The current research intends to measure the causal link between L2 learning intentions and self-regulatory capacity. Consequently, the researcher initiated a research design that adopted and modified the item pool for each subscale from Gardner (1985), Dai and Tseng (2011) and Liu (2008) in the following steps:
preparing a first version of the instrument; piloting this version; based on the pilot results designing the final version; and finally administering the instrument to a sample of language learners to validate it.
In order to measure students’ L2 learning intention and its influence on self-regulatory capacity, a two –phase study was designed to examine the interaction among variables. A pilot survey was conducted to validate the modified items questionnaire and corroborate the theoretical constructs of L2 intentions. The researcher intended to investigate whether the items of learning intentions could be categorized into five subscales as proposed in the previous research. Each item was treated as a variable, and factor analysis was done through subjecting the variables to Principal Component Analysis. In addition, reliability test was implemented to ensure consistency within the subscales. Subsequently, based on the valid and reliable data, a formal questionnaire was carried out and the collected data received multiple regression analysis to identify the factors that best predict self-regulatory capacity.
Participants The Pilot Study
The pilot study was conducted to finalize the intention inventory. The pilot study involved 116 Taiwanese senior high school students in Taichung. Judging from the
31
rank status of the school and their overall scores of the entrance exam, participants were of intermediate level. They were recruited from three intact classes, with two classes majoring in natural science and one in social science. Participants were all from the second semester of their sophomore year in high school.
The Formal Study
A total of 412 students participated in the formal study. The researcher recruited a bigger sample when considering the completion rate of the questionnaire and the sampling errors. Excluding the incomplete accomplishment of the data, there were 405 participants, including 162 males and 243 females, from 10 intact classes of four public high schools in central Taiwan. English was a compulsory subject for all the participants. Among the participants who had completed the items, 126 of them were in the first year of senior high school and 279 of them were the second graders. The participants have been learning English as a foreign language at school for minimally five years since students in Taiwan are required to take English as a compulsory subject starting from the fifth grade in elementary school. The subjects were asked to complete a questionnaire which was designed to elicit information concerning their goal intentions, implementation intentions, and self-regulatory capacity. The participants were given clear instruction and sufficient time from the researcher to answer the questions in an appropriate manner, and they were assured that the results would be confidential and would not affect their grades.
Instruments
In order to confirm the psychometric properties of the scales, three sections of items were presented to collect information on learners’ learning intentions and self-regulatory capacity. The categorized item pools were demonstrated in Appendix
32
A, B, C, and D.
Scale of Goal Intentions in Language Learning
The measurement of goal intentions was adopted from the Attitude/ Motivation Test Battery (AMTB) (Gardner, 1985), consisting of integrative and instrumental orientations. Along with the mental representation (“I intend to become/achieve X”) proposed by Gollwitzer (1996, 1999), the questionnaire items were phrased as “I intend to learn English because…” and “My goal in learning English is to…” to further analyze goal intentions in language learning. AMTB has been regarded as a well-designed instrument and as having the structure that follows the psychometric principles governing questionnaire theory (Dörnyei, 2005). Items were arranged and modified to suit the L2 learning context in Taiwan.
Scale of Implementation Intentions in Language Learning
On the other hand, consistent with the definition of implementation intention, the researcher specified when, where, and how learners exert intentions in the questionnaire items of language learning. The when and where components were merged as one situational orientation for its context-oriented character in task completion (“If there are many other things to do, I still try to spare time to learn English”), whereas the how elements were identified as strategic orientation for its tactical nature in dealing with language learning tasks and creating personal efficient means (“When studying English, I know how to use appropriate learning techniques”).
In the current study, the what aspect was added to elaborate on the decision-making process for improving the proficiency level (“If there are some chances to meet English speakers, I will try to think of some topics to talk with them to improve my oral skills”) as content orientation. Most of the items were adopted from Dai and
33
Tseng (2011). As mentioned earlier in Chapter 2, the research design of Dai and Tseng (2011) did not fully enumerate the intentional dimensions through confirmatory factor analysis. To find out possible unexplored variables in the motivational process, the collected pilot data would receive explanatory factor analysis. The items were rated for agreement on a 6-point Likert scale with anchors 1: strongly disagree and 6:
strongly agree.
Scale of Self-Regulatory Capacity in Language Learning
As for the measurement of self-regulatory capacity, the questionnaire contained items from five aspects: Commitment Control (“When studying English, I can effectively solve the problems I encounter”), Metacognitive Control (“When learning English, I think my methods of controlling concentration are helpful”), Satiation Control (“Once the novelty of learning English is gone, I easily become impatient”), Emotional Control (“When I feel stressed, I know how to handle it.”), and Environmental Control (“When learning English, I know how to set up an environment that can best facilitate my learning.”). There were total 18 items: 5 for Commitment Control, 4 for Emotional Control, 4 for Metacognitive Control, 2 for Satiation Control, 3 for Environmental Control, as shown in the Appendix N. The items were rated for agreement on a 6-point Likert scale with anchors 1: strongly disagree and 6: strongly agree. The SRClang scale was properly developed by Liu (2008) based on Tseng, Dörnyei, and Schmitt (2006) via Principal Axis Factoring only to find that the descriptors in the scale answered to one latent variable. It was hence suggested that self-regulation is unidimensional. This measure aims to measure the underlying capacity to regulate the learning behaviors. Self-regulating capacity would be conceptualized more like aptitude than as a series of discrete events (Winne &
Perry, 2000). A total score of self-regulatory capacity would be obtained as the sole
34
dependent variable. The reliability for SRClang scale was reported to be .93 for the pilot and the formal study, which can be regarded as internally consistent construct for further investigation.
Procedures
Before the formal study was done, the pilot version of the modified questionnaire should be tested. In the pilot study, goal intentions, composed of integrative orientation and instrumental orientation proposed by Gardner (1985) and combined with the structure of Gollwitzer (1999), was approached through 16 items, 7 items for integrative and 9 for instrumental. As for implementation intentions, 6 items for content orientation, 7 items for situational orientation, and 7 items for strategic orientation were adopted based on Dai and Tseng (2011). Self-regulatory capacity was measured by implementing Liu’s (2008) scale, for which self-regulation was realized as one single psychometric trait with high internal consistency (α = .93). The participants received the Chinese version of the questionnaire and were told that there was no time limit in completing it; they were also encouraged to report the items which were worded inappropriately.
Item Analysis
After the data were gathered, two kinds of item analysis were conducted:
Principal Component Analysis (PCA) and reliability analysis. With regard to the PCA, items were grouped in different subgroups and, having high within-correlations, considered to represent the same dimension (Tacq, 1997).
For the measurement of goal intentions, four components were extracted at first.
The items were grouped based on the loadings. Item 14 (designed as instrumental orientation) were deleted because this item switched its component during the
35
analysis, as observed in the first column in Table 2. For item 1, 3, 8, and 13, cross-loadings were discovered among components, which could be regarded as unsuitable for further survey.
Table 2
First Rotated Component Matrix for Goal Intentions in the Pilot Study Component
Integrative Instrumental Unknown Unknown
G15 .875 .095 .032 .151
G10 .828 .091 .217 .171
G14 .765 .072 .262 -.153
G16 .733 .156 .237 .183
G12 .709 -.034 .315 .371
G6 .640 .052 .151 .507
G4 .146 .904 .038 .160
G5 .095 .896 .084 .104
G7 .007 .847 .097 .258
G9 .074 .726 .306 -.207
G11 .081 .562 .088 -.424
G8 .318 .124 .799 .051
G13 .324 .161 .767 .162
G3 .114 .379 .543 .387
G1 .369 .091 .340 .684
G2 .532 .170 .124 .617
a. Rotation converged in 5 iterations.
After the removal of the five items from the pool for goal intentions, another Principal Component Analysis of the remaining 11 items was conducted. This time, two components were extracted, with the two explaining 68% of the variance, and the results demonstrated that the items could be perfectly grouped as integrative and
36
instrumental orientations, 6 and 5 items respectively, as shown in Table 3.
Table 3
Second Rotated Component Matrix for Goal Intentions in the Pilot Studya
Item No.
Component
Integrative Instrumental G12 To appreciate English art and literature .852 -.018 G10 To meet and converse with more and varied people .850 .095
G15 To take part in cultural activities .841 .061
G6 To better understand English people and culture .820 .039 G16 To be more at ease with fellow English speakers .792 .159
G2 To better understand the world .765 .148
G5 To be a more qualified job candidate .149 .892
G4 To find a decent job .203 .887
G7 I will need it for my future career. .156 .846
G9 It will assist me to apply for a better major .074 .778
G11 To achieve an “A” in the class -.101 .578
a. Rotation converged in 3 iterations.
Total Variance Explained: 67.63%
As for the items measuring implementation intentions, five components were extracted in the beginning. The items were grouped based on the loadings. Item 18 (designed as situational orientation) and item 7 (designed as content orientation) were deleted due to its different underlying conceptualization with other higher-loaded items for situational orientation, as observed in the first column in Table 4. In addition, there was no common ground among the higher-loaded items for the fifth column at the semantic and theoretical level, and, hence, item 12 , 13 ,19 were deleted.
37
Table 4
First Rotated Component Matrix for Implementation Intentions in the Pilot Studya
Item No.
Component
situational interactive strategic content unknown
I_15 .773 .110 .044 .147 .054
I_16 .748 -.156 .086 .386 .090
I_18 .692 .457 .064 .174 .196
I_10 .691 .244 .167 .209 .202
I_14 .660 .049 .067 .071 .526
I_7 .562 .370 .313 .046 .039
I_9 .471 .052 .281 .430 .180
I_1 .029 .827 .202 .276 .102
I_20 .079 .802 .099 .200 .226
I_17 .465 .693 .153 .075 -.004
I_11 .169 .421 .350 .126 .322
I_6 .140 .126 .878 .134 .006
I_5 .249 .222 .798 .216 .165
I_3 .252 .195 .151 .832 .052
I_4 .272 .360 .245 .717 .111
I_2 .171 .356 -.048 .498 .493
I_12 .176 .150 .119 .452 .403
I_13 .303 -.012 -.052 .100 .695
I_19 .043 .315 .304 .105 .649
I_8 -.037 .147 .558 .078 .592
During the first component analysis, an interesting finding emerged from this process. The second column of the component matrix attracted the most attention because of the high loadings of item 1, 17, and 20. By looking back into the statements of these three target observations, the researcher found an interesting similarity. Even though these three items were initially coded as different orientations,
38
but they were all associated with foreigners regarding the descriptions of the situational cue (the if-component). As the interaction with foreigners becomes more accessible, this newly-found component reflected relevant learning stimuli in contemporary EFL contexts. As Cullhane (2004) introduced Interaction Motivation relating the learner’s focus on intercultural interactions, discussion of the construct in the present investigation pertaining to the communicative and interactive objectives with foreigners would be termed as interactive orientation. The items were thus re-categorized as one new dimension based on their similarity in highlighting intercultural contact. They would be tested in reliability analysis and in the formal study to assure the construct validity.
After eliminating item 7, 12, 13, 18, 19, the remaining 15 items of implementation intention underwent the second component analysis and the categorization was demonstrated in Table 5. Four components were properly extracted from the item pool and the total variance explained reached 69%, with coefficients under 0.2 suppressed. In the rotated matrix, items were grouped together based upon the level of loadings and were shaded. With proper consideration, the researcher keep item 11 as one of the strategic variables despite its lower loadings because its strategic nature based on the original objective of item formation (If there are words I don’t know, I know how to handle them.). The overall modifications for the items were summarized in Table 6.
39
Table 5
Second Rotated Component Matrix for Implementation Intentions in the Pilot Study
Component
situational Interactive content strategic
I_16 If I have to be out,… .813 .314
I_15 If there are opportunities during break time,…
.808
I_14 If there are opportunities after school,…
I_2 If there are opportunities, I intend to learn as many words as
possible.
.387 .635
I_9 If there are opportunities, I try to make an English study plan for myself.
.409 .625 .249
I_6 When studying English, I know how to use appropriate learning techniques.
.871
I_5 When studying English, I know how to maintain my attention.
.245 .227 .210 .805
I_8 If I encounter difficulties when I read English, then I have tactics to solve the problems.
.264 .689
40
Table 6
Summary of Modifications Made from the Pilot Results
Item No.
Content Modifications (reasons)
G1 I intend to learn English to be able to read textbooks, novels, manuals in the original texts.
Deleted (Cross-loading) G3 I intend to learn English to obtain as many English
certificates as possible.
Deleted (Cross-loading) G8 My goal in learning English is because I may have an
opportunity to move to an English-speaking country.
Deleted (Cross-loading) G13 My goal in learning English is to apply for further
studies in foreign countries.
Deleted (Cross-loading) G14 I intend to learn English is to better understand my
favorite stars from English-speaking countries.
Deleted (Component switch)
I_18 If there are opportunities in my spare time, I will think of how to learn English.
Deleted (Cross-loading) I_7 If opportunities occur, I plan to keep a regular diary
to improve my English writing.
Deleted (Cross-loading) I_12 When studying English, I manage my environment in
order to make learning more efficient.
Deleted (Cross-loading) I_13 If I have English class, I will try to concentrate on
learning more effectively.
Deleted (Cross-loading) I_19 If I have to study English, then I will select the
appropriate time to make learning more efficient.
Deleted (Cross-loading) I_1 If there are opportunities to meet and converse with
English-speaking foreigners, then I will think of how to speak with them to improve my oral skills.
Strategic Orientation
Interactive Orientation I_17 If there are more opportunities to meet English
foreigners, I try to think of some topics to speak with them to improve my oral skills.
Content Orientation
Interactive Orientation I_20 If there are opportunities to meet English-speaking
foreigners, I will think of how to use these opportunities to improve my English ability.
Strategic Orientation
Interactive Orientation
Note. G1: item number 1 for measuring goal intentions; I_1: item number 1 for measuring implementation intentions
41
The next step was to conduct an internal consistency reliability analysis to determine the reliability of each subscale. The researcher computed Cronbach’s Alpha coefficients for the good-fit items and then determined the variables to form the most coherent scales. The reliability results of the final subscales were shown in Table 7.
Given that the mean Cronbach’s Alpha coefficients were above 0.85 and all the individual scale coefficients were well above 0.79, it was concluded that the proposed categorization performed well in terms of reliability. Goal intentions and implementation intentions also reached high internal consistency among the selected items, 0.85 and 0.89 respectively. Since the result of the reliability reached the significant level, it could be assumed that the 26 items, including 11 for goal intentions and 15 for implementation intentions, would be retained and renumbered for the formal study, as shown in Appendix I, J, K and L.
Table 7
The Internal Consistency Reliability of Subscales in the Pilot Study
Subscales Item Number Cronbach’s
alpha
Goal Intention (.854)
Integrative Orientation 2, 6, 10, 12, 15, 16 .904
Instrumental Orientation 4, 5, 7, 9 ,11 .855 Implementation
Intention (.896)
Content Orientation 2, 3, 4, 9 .807
Situational Orientation 10, 14, 15, 16 .821
Strategic Orientation 5, 6, 8, 11 .795
Interactive Orientation 1, 17, 20 .808
42
Data Analysis
To measure the relationship between goal intentions/ implementation intentions and self-regulation, the current study applied a multivariate statistics approach- multiple regression- to check the extent to which intentions could affect self-regulation. Specifically, this study would aim to understand which type of learning intentions and orientations has the strongest predictive power over self-regulatory capacity as the two research questions suggest:
1. To what extent can goal intentions affect the demonstration of self-regulatory capacity for high school students in Taiwan?
2. To what extent can implementation intentions affect the demonstration of self-regulatory capacity for high school students in Taiwan?
In order to answer the two research questions which were proposed to measure the effects of intentions in language learning on learners’ self-regulatory capacity, learners’ goal intentions and implementation intentions were adopted as predictor (independent) variables while their self-regulatory capacity was operated as a criterion (dependent) variable. In the current research, the subscales would be compared in terms of their effects on the total score of the dependent variable respectively. In view of the theoretical consideration from Sheeran et al. (2005), the strength and the activation of goal intentions have proved to be complementary of performance/
implementation effects. That is, implementation intentions function upon the basis of goal intentions. Consequently, a hierarchical regression was designed when the independent variables were entered in sequence in terms of the theoretical account.
There were two models in the investigation: (1) subscales of goal intentions entered to the analysis and (2) subscales of implementation intentions added to the first model.
Through the comparison between models, the predictive power and the impact of a
43
certain variable could be revealed. In addition, the relative weight of each independent variable on the dependent variable would be demonstrated through the analysis.
In the current study, the research purpose was to examine the relationship between intentions and self-regulatory capacity. In addition to measuring the effects of intentions via multiple regressions, the correlation among independent variables and the interface between self-regulatory capacity and each of the independent variables were examined to further demonstrate how the variables were related through Pearson Product-Moment Correlation.
44