IAVLE = initial appraisal of vocabulary learning experience; SRCvoc = self-regulating capacity in vocabulary learning; SVLI = strategic vocabulary learning involvement;
MVLT = mastery of vocabulary learning tactics; VOCkno = vocabulary knowledge;
PAVLT = postappraisal of vocabulary learning tactics.
Figure 2.9 Tseng and Schmitt (2008) model of vocabulary learning SVLI
.48*
SRCvoc
IAVLE PAVLT
VOCkno
MVLT
.62*
.46*
.56*
.68*
.67*
Comments on the Five Models Reviewed
The strengths of Tremblay and Gardner’s (1995) model lies in the addition of more elaborate motivation factors in the L2 learning motivation model with a suggestion of three mediators: goal setting, valence, and self-efficacy, between language attitudes and motivational behavior. Also, the model hypothesized achievement would be directly influenced by French language dominance and motivational behavior, and the hypotheses were supported. The weaknesses of this model may be on the hypotheses of language attitudes and French language dominance as the exogenous variables.
French language dominance, identified as the perceived performance and frequency of French use, would probably be better specified as an endogenous variable and has a
45
direct influence on achievement in the initial hypothesized model. However, this construct was hypothesized to have another indirect influence on achievement, with adaptive attribution, self-confidence, and motivational behavior as the mediators. The causal relationships seem not to reasonably present the learning process by hypothesizing language dominance as the antecedent of these three mediators, which should be presented conversely. Further, the model was discovered to have too many modifications and the researchers did not provide sufficient theoretical support and model fit indices. These two deficiencies may pose problematic challenges to the theoretical hypotheses in the earlier stage.
The strengths of Gardner, Tremblay, and Masgoret’s (1997) model are that it includes support for significant individual difference measures, such as language attitudes, motivation, anxiety, self-confidence, language aptitude, learning strategies, and field independence, all of which have been shown to correlate significantly with language achievement. Further, the items in many variables are both positively and negatively worded and measured. The weakness of this model may be attributed to the identification of the latent variables with an exploratory factor analysis (EFA).
Structural equation modeling is an approach in which the researcher generally specifies a priori knowledge based on theoretical grounds (Tremblay & Gardner, 1995). It is suggested that the researchers in this study use confirmatory factor analysis (CFA), not exploratory factor analysis, to verify and validate the factors investigated in the model because the researchers need to test the hypothesized causal relationships in the measurement model as well as in the structural model. Particularly when the factors have been disclosed to be important motivational components in prior research, an EFA may not be the most appropriate manner in which to specify the factors. The second flaw of the study is that the construct self-confidence,
46
including three indicators—language anxiety, self-confidence, and self-rated proficiency (Can Do), with the latter two revealing similar affective information for one problem and an identical variable name for both one indicator and the latent variable for another.
The strengths of Csizér and Dörnyei’s (2005) model are that it consists of both macro-perspective and micro-perspective social influences, i.e. cultural interest, attitudes toward L2 speakers and language learning milieu. Further, most of the latent variables concern the affective antecedents of motivation intention, providing more comprehensive facets for the antecedents of motivation formation. The final results seem to contradict with Csizér and Dörnyei (2005) earlier statement, “… both Integrativeness and Instrumentality are hypothesized to be directly linked to the motivated behavior (i.e. Language Choice) does not need much justification because these variables have been the most often researched concepts in this field in this respect” (p. 26). However, instrumentality in the final model becomes an antecedent of integrativeness, which the authors use to link their interpretation from self perspectives, depending on the extent of internalization of extrinsic motives:
internalized instrumental motives are associated with the ideal L2 of integrativeness, whereas noninternalized instrumental motives are associated with the ought-self. The distinction of a single construct into two types of selves may be problematic because the modification of the construct, instrumentality, with respect to the causal relationships, seriously opposed its original proposed causal hypotheses. Second, the interpretation of instrumentality in terms of the two types of self does not have empirical support from the current data; therefore, it may appear awkward to explain instrumentality in this respect. The researchers seem to intentionally attribute the impact of instrumentality on integrativeness, which may not do justice to the 13-14
47
year-old schoolchildren because they may not have an impending need to associate their choice of learning a foreign language with utilitarian benefits or obligations (Warden & Lin, 2000). Therefore, hypothesizing instrumentality as an antecedent or a mediating factor of integrativeness in this model may be controversial.
The strengths of Yashima, Zenuk-Nishide, and Shimiza’s (2004) study lie in the small hypothesized model with a clear intentional behavior—willingness to communicate in L2 and the addition of the new construct, “international posture,” in the SEM model. Generally, the two studies show that learners’ perceived communication competence is most strongly related to WTC, which, in turn, results in the frequency and the amount of communication. Though this study takes contextual variables into account, such as frequency and the amount of communication with host nationals, the focus is concerned with the “intergroup” construct within a multicultural setting rather than the “interpersonal” interactions within a monoculture.
The intergroup interactions may not be likely to motivate EFL learners who generally do not have much L2 contact with native speakers of English. In addition, those who could afford to participate in student exchange programs tend to have better economic status and to have stronger motivation to learn the second language with respect to having opportunities to join the community of the target language. It is indicated that the higher parental social economic status was found to be associated with their involvement in the students’ learning at home and at school (Ma, 2000; Phillipson, 2009). Thus, parents may likely provide more economic and social support to the students in the exchange program. The factor, international posture, is an interesting construct, but it may not be a motivational component to most common EFL learners.
Despite the rare interactive environment, the international posture has shown significant effects on motivation to learn an L2, willingness to communicate, and
48
frequency of communication, which seems to indicate that some of the mediating constructs are not necessary.
The strengths of Tseng and Schmitt’s (2008) article lie in that they provide a rather comprehensive design of a structural model of motivated vocabulary learning, in contrast with previous studies focusing on the internal structure of motivation constructs, such as integrative orientation, instrumental orientation, and integrative motive (Gardner, 1985), the ideal L2 self, ought-to-L2 self, and L2 learning experience (Dörnyei, 2005), and the distinction between intrinsic and extrinsic motivation (Noels, et al., 2001; Wang & Guthrie, 2004; Wigfield & Guthrie, 1997).
Most previous studies did not provide sequential causal links among variables, Tseng and Schmitt’s study offers a model hypothesizing causal relations between the initial appraisal of vocabulary learning experiences, strategic learning involvement, and mastery of vocabulary learning tactics. The hypothesized model consists of six latent variables which showed not only the initial motivational state influencing the process of task performance, but also a retrospection of task performance which is likely to influence in turn the initial appraisal. The most powerful strength of this model lies in its cyclical design of causal links, indicating learners who perceived themselves as achieving the learning goal and making proper attribution for their successful learning are also more likely to sustain their high self-efficacy, positive attitudes, and a favorable emotional climate for the subsequent task execution (Dörnyei, 2001;
Pintrich & Schunk, 2002; Weiner, 1986 1992). This model seems to help learners energize and upgrade themselves through the whole loop and its vitality may be sped up and intensified once the learners become familiar with the task performance in the loop. However, the model also has its flaws. First, a rather comprehensive model not only should include constructs concerning on initial appraisal of learning experiences,
49
learning behavior, vocabulary knowledge and post-appraisal learning tactics, but also should consist of constructs related to how students plan to execute their learning behavior, which will help to tackle difficulties which may have emerged during the process. Second, the initial appraisal of a student’s vocabulary learning experience may need to include social influences as researchers recommended in the 1990s.
The common deficiency that the five models reviewed share lies in the fact that none of them clearly specified learning processes in terms of discrete actional phases, given a consideration of macro top-down methodological design. Though Tseng and Schmitt (2008) delineates their model a process of vocabulary learning, the initial appraisal of the vocabulary learning experiences doesn’t seem to cover sufficient aspects of societal impacts on L2 learners’ motivation, neither do Tremblay and Gardner’s (1995). Second, no model among these five placed an emphasis on the intention construct, particularly on implementation intention, a construct that facilitates learners to specify when, where, and how to plan their studies which make their L2 learning behavior easier to execute and more effective in proceeding towards their goals. Another important perspective that several researchers have maintained (Landy & Becker, 1987; Tremblay & Gardner, 1995; Weiner, 1984) is that no single theory can effectively explain the dynamic characteristics of L2 learning motivation.
However, a major theoretical framework should be grounded so that it will and should provide an anchor to the constructs composed in the motivational model. In addition, using a structural equation modeling design in essence specifies a priori the direction of the causal relationships (Hair et al., 2010; Kline, 2005; Tremblay & Gardner, 1995), based on theoretical review. Despite the addition of a specific theoretical grounded construct into a designed model, the overall development of the whole model should also be constructed on a major theoretical framework that would provide the proposed
50
model with a unified sense of unity, instead of putting all the significant factors together into a model on one hand and hypothesizing the causal relationships without consecutive sequential effects on the other. Keeping the rules of thumb in mind, the current study will be grounded on Ajzen’s (1991, 2005) theory of planned behavior, reinforced by Gollwitzer’s (1993, 1999) theory of implementation, and Tseng and Schmitt’s (2008) process model of vocabulary learning in an attempt to construct a motivational model for second/foreign language learning.