (e.g. Gao and Zhou 2008 ; Gao et al. 2011 ; Liu and Gao 2012 ); the relationship between L2 motivation, coursebooks, and classroom activities (e.g. Gao 2004 ; Li and Wu 2007 ) including the impact of motivational strategy training (e.g. You 2010 ); as well as the existence of unique Chinese models of L2 motivation (e.g. Zhou 1992 ; Qin and Wen 2002 ; see also Chen et al. 2005 ).

By means of illustration, Gao et al. (2003b) have conducted one of the largest-scale investigations of this kind, involving 2,278 undergraduates from 30 Chinese universities in 29 provinces, autonomous regions, and municipalities. They classified Chinese learners’ L2 motivation into seven types: (i) intrinsic interest, (ii) immediate achievement, (iii) learning situation, (iv) going abroad, (v) social responsibility, (vi) individual development, and (vii) information medium. It was then shown that the students’ motivational intensity varied according to motivation type ( Gao et al. 2003a ), and subsequent longitudinal studies ( Zhou et al. 2011 ) examined the changes in motivational types and intensity of university students during their studies. This research programme stood out from the more general trend of empirical studies in China that tended to follow some version of Gardner and his associates’ classic approach (see Gardner 1985 ), typically involving small sample sizes in which secondary schools were underrepresented, and only few studies focusing on the latest theoretical insights in the field.

L2 motivational self system

One of the main objectives of this study was to conduct a theory-driven survey of the motivational characteristics that underlie the learning of English in China. For the purpose of directing the selection of motivational aspects to be included in our instrument, we adopted the L2 Motivational Self System ( Dörnyei 2005 , 2009 ), partly because it has generally been accepted as an integrative synthesis of several previous constructs and approaches in L2 motivation research, and partly because over the past decade it has been successfully utilized in quantitative surveys in diverse learning environments such as Germany ( Busse 2013 ), Hungary ( Kormos and Csizér 2008 ; Csizér and Lukács 2010 ), Indonesia ( Lamb 2012 ), Japan, China, and Iran ( Taguchi et al. 2009 ), Pakistan ( Islam et al. 2013 ), Saudi Arabia ( Al-Shehri 2009 ), and Sweden ( Henry 2009 , 2010a ).

Derived from possible selves theory in social psychology (e.g. Markus and Nurius 1986 ; Higgins1987 ), the L2 Motivational Self System comprises three principal components: the Ideal L2 Self , the Ought-to L2 Self , and the L2 Learning Experience . The Ideal L2 Self concerns a desirable self image of the kind of L2 user one would ideally like to become in the future. If learners see a discrepancy between this and their actual self image, the unease that this difference generates will act as a potent motivational source. The Ought-to L2 Self reflects an ‘imported’ self image, that is, the attributes that one believes one ought to possess to meet the expectations of others and to avoid possible negative outcomes. The L2 Learning Experience represents situated, ‘executive’ motives related to the immediate learning environment/experience such as the impact of the teacher, the curriculum, the peer group, or the experience of success. Within this model, two dimensions of the traditionally conceived instrumentality were proposed ( Dörnyei 2005 , 2009 ): promotional instrumentality and preventional instrumentality , with the former referring to instrumental motives with a ‘pulling power’ (e.g. learning English to facilitate professional achievement) and the latter subsuming instrumental motives with an avoidance focus (e.g. studying in order not to fail an English course or disappointing one’s parents). Making this distinction is important because the first type is related to the Ideal L2 Self, while the second to the Ought-to L2 Self.

The L2 Motivational Self System has already been used successfully for assessing the L2 motivation of Chinese learners of English: Taguchi et al. (2009) conducted a study involving a comparative analysis of motivation in Japan, China, and Iran, and their structural equation modelling approach supported the validity of the three-component model of the construct in these learning environments. In a follow-up mixed methods study, Magid (2009) highlighted the significance of certain family aspects (most notably, responsibilities and pressures) and the notion of ‘face’ as significant factors in understanding language learning motivation in a specifically Chinese context. In a recent study, Li (2014) examined the differences in the motivation of Chinese learners of English in a foreign language context (China) and in a second language context (New Zealand). The results confirmed the validity of the L2 Motivational Self System in both environments, but also revealed interesting contextual variations (e.g. stronger idealized self images in New Zealand and higher levels of preventional instrumentality in China).

Research paradigm

The current survey was aimed at obtaining a comprehensive and representative overview of English language learning motivation in China. In order to achieve this purpose, we placed special emphasis on the systematic nature of sampling. As will be seen below, our sampling frame involved a methodical permutation of the main contextual parameters discussed earlier (e.g. geographical region, educational strand, and university type), to which we added two further categories: English major vs. non-English major and gender. The outcome of the design process was a stratified survey scheme that includes an elaborate system of strata, with 36 different clusters at the lowest level (e.g. university student–East–key university–English major–female; or secondary student–West–rural–male). We believe that this framework has taken into consideration most of the important independent variables that have been found in past studies to impact the motivation to learn English in China, and thus, it covers sufficient diversity to allow for the emergence of a comprehensive picture.

METHOD

Participants

Our survey investigated two different L2 learner populations—secondary and university students—evenly distributed across the three main regions of China: eastern, central, and western. Altogether 10,569 students were examined, and after eliminating some inadequate data (see below) the final sample size was 10,413. Table 1 presents a detailed summary of the subsamples according to the strata of the sampling frame. All the secondary students were selected from the first-, second-, and final-year populations (Years 11–13 in the UK system), with an average age of 16.5 years. Some came from urban schools located in educational centres in major cities (e.g. Yinchuan, the capital city of Ningxia Hui Autonomous Region in the west) others from rural schools situated in small towns (e.g. Huling Town in the east). The university student sample was evenly distributed across two university types—key and ordinary—and was made up of freshmen and sophomores, with an average age of 19.6 years. In each higher education institution we selected separate groups of English majors and non-English majors, the latter studying varied subjects. Looking at the whole sample in Table 1 , we can see that all the 36 lowest-level subgroups are represented with appropriate sample sizes—with a mean of 281 participants per subgroup—and only five cells have fewer than 100 participants. Because of the limited resources available, we could not conduct random or systematic sampling within each strata of the sampling frame, but it is believed that the robust coverage ensures that no major motivational trends have gone unnoticed.

Table 1:

Detailed demographical summary of the investigated sample (figures indicate the number of students assessed in each category)

Educational strand		Region		Type of institution/course		Sex a
Secondary schools	4,508	East	1,456	Urban	857	Male	352
						Female	481
				Rural	599	Male	283
						Female	305
		Centre	1,484	Urban	751	Male	381
						Female	352
				Rural	733	Male	372
						Female	333
		West	1,568	Urban	778	Male	296
						Female	436
				Rural	790	Male	363
						Female	413
Universities	5,905	East	1,928	Key universities	1,067
				English majors	401	Male	59
						Female	330
				Non-English majors	666	Male	430
						Female	219
				Ordinary universities	861
				English majors	377	Male	30
						Female	338
				Non-English majors	484	Male	224
						Female	243
		Centre	2,059	Key universities	963
				English majors	503	Male	132
						Female	354
				Non-English majors	460	Male	262
						Female	183
				Ordinary universities	1,096
				English majors	496	Male	29
						Female	458
				Non-English majors	600	Male	265
						Female	331
		West	1,918	Key universities	983
				English majors	490	Male	74
						Female	396
				Non-English majors	493	Male	253
						Female	234
				Ordinary universities	935
				English major	390	Male	53
						Female	333
				Non-English majors	545	Male	377
						Female	155

^a Some questionnaires had missing gender data.

Instrument

The research instrument used was a questionnaire specifically developed for the study following the principles in Dörnyei (2010) . It contained 73 six-point Likert scale items and seven background questions. The motivation domain targeted by the questions included three domains, (i) aspects of the L2 Motivational Self System, (ii) Intended Effort, and (iii) Language Learning Vision, but because of space limitations, this article only reports on the first two areas. In order to measure Intended Effort and the principal components of the L2 Motivational Self System, we drew on a previous questionnaire used by Taguchi et al. ( 2009 ; see also Dörnyei 2010 ) and we also included several newly designed items based on initial interviews with 10 Chinese L2 learners. A summary of the variables discussed in this article can be found in Table 2 .

Table 2:

Information about the motivational variables

Variables	Number of items	Sample items
Ideal L2 self	5	I can imagine myself in the future having a discussion with foreign friends in English.
Instrumentality-promotion	6	Studying English can be important to me because I think I’ll need it for further studies.
Cultural interest	5	I like English films.
Travelling	5	I study English because with English I can enjoy travelling abroad.
Ought-to L2 self	6	I consider learning English important because the people I respect think that I should do it.
Instrumentality-prevention	5	I will study English harder when thinking of not becoming a successful user of English in the future.
Parental expectations	5	I have to study English, because, otherwise, I think my parents will be disappointed with me.
Language learning experience (Attitudes to L2 learning)	5	I find learning English really interesting.
Intended effort	5	Even if I failed in my English learning, I would still learn English very hard.

The English version of the questionnaire was translated into Chinese both by the first author and a professional English–Chinese translator. The Chinese version was then back-translated by a professional Chinese–English translator, resulting in the modification of several items. As the next step, three Chinese university students were asked to think aloud when completing the questionnaire, and based on their comments, further minor amendments were made to the Chinese wording of certain items. Prior to the main administration, the Chinese version of the questionnaire was piloted among 208 Chinese L2 learners (109 university students and 99 secondary students) who were in the same age range as the participants in the main study. The questionnaire administrators were asked to take note of any problems raised by the respondents but no issues were reported. Based on the item analysis of this pilot run, some items were removed or rewritten before the questionnaire was finalized.

Data collection and analysis

Data collection took place between October 2012 and July 2013. Participants were recruited by an extensive search, using the researchers’ contact as well as a snowball sampling procedure. Once a potential location was identified, the first author approached the chosen institution by email or phone and provided information about the purpose and details of the survey. After obtaining permission, we contacted the vast majority (approximately 90%) of the teachers of the selected classes individually, asking for their co-operation (while the remaining 10% were contacted by colleagues and friends). Printed copies of the questionnaire were mailed or personally delivered to each institution, along with an administration manual for participating teachers. This manual provided exact guidelines on what information to give to the students prior to the survey as well as instructions on administering the questionnaire and clarifying uncertainties. Designated contact persons in each institution circulated the questionnaires among the participating teachers and collected the completed forms. Students filled in the questionnaires during class time; answering the questionnaire took 15 min on average. On balance, the procedures went smoothly, with only a few minor issues reported. The completed questionnaires were mailed back to the research team and specially commissioned assistants keyed in the raw data using Microsoft Excel. Finally, the data file was converted to SPSS 21.0.

Before we started the analyses, we ran several checks to spot any outliers and errors (see Dörnyei 2007 for discussion of how to conduct such probes), and also performed visual inspections of the hard copies of the questionnaires to find indications of meaningless questionnaire completion (e.g. making the marks on the sheets into a visual pattern such as an ‘X’). Large-scale surveys will always involve a small number of respondents who do not take the process seriously (see e.g. Dörnyei et al. 2006 ); in our sample we eliminated 156 questionnaires from the sample for various reasons, which was a low proportion (less than 1.5%) and was therefore considered acceptable. We then conducted a reliability analysis of the remaining data, and as shown in Table 3 , the Cronbach Alpha internal consistency reliability coefficients for all the nine multi-item scales rendered satisfactory levels in most clusters, and even the few subgroups where the coefficients did not reach the recommended .70 threshold, the differences were only a few percentage points. Thus, we may conclude that the instrument has delivered data of good psychometric quality, and the consistency of the coefficients in the various subgroups attests to the appropriate quality of the survey administration procedures across the learning contexts.

Table 3:

Reliability coefficients for the motivational variables (East/Centre/West)

Variables	Whole sample	Secondary students	University students
			English majors	Non-English majors
Ideal L2 self	.88 (.89/.88/.87)	.88 (.90/.89/.86)	.86 (.85/.86/.86)	.88 (.88/.87/.87)
Instrumentality-promotion	.73 (.77/.72/.70)	.75 (.79/.74/.71)	.71 (.73/.70/.69)	.70 (.70/.69/.69)
Cultural interest	.77 (.79/.76/.75)	.78 (.80/.78/.75)	.72 (.74/.71/.71)	.76 (.77/.77/.71)
Travelling	.82 (.85/.82/.80)	.83 (.86/.82/.80)	.79 (.80/.81/.76)	.83 (.85/.82/.81)
Ought-to L2 self	.74 (.76/.74/.73)	.74 (.77/.73/.73)	.76 (.76/.75/.76)	.73 (.75/.74/.70)
Instrumentality-prevention	.71 (.73/.69/.70)	.72 (.73/.69/.72)	.74 (.76/.73/.74)	.68 (.71/.66/.66)
Parental expectations	.73 (.75/.74/.72)	.71 (.74/.70/.68)	.75 (.75/.77/.73)	.75 (.75/.75/.76)
Attitudes to L2 learning	.88 (.89/.89/.88)	.90 (.91/.90/.89)	.86 (.87/.86/.85)	.87 (.86/.87/.87)
Intended effort	.81 (.81/.80/.80)	.82 (.83/.81/.80)	.80 (.79/.80/.80)	.78 (.77/.79/.79)

RESULTS AND DISCUSSION

Descriptive results

Table 4 presents descriptive statistics of the nine motivational variables as measured in the whole sample and in all the subsamples. The reason for offering such a comprehensive summary of the results is to make it possible for the dataset to serve as baseline data for comparisons in future investigations. Regarding the whole sample, the mean values range from 3.21 to 4.32 on a 6-point scale, and in seven of the nine scales they exceed the midpoint, 3.50, which indicates a generally favourable disposition towards learning English. This is also reflected in the fact that the highest scale value was obtained for Intended Effort ( M = 4.32), indicating solid commitment, and the robustness of this positive disposition is further evidenced by the fact that of the 52 subgroup coefficients presented in Table 4 for Intended Effort, only 8 are below 4.2, and of these only 1 is (slightly) below 4. We can therefore conclude that Chinese students in general are favourably disposed towards studying English as a foreign language.

Table 4:

Descriptive information about the motivational variables

Variables		Whole sample	University students			Secondary students (Urban/Rural)
			Whole sample (Key/Ordinary)	English majors (Key/Ordinary)	Non-English majors (Key/Ordinary)
Ideal L2 self		3.84	3.97 (4.04/3.88)	4.23 (4.30/4.15)	3.75 (3.82/3.68)	3.67 (3.79/3.54)
	East	3.83	4.02 (4.10/3.92)	4.21 (4.29/4.12)	3.90 (3.99/3.77)	3.58 (3.84/3.22)
	Centre	3.88	3.98 (4.00/3.97)	4.21 (4.24/4.18)	3.77 (3.75/3.79)	3.74 (3.79/3.68)
	West	3.80	3.89 (4.02/3.75)	4.27 (4.36/4.15)	3.57 (3.67/3.47)	3.68 (3.74/3.63)
Instrumental-promotion		3.89	3.98 (4.06/3.91)	4.04 (4.10/3.98)	3.94 (4.02/3.86)	3.76 (3.86/3.65)
	East	3.91	4.09 (4.18/3.98)	4.05 (4.09/4.00)	4.12 (4.24/3.96)	3.66 (3.92/3.29)
	Centre	3.93	4.02 (4.07/3.98)	4.08 (4.15/4.01)	3.96 (3.98/3.94)	3.81 (3.87/3.75)
	West	3.82	3.84 (3.90/3.77)	3.98 (4.05/3.90)	3.71 (3.76/3.67)	3.80 (3.77/3.83)
Cultural interest		4.06	4.17 (4.26/4.07)	4.40 (4.46/4.33)	3.98 (4.08/3.87)	3.91 (4.13/3.66)
	East	4.14	4.27 (4.34/4.19)	4.41 (4.44/4.39)	4.18 (4.28/4.04)	3.96 (4.27/3.52)
	Centre	4.09	4.18 (4.27/4.10)	4.39 (4.45/4.33)	3.98 (4.08/3.91)	3.97 (4.10/3.84)
	West	3.94	4.05 (4.16/3.93)	4.40 (4.50/4.28)	3.75 (3.83/3.68)	3.80 (4.00/3.61)
Travelling		3.88	3.87 (3.90/3.85)	4.08 (4.08/4.07)	3.71 (3.75/3.67)	3.89 (4.04/3.71)
	East	3.96	3.96 (3.95/3.97)	4.10 (4.07/4.12)	3.87 (3.88/3.86)	3.95 (4.22/3.56)
	Centre	3.90	3.89 (3.92/3.87)	4.05 (4.06/4.05)	3.74 (3.76/3.72)	3.91 (3.95/3.86)
	West	3.79	3.77 (3.83/3.70)	4.09 (4.11/4.06)	3.50 (3.56/3.44)	3.81 (3.93/3.69)
Ought-to L2 self		3.49	3.52 (3.51/3.52)	3.52 (3.54/3.51)	3.51 (3.49/3.54)	3.45 (3.47/3.42)
	East	3.47	3.57 (3.56/3.58)	3.54 (3.55/3.52)	3.59 (3.56/3.62)	3.34 (3.47/3.16)
	Centre	3.51	3.55 (3.56/3.53)	3.58 (3.62/3.53)	3.52 (3.50/3.53)	3.47 (3.45/3.48)
	West	3.48	3.44 (3.42/3.46)	3.45 (3.45/3.46)	3.43 (3.39/3.46)	3.53 (3.50/3.56)
Instrumental-prevention		4.14	4.15 (4.15/4.15)	4.18 (4.18/4.19)	4.13 (4.13/4.12)	4.13 (4.12/4.13)
	East	4.07	4.16 (4.15/4.16)	4.15 (4.13/4.17)	4.16 (4.16/4.16)	3.96 (4.05/3.82)
	Centre	4.18	4.19 (4.10/4.27)	4.18 (4.09/4.27)	4.20 (4.10/4.27)	4.18 (4.21/4.14)
	West	4.17	4.11 (4.21/4.01)	4.22 (4.30/4.12)	4.03 (4.13/3.93)	4.24 (4.11/4.36)
Parental expectations		3.21	3.22 (3.19/3.26)	3.34 (3.29/3.38)	3.13 (3.10/3.16)	3.20 (3.21/3.18)
	East	3.20	3.27 (3.18/3.37)	3.32 (3.26/3.38)	3.23 (3.14/3.37)	3.11 (3.15/3.05)
	Centre	3.23	3.21 (3.21/3.21)	3.35 (3.28/3.42)	3.08 (3.14/3.04)	3.25 (3.19/3.32)
	West	3.20	3.18 (3.16/3.20)	3.33 (3.32/3.34)	3.05 (3.00/3.10)	3.23 (3.31/3.15)
Attitudes to L2 learning		3.72	3.80 (3.83/3.77)	4.03 (4.04/4.03)	3.61 (3.66/3.57)	3.62 (3.67/3.56)
	East	3.70	3.85 (3.86/3.84)	4.01 (3.99/4.02)	3.74 (3.78/3.69)	3.50 (3.75/3.13)
	Centre	3.74	3.83 (3.85/3.82)	4.08 (4.08/4.09)	3.60 (3.59/3.60)	3.60 (3.68/3.52)
	West	3.74	3.72 (3.80/3.64)	4.00 (4.03/3.97)	3.48 (3.56/3.41)	3.75 (3.57/3.94)
Intended effort		4.32	4.40 (4.42/4.37)	4.60 (4.61/4.59)	4.23 (4.26/4.20)	4.21 (4.27/4.15)
	East	4.25	4.38 (4.40/4.35)	4.55 (4.55/4.55)	4.27 (4.32/4.20)	4.07 (4.28/3.77)
	Centre	4.33	4.43(4.35/4.49)	4.61(4.51/4.70)	4.26(4.16/4.33)	4.21(4.34/4.07)
	West	4.37	4.38(4.51/4.25)	4.63(4.74/4.50)	4.16(4.27/4.07)	4.35(4.18/4.52)

A closer look at the figures in Table 4 reveals the consistent trend that the Ought-to L2 Self domain produced some of the lowest scores. Within this domain, while the Instrumental-Prevention scores (i.e. the desire to avoid failure) are reasonably high, the Ought-to L2 Self scale’s overall mean (3.49) slips into the negative domain and the level of Parental Expectations is the lowest of all scales. These results go against the common stereotype that the typical disposition of Chinese learners is less individualistic and more societally determined. Indeed, the Ideal L2 Self scores are considerably higher across the board than the Ought-to L2 Self ones. In the next sections we shall return to this issue in the light of more nuanced analyses, and we shall also examine other prominent contrasts showing in Table 4 in focused comparisons.

Comparisons across genders

Table 5 presents the scores of the nine motivational dimensions broken down by gender. The figures show a consistent pattern, with female L2 learners scoring significantly higher than their male counterparts. In a review of gender differences in L2 motivation, Henry (2010b) points out that although empirical studies from different sociocultural contexts have, with few exceptions, revealed systematic gender differences in L2 motivation, little focused analysis has been directed at finding an explanation for the widely observed disparity. He proposed that the core difference between male and female learners may lie in their self-construal, with women’s possible selves being characterized by more interdependence and interpersonal qualities than those of men. In a follow-up study, Henry and Cliffordson (2013) further argued that females’ greater concern with interpersonal interaction and investing in self–other relationships makes it easier for them to envision themselves in future L2 communication situations, which in turn allows for the development of more elaborate and phenomenologically more robust motivational future self guides.

Table 5:

The motivational findings broken down by gender (M = male, F = female) ^a

Variables		Whole sample	University students						Secondary students
			Whole		English majors		Non-English majors
	M	F	M	F	M	F	M	F	M	F
Ideal L2 self	3.61	4.00	3.75	4.09	4.24	4.22	3.65	3.89	3.45	3.86
Instrumental-promotion	3.76	3.98	3.91	4.03	4.00	4.04	3.89	4.00	3.60	3.90
Cultural interest	3.82	4.23	3.94	4.31	4.30	4.42	3.87	4.12	3.68	4.11
Travelling	3.59	4.08	3.62	4.03	3.90	4.10	3.56	3.92	3.57	4.17
Ought-to L2 self	3.45	3.51	3.51	3.52	3.57	3.51	3.49	3.54	3.39	3.50
Instrumental-prevention	4.03	4.23	4.06	4.21	4.03	4.21	4.07	4.21	3.99	4.26
Parental expectations	3.19	3.22	3.16	3.25	3.34	3.33	3.13	3.12	3.22	3.18
Attitudes to L2 learning	3.45	3.93	3.57	3.94	3.96	4.05	3.49	3.77	3.31	3.90
Intended effort	4.06	4.50	4.16	4.54	4.37	4.64	4.11	4.39	3.96	4.43

^a Values in bold italics indicate non-significant differences.

There are two interesting exceptions to the robust gender-disparity trend in our data. First, we find several non-significant results in the two scales that directly reflect societal expectations (Ought-to L2 Self and Parental Expectations). This makes sense because these expectations are external to the learners and therefore are less determined by their gender. Secondly, we find fewer differences amongst university English majors than amongst the other students. This indicates that when it comes to real dedication—that is, a commitment that someone expresses by choosing English as their main university degree subject—the high level of motivation overrides any gender differences although the latter do not completely disappear.

Comparisons across geographical regions

Regional differences in the motivational scores are presented in Table 6 . Looking at the figures, the first general observation we can make is that in terms of its motivational set-up, China appears to be geographically bipolar : the scores obtained from the central part never represent an extreme but are always either in between East and West or have the same level as one of these. Thus, our data suggest a distinct East–West disparity with an important caveat: although in our survey the diversity was manifested along geographical dimensions, this motivational variation is rooted in the economic stratification of China as mediated through uneven resource allocations and through social class factors. The role of social class in applied linguistics has recently been highlighted by David Block (2014) , and a study by Kormos and Kiddle (2013) has provided empirical evidence that class issues combined with socio-economic factors have a medium-size effect on motivational variables.

Table 6:

Comparison of the motivational scores for the three geographical regions (E = East; C = Central; W = West) for the whole sample as well as broken down by school/course type

Variables		M : E/C/W	SD : E/C/W	F	Sequence a	Effect size b
Ideal L2 self	Whole	3.83/3.88/3.80	1.12/1.08/1.09	5.40**	W,E<E,C	.001
	Sec. school	3.58/3.74/3.68	1.21/1.15/1.09	6.90**	E<W,C	.003
	Eng. majors	4.21/4.21/4.27	.92/.95/.97	1.25	–	–
	Non-Eng. majors	3.90/3.77/3.57	1.03/1.03/1.08	27.49***	W<C<E	.017
Instrumental- promotion	Whole	3.91/3.93/3.82	.89/.82/.85	16.87***	W<E,C	.003
	Sec. school	3.66/3.81/3.80	.97/.87/.89	12.96***	E<W,C	.006
	Eng. majors	4.05/4.08/3.98	.79/.77/.78	4.13 *	W,E<E,C	.003
	Non-Eng. majors	4.12/3.96/3.71	.77/.77/.84	75.11***	W<C<E	.044
Cultural interest	Whole	4.14/4.09/3.94	.93/.91/.94	44.74***	W<C<E	.009
	Sec. school	3.96/3.97/3.80	1.01/.99/.97	14.24***	W<E,C	.006
	Eng. majors	4.41/4.39/4.40	.77/.76/.77	.22	–	–
	Non-Eng. majors	4.18/3.98/3.75	.87/.87/.89	65.16***	W<C<E	.039
Travelling	Whole	3.96/3.90/3.79	1.09/1.00/1.02	23.57***	W<C<E	.005
	Sec. school	3.95/3.91/3.81	1.19/1.06/1.04	6.45**	W<C,E	.003
	Eng. majors	4.10/4.05/4.09	.91/.93/.89	.56	–	–
	Non-Eng. majors	3.87/3.74/3.50	1.05/.98/1.01	37.05***	W<C<E	.022
Ought-to L2 self	Whole	3.47/3.51/3.48	.87/.83/.86	2.37	–	–
	Sec. school	3.34/3.47/3.53	.92/.84/.88	17.20***	E<C,W	.008
	Eng. majors	3.54/3.58/3.45	.84/.83/.86	4.99**	W<E,C	.004
	Non-Eng. majors	3.59/3.52/3.43	.79/.81/.82	10.20***	W<C,E	.006
Instrumental- prevention	Whole	4.07/4.18/4.17	.90/.84/.88	16.58***	E<W,C	.003
	Sec. school	3.96/4.18/4.24	.96/.89/.90	38.45***	E<C,W	.020
	Eng. majors	4.15/4.18/4.22	.87/.83/.88	1.29	–	–
	Non-Eng. majors	4.16/4.20/4.03	.83/.77/.83	12.80***	W<E,C	.008
Parental expectations	Whole	3.20/3.23/3.20	.93/.92/.93	1.16	–	–
	Sec. school	3.11/3.25/3.23	.96/.91/.92	10.29***	E<W,C	.005
	Eng. majors	3.32/3.35/3.33	.90/.95/.93	.24	–	–
	Non-Eng. majors	3.23/3.08/3.05	.90/.88/.92	12.80***	W,C<E	.008
Attitudes to L2 learning	Whole	3.70/3.74/3.74	1.05/1.03/1.10	1.56	–	–
	Sec. school	3.50/3.60/3.75	1.14/1.10/1.16	19.63***	E<C<W	.009
	Eng. majors	4.01/4.08/4.00	.91/.89/.94	2.18	–	–
	Non-Eng. majors	3.74/3.60/3.48	.95/.97/1.08	18.32***	W<C<E	.011
Intended effort	Whole	4.25/4.33/4.37	.87/.84/.90	17.17***	E<C,W	.003
	Sec. school	4.07/4.21/4.35	.96/.89/.92	35.73***	E<C<W	.020
	Eng. majors	4.55/4.61/4.63	.77/.78/.81	2.21	–	–
	Non-Eng. majors	4.27/4.26/4.16	.77/.77/.89	5.04**	W<C,E	.003

^a ‘<’ indicates significant difference; ‘,’ indicates non-significant difference.

^b Eta ² .

* p < .05; ** p < .01; *** p < .001.

The second statement we can make on the basis of the results is that the regional differences are not as strong as one would expect on the basis of the economic disparities discussed earlier. Although, as mentioned above, with a high sample size such as ours, most coefficients are likely to reach significance, the actual effect sizes are rather small, indicating that many of the observed differences are of relatively modest magnitude.

Within the geographical variation we do find, however, some meaningful contrasts. First, the data indicate that the geographically least affected subgroup is the university English majors: out of the nine variables they display non-significant results in seven, and the other two cases are also marginal differences. Therefore, similar to gender variation, we find here again that the commitment to pursue English studies overrides other influences. In contrast to this absence of variation, non-English majors, that is, the broader university population, followed a uniform trend in all the nine variables: the more developed East was associated with higher levels of English-related disposition than the western region. In fact, the three strongest East–West contrasts observed in our survey can be found within this subgroup: Instrumental-Promotion: 4.12 vs. 3.71; Cultural Interest: 4.18 vs. 3.75; and Travelling: 3.87 vs. 3.50. This pattern is fully consistent with the widely held view that the developed eastern parts of China exceed their western counterparts in terms of being integrated in the globalized international world whose lingua franca is Global English, which in turn fuels English learning motivation.

Comparisons across teaching contexts

The various subgroupings in the data set allow for a large number of potential analyses, but in our study there are four key contrasts regarding educational contexts that are at the heart of most of the variation: the motivational difference between students in (i) secondary schools and universities, (ii) ordinary and key universities, (iii) English and non-English degree courses, and (iv) rural and urban secondary schools. Table 7 presents t test statistics of the observed differences in these four key comparisons.

Table 7:

t-test statistics comparing the scores of various teaching contexts

Variables	Secondary vs. university students		Key vs. ordinary university students		English vs. non-English majors		Urban vs. rural secondary students
	t -value	Effect size a	t -value	Effect size a	t -value	Effect size	t -value	Effect size a
Ideal L2 self	−13.53***	.02	5.86***	.01	18.28***	.05	7.44***	.01
Instrumental-promotion	−13.23***	.02	7.07***	.01	4.83***	.004	7.68***	.01
Cultural interest	−13.91***	.02	8.34***	.01	19.54***	.06	16.13***	.05
Travelling	.59	–	2.22 *	.001	14.52***	.03	10.20***	.02
Ought-to L2 self	−4.10***	.002	−.51	–	.54	–	1.96 *	.001
Instrumental-prevention	−1.52	–	.01	–	2.49 *	.001	−.41	–
Parental expectations	−1.27	–	−2.96**	.001	8.72***	.01	1.27	–
Attitudes to L2 learning	−8.49***	.01	2.49 *	.001	16.87***	.05	3.15**	.002
Intended effort	−10.39***	.01	2.19 *	.001	17.51***	.05	3.95***	.003

^a Effect size: Eta ² .

* p < .05; ** p < .01; *** p < .001.

Regarding the Ideal L2 Self we see a very consistent pattern: the more advanced or specialized one’s education, the stronger ideal language image the student entertains. Thus, within secondary school students, pupils in urban schools exceed their rural counterparts, university students in general exceed secondary school pupils, and within university students key universities and English degree courses have the advantage over ordinary universities and non-English degree courses. All these contrasts are statistically significant, and the two ends of the ‘pecking order’—rural secondary school pupils at the low end and English majors in key universities at the high end—display a substantial discrepancy of 3.54 vs. 4.30, with the other permutations falling neatly in between these two extremes. Furthermore, we find exactly the same trend in four other variables: Instrumental-Promotion (3.65 vs. 4.10), Cultural Interest (3.66 vs. 4.46), Attitudes to L2 Learning (3.56 vs. 4.04), and Intended Effort (4.15 vs. 4.61). This consistent pattern suggests a link to the pattern previously observed regarding regional differences: the students’ education increasingly opens up their horizon onto the global world beyond China, bringing about an increasing appreciation of Global English.

Interestingly, the Ought-to L2 Self presents a strikingly different picture. Here the various comparisons are either non-significant or marginal (with effect sizes of 0.001 and 0.002) and the same nondescript trend is also true of the other two variables in the broad Ought-to domain, Instrumental-Prevention, and Parental Expectations. There is only one exception, but that is highly meaningful: the values for Parental Expectations in the university English majors’ subgroup are significantly higher than for non-English majors (3.34 vs. 3.13). This indicates that while students’ perceptions about their parents’ and peers’ expectation concerning English are fairly even and—as we have seen earlier—not very high across the country, when it comes to the students’ principal degree course—that is, to their future career—the parental expectancy level suddenly goes up. This is clearly related to a phenomenon typical of China, the principle of ‘reciprocal duty’ in society; as Magid (2009) explains, parents feel obliged to provide their children with the best possible education, and in return, it is the children’s responsibility to take care of their parents when they become old. Accordingly, when English is directly associated with their children’s future career, parents take an increasing interest in good results than when English is just a background subject.

Motivational impact on intended effort

The ultimate question about motivation in SLA is always how it can be translated into actual learning behaviours. In a large-scale survey such as ours, the closest we can get to examining this issue is by examining the association of motivational variables with a self-report behavioural measure, Intended Effort. Table 8 presents the motivation–effort correlations obtained for the three principal components of the L2 Motivational Self System. Looking at the table, the most obvious pattern is the consistent rank order of the three components of the L2 Motivational Self System across all the subgroups. The association with Intended Effort is highest with Attitudes to L2 Learning (.67 for the whole sample), followed by the Ideal L2 Self (.51), with the Ought-to L2 Self considerably lagging behind (.38). Of course, correlations cannot indicate cause–effect relations, but it is still important to note that for Chinese students the desire to invest time and energy in language learning seems to be associated first and foremost with the evaluation of the learning process. This result is consistent with findings in several previous studies (e.g. Kormos and Csizér 2008 ; Taguchi et al. 2009 ; Papi 2010 ; Lamb 2012 ; Islam et al. 2013 in Iran and Japan).

Table 8:

Correlations between the three principal motivational variables and intended effort ^a

Variables	Whole sample (male/female)	East (male/female)	Centre (male/female)	West (male/female)
Ideal L2 self	.51 (.51/.47)	.54 (.54/.50)	.49 (.48/.45)	.52 (.52/.48)
Secondary school	.49 (.50/.44)	.53 (.56/.45)	.45 (.41/.42)	.50 (.53/.44)
Urban	.48 (.45/.46)	.46 (.46/.45)	.39 (.34/.36)	.57 (.59/.54)
Rural	.51 (.53/.42)	.55 (.56/.41)	.50 (.48/.47)	.46 (.51/.36)
University	.52 (.51/.50)	.51 (.47/.51)	.52 (.57/.47)	.54 (.51/.51)
Key university	.50 (.47/.50)	.47 (.41/.49)	.55 (.59/.50)	.50 (.41/.51)
Ordinary university	.54 (.56/.50)	.56 (.53/.54)	.50 (.53/.46)	.56 (.58/.51)
Ought-to L2 self	.38 (.45/.33)	.41 (.44/.39)	.35 (.44/.27)	.37 (.45/.32)
Secondary school	.42 (.46/.39)	.46 (.49/.43)	.39 (.41/.32)	.40 (.45/.37)
Urban	.38 (.39/.36)	.41 (.39/.42)	.33 (.38/.24)	.39 (.44/.38)
Rural	.47 (.51/.41)	.49 (.53/.44)	.45 (.46/.41)	.41 (.46/.35)
University	.33 (.43/.28)	.33 (.34/.33)	.32 (.47/.25)	.35 (.45/.29)
Key university	.33 (.39/.28)	.30 (.30/.32)	.35 (.45/.26)	.34 (.43/.29)
Ordinary university	.34 (.47/.29)	.36 (.41/.34)	.30 (.50/.25)	.37 (.48/.29)
Attitudes to L2 learning	.67 (.67/.64)	.68 (.68/.65)	.67 (.67/.63)	.68 (.67/.65)
Secondary school	.70 (.69/.67)	.71 (.70/.67)	.69 (.66/.66)	.70 (.69/.67)
Urban	.67 (.65/.66)	.68 (.67/.67)	.64 (.61/.63)	.69 (.67/.67)
Rural	.72 (.70/.68)	.69 (.63/.64)	.72 (.69/.69)	.69 (.69/.63)
University	.64 (.64/.62)	.62 (.60/.62)	.64 (.66/.61)	.66 (.66/.64)
Key university	.61 (.62/.60)	.57 (.54/.58)	.66 (.68/.63)	.63 (.62/.61)
Ordinary university	.67 (.67/.64)	.68 (.70/.67)	.63 (.64/.60)	.69 (.68/.67)

^a All coefficients are significant.

Within the set of correlations between Attitudes to L2 Learning and Intended Effort, a further powerful trend appears: secondary school pupils display generally higher coefficients than university students (the correlation for the whole secondary school sample is .70 in contrast to the coefficient of .64 in the whole university sample, with the same trend reflected in the subgroups). This indicates that the quality of the learning experience plays a more decisive role for younger learners than for their university counterparts.

With regard to the two future self dimensions (ideal and ought-to), although the correlations with Ideal L2 Self display a fair amount of variation across the various subgroups, all the coefficients are significant without any obvious internal trends. This confirms the conclusions of Dörnyei and Chan (2013) and You and Chan (in press) that the ‘vision’ of one’s idealized persona is a valid and potent motivator for Chinese learners of English. Regarding the Ought-to L2 Self, the correlation coefficients are considerably lower than those for the Ideal L2 Self. This is in line with the earlier observation that the mean values for Ought-to L2 Self were the lowest amongst the three components of the L2 Motivational Self System, indicating only a moderate influence on Chinese learners exerted by societal and peer expectations. This limited role of the ought-to language self has been documented in several past studies, and based on these and their own results, Dörnyei and Chan ( 2013 , p. 454) concluded:

There is, thus, a tentative conclusion emerging from the existing body of research that, while externally sourced self-images (i.e. the images that are usually categorized under the rubric of the ought-to self) do play a role in shaping the learners’ motivational mindset, in many language contexts they lack the energizing force to make a difference in actual motivated learner behaviours by themselves.

Within the set of Ought-to L2 Self coefficients we find a salient pattern across the subsamples: the correlations for males are considerably higher than for females (of the 28 comparisons shown in Table 8 , a total of 26 display higher male scores, and the mean coefficients across all the subgroups are .44 vs. .33 for males and females, respectively). This consistent trend suggests that Chinese male learners are more affected by external societal expectations than their female counterparts, and therefore, the latter are more ‘responsible’ for the lower overall impact of the ought-to self dimension.

The ‘Chinese imperative’

Finally, let us examine a specific motive that has been labelled in the literature as the ‘Chinese Imperative’ by Chen et al. (2005) . Based on a survey of Taiwanese learners, the researchers identified a powerful factor that was rooted in the surrounding product-oriented institutional structure and ‘a society emphasizing and even praising exam results’ (p. 625). They claimed that this motive represented a uniquely Chinese factor that stemmed from the traditional Chinese striving for personal advancement and family welfare through success in exams. In this sense, this motive might involve more than the fear of educational failure or an aspiration for academic achievement: rather, it might be seen as a highly internalized aspect of the Chinese achievement-related mindset, equating value with exam success, further enforced by societal, educational, and familial expectations.

To examine this potential motivational aspect, Table 9 elaborates on the Ought-to L2 Self dimension already presented in Table 8 by examining the correlations of Intended Effort with Instrumentality-Prevention and Parental Expectations, two components that are central to the Chinese Imperative. Taken together the data in Tables 8 and 9 , we can conclude that the desire to avoid academic failure (assessed by Instrumental-Prevention) is indeed a powerful factor throughout all the subgroups. According to Magid (2009) , this is related to the Chinese concept of ‘losing face’ and can be indeed seen as East Asia–specific (although it is also present in other cultures, e.g. in the Middle East). However, our data also indicate that other components of the Chinese Imperative construct—parental and societal expectations—play only a small to moderate role in determining Intended Effort, and earlier we found the same twofold nature of the ought-to dimension in Table 4 , where only Instrumental-Prevention displayed high scale means, whereas the endorsement of the Ought-to L2 Self and Parental Expectations was the lowest of all the motivational variables assessed. Thus, our data from Mainland China confirm the significance of avoiding losing face as a powerful motivational factor, but there is no support for a socially embedded Chinese Imperative construct.

Table 9:

Correlations between instrumental-prevention and parental expectations with intended effort*

Variables	Whole sample (male/female)	East (male/female)	Centre (male/female)	West (male/female)
Instrumental-prevention	.50 (.52/.46)	.49 (.50/.47)	.45 (.49/.41)	.54 (.58/.48)
Secondary school	.55 (.58/.49)	.56 (.58/.50)	.47 (.50/.38)	.60 (.64/.53)
Urban	.50. (49/.48)	.48 (.46/.49)	.38 (.39/.35)	.60 (.62/.55)
Rural	.61. (65/.50)	.64 (.66/.54)	.54 (.58/.41)	.58 (.64/.46)
University	.45. (45/.44)	.40 (.35/.43)	.45 (.47/.44)	.49 (.51/.45)
Key university	.41. (39/.41)	.34 (.28/.39.)	.42 (.44/.39)	.46 (.47/.43)
Ordinary university	.49 (.51/.46)	.47 (.44/.46)	.46 (.51/.46)	.51 (.54/.46)
Parental expectations	.22 (.26/.19)	.25 (.26/.25)	.20 (.26/.14)	.22 (.27/.18)
Secondary school	.24 (.30/.22)	.30 (.34/.29)	.20 (.26/.13)	.22 (.27/.20)
Urban	.23 (.27/.21)	.26 (.28/.25)	.16 (.28/.04 a )	.26 (.31/.28)
Rural	.26 (.32/.24)	.35 (.42/.36)	.26 (.29/.23)	.22 (.29/.18)
University	.20 (.24/.17)	.18 (.15/.19)	.21 (.28/.15)	.22 (.27/.17)
Key university	.19 (.20/.16)	.15 (.12/.17)	.19 (.27/.10)	.22 (.20/.20)
Ordinary university	.22 (.29/.17)	.22 (.23/.21)	.22 (.29/.17)	.23 (.34/.13)

*All coefficients are significant unless otherwise marked.

^a Non-significant score.

CONCLUSION

Drawing on a large-scale stratified questionnaire survey, this article has reported on the most salient aspects of the motivational disposition of Chinese learners of English. Besides providing a comprehensive empirical description of the current motivational set-up, the detailed information presented in this article can also serve as a baseline for future research that focuses on social and geographical variation as well as temporal evolution within the Chinese context. In addition, it is hoped that the wealth of data from the world’s largest L2 learning community with considerable social, geographical, and educational implications can offer lessons that are relevant to second language acquisition in other parts of the world, thereby contributing to applied linguistics in general.

Despite the inevitable variation in the data as a function of the complex sampling frame, some broad and generalizable tendencies have emerged. To start with a perhaps unusual point, one of the most important findings of the survey was the fact that no extraordinary results have emerged. That is, the motivational set-up of the world’s largest L2-learning community has turned out to be broadly compatible with results obtained from other countries. This does not mean that certain motivational aspects do not carry unique weight in the Chinese context, but our overall experience is not consistent with the argument that the motivation models developed in Western cultural contexts do not sit easily within East Asian educational cultures. To the contrary, the L2 Motivational Self System appeared to offer a framework that seemed no less relevant and usable in China than in the other learning environments in which it has been successfully applied in the past. Related to this point, it is also noteworthy that we did not find a unique ‘Chinese Imperative’ factor that was conceptually different from Western notions of extrinsic motivation linked to the fear of academic failure: in our study, the only powerful aspect of this dimension was Instrumental-Prevention, with societal and peer expectations only playing a small to moderate part.

Turning to what we have found, the first point to highlight is the generally strong endorsement of motivation to learn English in China. Learners across the board tended to have positive ideal self images associated with English, equally positive Attitudes towards L2 Learning, and reported high levels of Intended Effort that they were ready to invest in the process. Significantly, our findings present a picture that is different from the widely held belief that Chinese learners are primarily instrumentally motivated. While instrumental factors (both promotion and prevention) do play an important part in motivating learning, they cannot be singled out as the principal parameters in any of the main Chinese learning subsamples assessed. Instead, language learning motivation has been determined by the joint operation of the established components of the L2 Motivational Self System; this is well reflected by the fact that Attitudes to L2 Learning explained the most variance in Intended Effort, followed by the Ideal L2 Self, with the Ought-to L2 Self only occupying the third place.

Regarding any internal variation within our total sample, we have observed several salient patterns. There was a definite gender division in the endorsement of the motivational scales, with female learners exceeding their male counterparts, but this distinction faded away in the scales belonging to the ought-to dimension; in fact, in terms of the correlations with Intended Effort, male learners did better than females on the Ought-to L2 Self scale. The gender difference was also partially overridden in the, by definition, most committed sample, university students who chose English as their main degree course.

We also found salient East–West variation. Our data point to the fact that the motivational set-up of Chinese learners of English is geographically bipolar, with the levels observed in the East typically being higher than those in the less developed West. Here again English majors are less affected, but non-English majors displayed a consistent East–West division in several key variables. We assumed that this pattern was related to the more globalized nature of the eastern regions, which fuels the learning of Global English. It is also likely that the same reason lies behind a great deal of the observed teaching-context-based variation. Here we encountered a very consistent pattern: the more advanced or specialized one’s education was, the stronger ideal language image the student entertained. A likely explanation for this trend is that the students’ education increasingly opens up their horizon onto the global world beyond China, bringing about an increasing appreciation of Global English. This may also explain the fact that for the younger age group—secondary school pupils—situated motives associated with the actual learning environment (as measured by Attitudes to L2 Learning) played a more decisive role in energizing L2 learning than for their university counterparts, explaining in some subgroups more than 50% of the variance in Intended Effort.

Interestingly, the Ought-to L2 Self dimension was by and large exempt from any geographical variation, pointing to a fairly even—and as our results showed—only moderate influence. However, Parental Expectations stood out in one specific subgroup: university English majors. We argued that this was related to the principle of ‘reciprocal duty’ that is prevalent in China, namely that parents feel obliged to provide their children with the best possible education, and in return, it is the children’s responsibility to take care of their parents when they become old. This would explain the special interest on the parents’ part when English was directly associated with their children’s main degree course, that is, their future career.

Conflict of interest statement . None declared.