Abstract
The second-language (L2) Motivational Self System (L2MSS) proposed by Zoltán Dörnyei described how L2 learners are motivated through envisioning their ideal future L2 self. The L2MSS framework for studying L2 motivation has had mixed results. The aim of this study was to evaluate measuring future selves as L2 motivators consistent with the L2MSS in Saudi female college students enrolled in a higher education English language program, and consider how these findings are applicable to collectivist cultures. A 20-item instrument was developed that showed evidence of validity and reliability in measuring ideal L2 future self and two ought-to ideal future selves as sources of L2 motivation. However, given the collectivist nature of Saudi culture, these measurements were positively correlated, giving rise to the question as to the utility of such a measurement. Future research on further validating the instrument and exploring ways to increase L2 motivation among learners in collectivist cultures is discussed.
Plain language summary
The second-language (L2) Motivational Self System (L2MSS) proposed by Dörnyei described how L2 learners are motivated through envisioning various future L2 selves, or images of themselves in the future using the L2 proficiently. Future selves described by Dörnyei include the ideal L2 self (IL2)—what the learner envisions for themselves in the future—and the ought-to L2 self (OL2)—what the learner envisions others’ expectations are for their future L2 self. The success of attempts to develop measurement instruments for level of motivation arising from IL2 and OL2 have been mixed, especially in collectivist cultures as defined by Hofstede. This study sought to develop a valid and reliable instrument to measure L2 motivation in undergraduates majoring in English at a woman’s university in Saudi Arabia using the L2MSS framework. After a pilot test of the instrument in a smaller sample to ensure it was performing adequately enough to be used in a larger research study, the revised instrument was administered to over 200 additional individuals, for a total sample of 254. Using a statistical approach called factor analysis, the instrument was found to be somewhat valid in measuring IL2 and OL2, but it was difficult to tell the measurements apart, because they appeared to be measuring almost the same thing. In addition, the PE subscale seemed to be also measuring the IL2. Results from interviews confirmed that learners would adjust their L2 motivation level as influenced by their parents and additional significant others, which may have been responsible for these similar measurements. As Saudi Arabia is a collectivist culture, while it may be possible to measure different future L2 selves as sources of L2 motivation in this group, the utility of the result may be questionable.
Keywords
Introduction
Early theories behind motivation for second language acquisition (SLA) were predominately developed by researchers Gardner and Lambert (1972) and Dörnyei (2009), and sought to explain motivation for second language (L2) learning in multiple contexts. As their theories became influential in L2 motivation research, it became clear that the concept of future L2 selves introduced by Dörnyei was assumed to translate to all L2 learning contexts, specifically with respect to Dörnyei’s idea of the ideal L2 self (IL2) (Dörnyei, 2009). He called his theory the L2 Motivational Self System (L2MSS) (Dörnyei, 2009), and it seeks to explain intrinsic motivation to learn an L2 that originates from the concept of future selves (Dörnyei, 2009; Moskovsky & Alrabai, 2009). Through much research, while it was found that while measurement of the IL2 appeared possible in different contexts, the measurement of other L2 selves conceptualized by Dörnyei seemed to be challenging in studies where participants were from collectivist cultures as characterized by Hofstede (2011), such as Iran and Japan (Taguchi et al., 2009).
The purpose of the present study was to measure the L2 motivation of female college students studying in an English program in the Kingdom of Saudi Arabia (KSA), an Arabic-speaking country. The aim of this study was to use the L2MSS framework for measuring L2 motivation to learn English in this sample, and consider how these findings are applicable to collectivist cultures. This section provides a short review of the evolution of Gardner’s and Dörnyei’s theories of L2 motivation, and describes the challenge of measuring future L2 selves in learners from different cultures. Briefly, Gardner’s theories led to research supporting practical teaching methods to decrease the time to L2 mastery for learners, while Dörnyei’s theories focused more on how to help learners maintain long-term motivation to master the L2.
Literature Review
Original Work by Gardner
The foundation for SLA motivation research was established largely by research led by Canadian social psychologist, Robert Gardner (Gardner, 1986; Gardner & Lambert, 1972). As Canada has both French-speaking and English-speaking regions, and Gardner’s team worked extensively with English-speaking students learning L2 French to adapt to the French-speaking region (Gardner & Lambert, 1972). Gardner’s group focused on studying many aspects of SLA, including ways to improve the L2 learning experience (L2LE) through an examination of learner motivation and attitude (Gardner, 1986). Gardner’s (1985) theory was expressed in the Attitude/Motivation Test Battery (AMTB), which he advocated for as a prescreen for learners to help instructors better understand their attitudes toward the L2 learning context. The AMTB measures “motivation intensity” as based on the learner’s level of “integrativeness” coupled with their attitude toward the learning situation, as Gardner described it (Gardner, 1985). Integrativeness was defined as attitude toward integrating into the new L2 community (Gardner, 1985).
Integrativeness as a Second-Language Motivation
In Gardner’s context, L2 learners were studying a language for the purpose of integrating into a specific L2 community to which they had social access (Gardner, 1985). However, this is not the setting for many L2 studies; a common setting is an “English as a second language” (ESL) course in any non-English-speaking country training learners who migrate to a future unknown English-speaking country (Teimouri, 2017). An ESL class in a non-English-speaking country lacks a specifically-defined L2 community into which the learners are planning to integrate, which is also true of L2 English learners in KSA, where the community speaks Arabic, and English L2 learners likely do not have a target English-speaking country selected for migration. This precludes the utility of integrativeness as a source of L2 motivation, so in this situation, other theories should be applied.
Identity-Based Second-Language Motivation
As a response, Dornyei (1994) sought to define a more general description of motivation in L2 learning, and further promoted its characterization during the L2 learning process so as to improve sustained motivation throughout the entire SLA journey. As a result, Dörnyei and Ushioda (2011) conducted an extensive line of research seeking a more standard way of understanding L2 motivation. As described earlier, Dörnyei developed the L2MSS, which took into account the self, the context, and any other motivations relating to L2 learning (Dörnyei, 2009; Dörnyei & Ushioda, 2011; MacIntyre et al., 2009).
The Ideal L2 Self (IL2)
The L2MSS was based on Dörnyei’s observation (MacIntyre et al., 2009) that L2 learners invariably expressed hopes and dreams associated with what they aspired to do as a future L2 speaker. Based upon the concept of possible ideal selves put forth by Markus and Nurius (1986), Dörnyei and colleagues observed that having a strong concept of a future L2 self can be a powerful motivator for the L2 learner (MacIntyre et al., 2009). This concept of IL2 serves as a motivator by forcing the learner to compare the IL2 with the current L2 self, and if the gap is large, strive to reduce it through L2 achievement (Dörnyei, 2009; MacIntyre et al., 2009).
Dörnyei’s point about the IL2 was also based on a concept put forth by Higgins (1987) called the self-discrepancy theory. This theory holds that in addition to a real self concept, there is also the idea of an ideal self-concept, and it is the discrepancy between the two that results in motivation to reduce the discrepancy. In other words, the less a person feels their current self-concept measures up to their ideal self-concept, the more vulnerable they feel, and they seek to reduce that vulnerability by trying to achieve their ideal self-concept (Higgins, 1987).
The Ought-to L2 Self (OL2)
Higgins’s (1987) self-discrepancy theory inspired Dörnyei to also consider not only the individual learner’s IL2, but what the learner might believe is expected of them for their future L2 self by way of societal pressure—the likely source of the emotional vulnerability felt when there is a discrepancy between the perceptions of ideal and real self-concept. In addition, based on discussions of an ideal “ought to” self by Markus and Nurius (1986), Dörnyei also suggested that if there was a discrepancy between the current self and what “others” have as a concept for the L2 learner, this should also serve as a powerful motivator for L2 learning. He termed this type of a motivator as the ought-to future L2 self (Dörnyei, 2009). This implied that the learner envisioned ought-to L2 future selves, or future selves who represent how the learner ought-to perform in the L2 according to groups in society. It also implied that there could be a variety of ought-to L2 future selves, as they responded to societal pressure from different groups (Dörnyei, 2009).
L2MSS Instrumentation and Measurement
Unlike Gardner, Dörnyei did not develop an instrument specifically to measure L2 motivation according to this L2MSS, but he philosophized about how factors might be determined in an instrument by pointing out that there may be multiple conceptualizations of “others” as sources of societal pressure, so potential multiple ought-to future L2 selves (Dörnyei & Ushioda, 2011). Dörnyei’s writing also points out that there may be influences from the L2LE that need to be figured into L2 motivation, but did not give explicit recommendations on measurement (Dörnyei, 2009; Dörnyei & Ushioda, 2011).
Dörnyei’s L2MSS gave rise to researchers developing instruments with the intention of measuring IL2, and various ought-to future L2 selves they felt might influence L2 motivation in their particular sample, as well as other influencing factors (Al-Hoorie, 2016; Moskovsky et al., 2016; Taguchi et al., 2009; Tahmouresi & Papi, 2021). This research showed that measuring IL2, ought-to future L2 selves, and other constructs of L2 motivation under the L2MSS was challenging, and difficult to generally characterize across populations. This was partly due to differences in the sample and the L2LE from study to study, but another challenge was introduced by the fact that each research team developed different instruments reflecting different components of L2 motivation.
Previous Findings
In L2 motivation studies by Taguchi et al. (2009), for example, IL2 was measured and found to be correlated with their measurements of integrativeness and “instrumentality,” which is a synonym for a term defined by Gardner, “instrumental motivation” (Masgoret & Gardner, 2003). In Gardner’s context, the French L2 learners inspired by practical considerations, such as wanting to do work or studying in the L2, were instrumentally motivated (Masgoret & Gardner, 2003). Although IL2 was measured in multiple studies, the approaches differed, so it is hard to compare even these across studies (Al-Hoorie, 2016; Moskovsky et al., 2016). Regression analysis revealed some evidence of IL2 being associated with L2 proficiency and other influences on L2 motivation, but without a standard measurement of L2MSS future selves, it is difficult to establish repeatable findings (Al-Hoorie, 2016; Moskovsky et al., 2016).
Further, some studies have had difficulty measuring ought-to future L2 self influences. Although various measurements have been developed, and many have been shown to be associated with L2 achievement, no particular measurement approach has dominated (Alqahtani, 2017; Taguchi et al., 2009). In fact, Papi et al. (2019) observed that it was particularly challenging to separate out IL2 from other sources of L2 motivation, specifically the ought-to L2 future selves. The authors observed that although studies repeatedly found IL2 to be associated with L2 achievement, the various ways the ought-to L2 future self had been measured did not also seem to be associated with L2 achievement, suggesting a potential measurement issue (Papi et al., 2019).
Challenges Measuring Collectivist Cultures
Papi et al. (2019) also noticed a pattern, in that this particular challenge measuring the ought-to L2 future self was seen more often in cultures considered to be collectivist in the framework of Hofstede’s individual versus collectivism (I/C) theory (Hofstede, 2011; Huppert et al., 2019) such as Japan and Iran (Papi et al., 2019). The authors implied that the lack of relationship between the measured ought-to future L2 self factors and L2 achievement reflected ambiguity in the learner as to what the so-called “others” feel in the collectivist culture, and therefore, it was challenging to separately measure ought-to L2 future self motivations and IL2 in those settings (Papi et al., 2019).
The Hofstede Model characterizes national cultures by six dimensions, of which I/C is one (Hofstede, 2011). The term “collectivism” refers to the I/C dimension, which can be characterized and scored on a 100-point system, where 0 represents collectivism, and 100 represents individualism (Hofstede, 2011). Although the model has been criticized as biased toward higher-income countries, the model does perform well on the I/C dimension in characterizing the relative level of collectivism and individualism in cultures, as the 100-point scale provides a large range to allow for a wide distribution (Huppert et al., 2019). Existing studies of education have used the framework of collectivism versus individualism to better understand micro-political dynamics (Or & Berkovich, 2023). Those in individualist environments have independent self concepts, goals that are independent from one another, behaviors driving by beliefs in themselves, and social interactions driven by rationality (Or & Berkovich, 2023). By contrast, those in collectivist environments have interdependent self concepts, goals that depend on one another, behaviors driven by norms, and social interactions based on relationships rather than rationality (Or & Berkovich, 2023). Studies concerning the individualist-collectivist divide in education between countries hypothesize that these country-level orientations may impact educational delivery, although evidence suggests that educational success in each country is multi-causal (Toyama & Yamazaki, 2022).
Reflecting upon existing research, much of Gardner’s and Dörnyei’s studies were based in higher-income countries with higher individualism scores, such as Canada and Hungary, while the studies citing challenges identifying ought-to L2 future selves were from Japan and Iran, with higher collectivism scores (Hofstede Insights Organisational Culture Consulting, 2021). It is possible that the pattern being seen where the ought-to L2 future self is especially difficult to measure is more prominent in countries with higher collectivism, because the IL2 and ought-to L2 future self are less separated. If this is indeed the case, a study of L2 English learners set in the Arabic-speaking KSA, which has an even lower collectivism score than Iran and Japan, would likely have difficulty finding separate constructs for IL2 and various ought-to L2 future selves (Hofstede Insights Organisational Culture Consulting, 2021). In a study of the individualist-collectivist divide in education across countries, individualist countries received high scores indicating high levels of individualism (e.g., United States (US) = 91, United Kingdom (UK) = 89, and Australia = 90), while KSA—like the other Arab countries in the study—received the low score of 38, indicating it is collectivist (Toyama & Yamazaki, 2022).
Measuring the L2 Future Selves in the Kingdom of Saudi Arabia
Current attitudes are changing quickly in KSA largely due to the implementation of the national strategy of Vision 2030, which began in 2016 (Saudi Vision 2030, 2020). Prior to 2016, KSA had been on a long-term economic path that relied on revenue from the petroleum industry. The implementation of Vision 2030 sparked new interest in diversifying and internationalizing industries in KSA outside of the energy sector (Saudi Vision 2030, 2020). Although in the past, speaking English by a Saudi in KSA was seen as either boastful or denying of culture, recently, there has been a noticeable shift in KSA toward a positive view of English-speaking and L2 English learning (Alrabai, 2016; Faruk, 2023; Mahboob & Elyas, 2014). Historically, English was used by non-Saudi associates as part of service encounters, and was not seen as prestigious or necessary (Alrabai, 2016; Faruk, 2023). Hence, there was a lack of emphasis on L2 English study in KSA (Alrabai, 2016, 2018; Faruk, 2023).
By contrast, Vision 2030 set forth an ambitious agenda with goals focusing on reforms in the areas of education, tourism, entertainment, and international investment (Saudi Vision 2030, 2020). In addition to an emphasis on internationalizing the Saudi labor market, Vision 2030 seeks to address the gender gap in the Saudi workforce (Swaantje, 2018). Therefore, Vision 2030 has represented a push in KSA toward Saudis learning English, especially those who want to enter the workforce. The emphasis on educating women is seen in the high level of enrollment in the colleges of languages at women’s public universities in KSA. The largest of these serves about 60,000 students annually, and its college of languages is only one of the several colleges of languages and translation (COLTs) across the nation at KSA universities (Al-Jarf, 2008a). At the largest women’s university, over 2,000 students are enrolled in one of the four programs at the college of languages, three of which are focused on English.
Purpose and Research Questions
The purpose of the present study was to measure the L2 motivation of female students in the English program at a KSA public university. Evidence in the literature suggests that Saudi female learners show higher levels of L2 motivation than their male counterparts (Alrahaili, 2013; Moskovsky & Alrabai, 2009). Moreover, Saudi females tend to favor using English over Arabic in labor market settings, medical arenas, and technological fields (Al-Jarf, 2008b). Leaders in this setting have a general understanding that students are motivated toward L2 learning, but report that no quantitative studies have taken place specifically on L2 motivation. As KSA is a collectivist culture, it is not clear that the constructs of the L2MSS, which involve measuring separate future selves, could be measured accurately in this group. The study had two research questions: (1) How can an instrument be developed to measure future selves as L2 motivators consistent with the L2MSS in Saudi female college students enrolled in a higher education English program? and (2) how do the findings from the first research question apply to collectivist cultures?
Methods
To address both research questions, first, items in the L2 motivation instrument were designed according to a framework consistent with the L2MSS. Next, this instrument was distributed to two cohorts of female KSA college students who were English majors. A subset of the second cohort also participated in a follow-up interview about sources of L2 motivation. Finally, analytics were conducted to assess the performance of the instrument using factor analysis as well as other analyses aimed at evaluating validity and reliability. Further, information discussed in the interviews was considered in the interpretation. Prior to commencement of research activities, ethical approval was obtained from both the University of Sussex Cross-school Research Ethics Committee (C-REC) and the Princess Nourah bint Abdulrahman University Institutional Review Board (IRB).
Participants
Two cohorts of female KSA college students who were English majors completed the instruments: (1) a “Pilot Cohort” of female college students in KSA majoring in English intended to test the instrument when it was in draft format and (2) a “Main Cohort” of female college students majoring in English at a KSA public university who participated in a larger study in which the instrument was included. It was originally intended to analyze these cohorts separately, but due to the small sample size of the first cohort, the cohorts were pooled to increase statistical power.
Instruments
The instrument was to measure only L2 motivational constructs from the L2MSS as they would apply to these particular students. It was envisioned that there were three future selves to take into account as domains in the instrument: the learner’s ideal future L2 self (IL2), the learner’s estimation of their parents’ ideal future L2 concept (one type of ought-to L2 future self, parental expectation, abbreviated PE), and the ideal future L2 concept for the learner arising from others (such as teachers and peers, a second type of ought-to L2 future self, abbreviated OL2). While PE and OL2 are two different constructs of the ought-to L2 future self, PE was believed to be a separate construct, divergent from OL2, because in Saudi society, the parents are praised by the community for their child’s academic success (Al-Nafisah, 2000).
To design the instrument, items were selected from previous instruments used in the Arabic-speaking population (Alqahtani, 2017; Moskovsky et al., 2016; Ryan, 2009; Taguchi et al., 2009), and placed on each of the three domains (see Table 2 in the Results for items and sources). The first of these, by Alqahtani (2017), studied an instrument that was shown to have 11 robust factors/subscales. The second, by Moskovsky et al. (2016), developed a five-factor/subscale instrument, and provided evidence behind these factors. Ryan (2009) developed a 100-item instrument to study 18 motivational variables crossing a broad spectrum of theories in Japanese students learning English, and strong support for several motivational variables directly connected with IL2. Finally, Taguchi et al. (2009) studied a 10-factor/subscale instrument and found evidence for each factor. Because the current instrument was to be developed and delivered in Arabic, items already in Arabic were used verbatim, and items that had not been translated into Arabic in previous research were translated by the author (see Table 2 notes). The author’s translation was reviewed by several Arabic learners of English and Arabic-English bilingual professors for accuracy. The instrument included 4 items for PE, 10 items for IL2, and 9 for OL2 which were statements. Respondents were asked to assign a score for each item on a scale of 1 = totally disagree, 2 = disagree, 3 = neutral, 4 = agree, and 5 = totally agree.
Procedure
Pilot Cohort
In 2018, two L2 motivation researchers and an academic in the school of education at the setting, who were the experts available at the time, reviewed the draft instrument and provided feedback to improve the items and to verify the English-to-Arabic translation. A revised online version of the instrument was created using web-based software Google Forms. An anonymous link was sent to female KSA English major students via WhatsApp by English professors known by the author who teach at the data collection learning site and other women’s colleges in KSA (not specified). There were 47 responses, and because all fields had been marked “required,” there were no missing data. After the pilot, it was determined that two of the items on the PE subscale were redundant, so one was dropped, leaving three items on the PE subscale, and 22 items on the entire instrument.
Main Cohort
Later in 2018, the author visited 13 level 3 and 4 English classes (equivalent to sophomore level at Western universities) taking place at the public university setting serving all students who had declared one of the three English majors available at the CoL: Linguistics (seven classes), English Literature (two classes), and Translation (four classes), representing approximately 400 students. She presented the study and invited the students to participate via a Google Forms link as was done with the Pilot Cohort. Because all fields had been marked “required,” there were no missing data, and 207 valid responses.
Respondents were asked if they wanted to identify themselves and participate in follow-up research, so a subset underwent subsequent interviews about their L2 motivation as has been recommended as a way to better understand the meaning behind the results of quantitative instruments (Islam et al., 2020; Ushioda, 2016). Interview results were used to shed light on the quantitative results.
Analysis
To answer the first research question, analysis was conducted on the quantitative data from the instrument. To assess reliability of factor subscales, a Cronbach α was conducted on items separately in each subscale (using the R language, package psych) (R Core Team, 2021), and at level of .70 was determined to be considered acceptable. After this, confirmatory factor analysis (CFA) was used to confirm factor loadings for each of the three factors (IL2, OL2, and PE) using tools from the psych package. Even though CFA was used, varimax rotation was also used, as there was significant uncertainty around the selected domains, and it was felt that the analysis could benefit from improved clarity on determining factor loadings (Forina et al., 2005). Items were considered loaded on a factor if the estimate was 0.6 or higher, and if the estimates on the other two factors were lower. The package nfactors was used to develop a scree plot, and the lavaan package was used to evaluate model fit (Rosseel et al., 2022). After evaluating factor loadings, each factor or subscale was scored by summing the remaining items in each factor, and Pearson correlations between summed scores of the subscales were evaluated for convergent and discriminant validity, with α set at 0.05 (using R package Hmisc). Plots of item responses were made with the likert package. A copy of the final instrument is attached at the end of this document (after the references).
To help validate the findings to the first research question as well as provide insight into the second research question, qualitative research by interview was conducted with a subset of the participants who completed the instrument. Interviews were conducted in Arabic, although respondents occasionally spoke English, and quotes relevant to the analysis were translated into English.
Results
As shown in Table 1, 47 records from the Pilot Cohort and 207 records from the Main Cohort were available for analysis, for a total sample of n = 254.
Sample Characteristics.
Note. First row percentages are of the total, and percentages in all other rows are column percentages.
As shown in Table 1, the Pilot Cohort was somewhat older than the Main Cohort, but the age range was small (18–27 years).
Factor Analysis Results
Table 2 presents the results from the CFA loadings, and Figure 3 presents the scree plot.
Factor Analysis Results.
Note. Factors were considered loaded if their estimate as 0.6 or greater on one factor, and bold indicates factor loading for each item that loaded. Codes were assigned to the items as follows: PE1-3 represent items on the “parental expectation (PE)” domain, O1 to O9 represent items on the “ought-to L2 self (OL2)” domain, and I1 to I10 represent items on the “ideal L2 self (IL2)” domain. Items O8 and O9 did not load. Item sources: aoriginal statement from Ryan (2009) translated into Arabic in Alqahtani (2017), boriginal statement from Taguchi et al. (2009) translated into Arabic in Alqahtani (2017), coriginal statement from Taguchi et al. (2009) translated into Arabic by the author, doriginal statement from Moskovsky et al. (2016) translated to Arabic by the author.
For the model shown in Table 2, the Comparative Fit Index (CFI) was 0.875, the root mean square error of approximation (RMSEA) was 0.0.092, and the standardized root mean square residual (SRMR) was 0.065, indicating that the model was admittedly not a very good fit. As part of previous research, factor loadings for each cohort were evaluated separately, and it was found that while IL2 item loadings were strong, OL2 and PE items had weak loadings, and were not reliably loading on intended domains (data not shown). When the cohorts were pooled as in the current analysis, the factor loadings for all domains were strong, with all items loading on their intended factors at 0.6 and greater with the exception of two items on the OL2 domain (see Table 2). For these reasons, those 2 items were removed for the remainder of the analysis, leaving 3 items on the PE subscale, 10 items on the IL2 subscale, and 7 items on the OL2 scale, for a total of 20 items. This three-factor loading was confirmed visually by the scree plot (see Figure 1).
Scree Plot. The optimal coordinates are identified as three, so it is reasonable to utilize three subscales.
To evaluate the distribution of responses by item among the items remaining in each domain, item distributions for IL2, OL2, and PE were plotted in Figures 2 to 4, respectively.
Distribution of responses per item for items remaining in ideal L2 self domain. As shown, responses were somewhat homogenous, in that there was between 69% and 86% of agreement on the items in this subscale.
Distribution of responses per item for items remaining in ought-to L2 self domain. Compared to the items on the ideal L2 self subscale, there was more variation in item responses, with agreement with the items ranging from 23% to 45%.
Distribution of responses per item for items remaining in parental expectation domain. Like with the ideal L2 self subscale, there was a high level of agreement with the items in this scale, ranging from 60% to 75%.
Figure 2 shows the distribution of responses for each IL2 item (1 bar per item). Items are labeled on the y-axis on the left side using the codes in Table 2. Per the legend, the dark green represents the percentage of responses of “5 = strongly agree” to that item, and the light green represents the percentage of responses of “4.” Together, these comprise the percentage who agreed, listed on the right side along the y-axis. Likewise, the dark gold represents percentage of responses of “1 = strongly disagree” by item, and light gold represents percentage of responses of “2.” They comprise the percentage who did not agree, listed on the left side along the y-axis. The center gray area represents percentage of responses of “3,” listed along the center y-axis. Figure 2 shows there was a uniformly high level of agreement with all the items in this sample; the lowest level of agreement was for item I9, which was 69%, and highest was for I3 (86%). This is in contrast to Figure 3 (formatted like Figure 2), which shows distribution of responses to OL2 items. Figure 3 shows a larger variance of item responses, with the percentage agreement ranging from 23% to 45%. Figure 4 (formatted like Figure 2) shows the distribution of item responses for PE, depicting a high level of percentage agreement (60%–70%) similar to the IL2 distributions seen in Figure 2.
Table 3 shows the results for the scored subscales for the sample.
Score Summary.
Note. sd = standard deviation.
As shown in Table 3, mean scores for both cohorts were predictably high for IL2 (Pilot Cohort mean 40.9; Main Cohort mean 42.2), where the maximum score possible was 50, and PE (Pilot Cohort mean 11.5; Main Cohort mean 11.8), where the maximum score possible was 15. For OL2, where the maximum score possible was 35, the scores were on average lower (Pilot Cohort mean 19.4; Main Cohort mean 20.8). Predictably, post-hoc t-tests conducted between the Pilot Cohort and the Main Cohort on these results were not statistically significant (data not shown), providing evidence that both cohorts were drawn from the same population. Per Table 3, for all subscales, standardized Cronbach α’s were greater than .70, so were deemed acceptable. However, Cronbach α results for subscales evaluated in the cohorts separately did not reliably reach at least .70 (data not shown).
Correlation Analysis Results
Table 4 shows the results of the correlation analysis among summed subscale scores.
Correlation Matrix for Subscale Scores.
Note. Numbers above the diagonal represent Pearson’s correlation coefficient r, and number below the diagonal are the corresponding p-values.
As shown in Table 4, all correlations were positive, and the highest correlation between subscales is between PE and OL2 and is statistically significant (r = .46, p < .0001). Although factor analysis results suggests that OL2 and PE are distinct constructs, given KSA culture, where children of all ages are seen as representing their parents in terms of reputation, it is not surprising that female college students studying L2 English in KSA would envision a future L2 self that would satisfy both their parents’ expectations as well expectations of other people in society at large. The second highest correlation which was also statistically significant was between IL2 and PE (r = .34, p < .0001), which unsurprisingly reflects that the formulated ideal L2 selves in these students was at least somewhat consistent with parental expectations. The lowest correlation which only approached statistical significance (p = .0767) was between IL2 and OL2, and this may reflect how KSA students would prioritize adjusting their future ideal L2 self to meet parental expectations more so than to meet the expectations of generic “others.”
Interview Results
In the main study, 24 participants were interviewed. With respect to sources of L2 motivation measured in the instrument, participants expressed a desire to align their envisioned future L2 self with that of the vision of their parents, even if it meant essentially following whatever their parents wanted as their education progressed. One participant, pseudonym Hala, expressed it well in this quote:
I didn’t choose law because my father’s approval is more important to me. After I graduate, I want to complete my master’s in English. If I don’t get accepted in the program because of my low proficiency, I would have no problem choosing a field other than English.
Therefore, if the student’s vision of themselves and what they perceived as their parent’s vision of them began to diverge, it became a source of tension, as expressed by participant with pseudonym Amira:
Whenever I get a bad grade, I conceal it from my father. Not because I’m afraid of him, but you know how fathers want their daughters to excel. He wants me to be excellent in English. To me, I think it’s normal to get a bad grade because I don’t expect to perform well all the time.
Hala represents an individual expressing an unclear IL2, yet a clear vision of what her father is expecting, which is her father’s approval of whatever she is studying. A similar sentiment is expressed by Amira. But by contrast, Jawaher (pseudonym) explained how she had a clear vision of her IL2, and was able to convince her parents to share a similar vision of her future L2 self:
My parents didn’t encourage me to major in English because they believed I had already knew (sic) English. They tried to convince me to choose another major. But I’m glad I chose English because it’s the best choice for me. Now they have changed their minds and became proud of me since they saw me studying hard and doing my best.
These findings are consistent with the positive correlations seen between the IL2 and PE subscales. They also explain to some degree why OL2 has a lower correlation with IL2 than PE, in that the generic “others” may be evaluating both the student (IL2) and the family (PE) together as a unit.
Discussion
This analysis showed that it was possible to develop an instrument to measure future L2 selves consistent with the L2MSS in Saudi female college students enrolled in a higher education English program. The instrument was able to measure the strength of the vision of the student’s ideal L2 self, as well as differentiate between future ought-to L2 selves envisioned to satisfy parental expectations versus expectations of society in general. However, as KSA is a strongly collectivist culture, the resulting measurements of the various L2 selves were directly correlated. In other words, although these sources of L2 motivation may be distinct, in highly collectivist cultures, it might be expected that they are strongly correlated, and therefore, do not function as totally independent sources of L2 motivation as might be seen in more individualistic cultures.
Although this analysis demonstrated strong and clear constructs for OL2 and PE, as described earlier, individual analyses of the cohorts separately did not reveal strong or clear factor loadings for these two subscales. It was only after the data were pooled into a larger dataset did the factor loadings increase in a consistent pattern and reveal the constructs. Although both cohorts were thought to be relatively homogenous in terms of L2 motivation, until the datasets were pooled for analysis, patterns of OL2 and PE did not emerge.
This challenge in detecting ought-to L2 motivational constructs in collectivist cultures has been reported by previous researchers (Papi et al., 2019; Teimouri, 2017). In a study of a diverse group of higher education students in the US studying L2 English, researchers designed an instrument based on the L2MSS that included two factors for IL2 and two for ought-to future L2 selves (Papi et al., 2019). Although they had a large sample size (n = 257), the authors noted that about 33% of the participants were from China (Hofstede score 20), and another 10% were from South Korea (Hofstede score 18), both of which are collectivist cultures (Hofstede Insights Organisational Culture Consulting, 2021; Papi et al., 2019). The other 57% of students were from Panama, Columbia, Turkey, and Iran, which are not Westernized cultures and have low Hofstede scores (Papi et al., 2019). They were able to identify four distinct future L2 self constructs in CFA, but the correlations between all of the constructs were positive, with several greater than r = .4, and most were highly statistically significant (p < .001). The current study only sought to identify three constructs, and though the group was more homogenous, a sample of approximately the same size (n = 254) was needed before distinct factors emerged in CFA. Even so, the correlations between subscales were positive, and two were statistically significant.
Results similar to those seen in the current study were also observed in previous research on male college students studying L2 English at an all-male university in KSA (Al-Hoorie, 2016). The researcher measured twelve different influences on L2 achievement, including one measurement of IL2 and one measurement of the ought-to L2 future self (Al-Hoorie, 2016). As with the current study, these two subscales had a weak positive correlation that was statistically significant (r = .14, p < .01). Like with the current analysis, this study included a sufficient sample size (n = 311) of a homogenous group to be able to detect these distinct but correlated factors in the analysis.
Although other authors have faced challenges developing instruments to measure L2MSS constructs in L2 learners from collectivist cultures, both the current study and the study of male KSA college students suggest that it is possible to develop a valid and reliable instrument that identifies distinct constructs of the IL2 and the various ought-to L2 future selves. However, in collectivist cultures, it is expected that these will all be correlated, as was found in the current study. This leads to the question of what the value is in being able to measure separate L2MSS constructs in a society where all of the societal pressure is directed in roughly the same direction, and the envisioned L2 future selves overlap significantly.
With respect to the current sample, when an unclear IL2 was expressed, but a clear understanding of parental expectations (which were invariably high) was understood, this could represent a learner with a low IL2 subscale value coupled with a high PE value. Participants in this situation clearly expressed tension due to the collective nature of the culture. By contrast, a high IL2 subscale score with a low PE score might represent a learner who felt unsupported by her parents. Though a learner in that circumstance might be motivated by their high IL2, given the collectivist nature of KSA culture, the lack of expression of parental expectation may serve as a significant demotivator. While many examples exist of KSA students working within their parents’ expectations to try to align them with their own by experimenting with different courses of study, in the end, culturally, the students generally expressed they perceived the most motivation and satisfaction when the self and parental expectations were aligned. Now that an instrument has been developed that can reliably measure these constructs in female Saudi English majors in college, there is increased potential for a better understanding the influence of various future L2 selves on L2 motivation and how these concepts can be used to increase L2 motivation in the educational setting.
Strengths and Limitations
This study has several strengths, but it is not without limitations. The main strength of this study is the successful development of an L2 motivation instrument measuring L2MSS constructs female higher education L2 English learners in KSA. The findings also shed light on challenges other researchers measuring L2MSS constructs in samples of participants from collectivist cultures have faced in measuring separate, distinct constructs, especially for the ought-to L2 future selves. A limitation of this study is that it is not clear how well the L2 motivation instrument would work in other populations, even in KSA, as the items are directed specifically at younger female L2 learners who are still under the strong influence of their parents. The findings may even be specific to students at the data collection learning site. Further, the usefulness of the instrument is questionable, since all future L2 selves are expected to be directly correlated in collectivist cultures, so being able to tell them apart is likely of limited utility. Future studies will need to be conducted to confirm the validity, reliability and utility of the instrument in populations both within and outside of KSA. Additionally, how the measurement of L2 motivation in learners from collectivist cultures can provide educators opportunities to increase L2 motivation in the classroom should be explored.
Conclusion and Recommendations
In conclusion, this study of female Saudi higher education students majoring in English found that they had a clear vision of their ideal future L2 selves, and that their vision of expectations placed on them for their future L2 self by their parents and by society at large could be measured as distinct L2 motivations. The results also shed light on why measuring future L2 selves in learners from collectivist cultures has been challenging in other research settings, and how different measurements of future L2 selves will predictably be directly correlated in collectivist cultures, where individualism is low. Future research is needed to verify the results of this study, and to explore how measurement of future L2 selves in learners from collectivist cultures can lead to educational strategies that can increase L2 motivation.
Footnotes
Acknowledgements
The author would like to thank Drs. Simon Williams, Roberta Piazza, and Evan Hazenberg, for providing guidance and scientific review.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
Ethics Statement
Prior to commencement of research activities, ethical approval was obtained from both the University of Sussex Cross-school Research Ethics Committee (C-REC, Reference #ER/DAMS20/2) and the PNU Institutional Review Board (IRB, Approval #18-0015).
Research Data
Reasonable requests for research data will be honored by the corresponding author.