順天堂グローバル教養論集第一巻20160325
29/158

25Motivational self-regulatory capacity in L2 writingscales. The decision to delete the items was made (a) when the correlation between the item and the re-spective subscale was below .40 and (b) if the dele-tion of that item increased Cronbach’s alpha for the respective subscale. Five items (7, 14, 23, 26, and 38) failed both criteria. These ve items were subse-quently deleted. To further reduce the number of items, I reviewed the contents of all items (including the deleted ones) by returning to the item pool. Three of the deleted items (38, 7, and 14) asked participants whether they would seek for others’ help in respective situations; two other items (20, 30) were also about others’ help. In addition, I noticed two other deleted items (23 and 26) because of the participants’ weak dis-crimination when they were asked whether they use specic self-motivational strategies. Six additional items (3, 8, 9, 15, 17, 40) fell in the criteria. A deci-sion was made to exclude these eight items. With the remaining 27 items, the internal reliabil-ity analysis was conducted for each subscale by computing Cronbach’s Alpha coefcients. The item reductions continued in order to create the most co-herent scales with higher internal consistencies. Seven items (2, 5, 12, 13, 25, 33, and 25) either asked the participants about two different aspects of their strategy use or were complexly worded with more than two phrases. As these items might have confused the participants, I considered deleting them by carefully examining each item’s relation-ship to the respective subscales. By following Tseng et al. (2006), I decided to re-tain four items per subscale, resulting in a total of 20 items for the nal version of the motivational SRC scale for academic essay writing (see Table 3 in the later section). Table 2 presents the reliability of ve subscales of motivational SRC in academic essay writing. The mean Cronbach’s Alpha coef-cient was .83 and all of the individual subscale coef-cients were above .70. 6. Preliminary exploration of motivational SRC 6.1. Uni-dimensionality of the scale Although not initially planned, I explored the uni-dimensionality of the nal scale by performing Ex-ploratory Factor Analysis through Principle Axis Factoring (PAF) on the remaining 20 items. The al-pha level was set at p < .05. A sizable number of correlations were higher than r = .32 (Tabachnick & Fidell, 1996). Though the PAF reported four factors loaded with eigenvalues greater that 1, the rst fac-tor explained 53.29 % of the item variance with ei-genvalue of 10.66, opposing to 1.34 for the second largest factor. This indicates that the uni-dimension-ality of the scale measured by the 20 items and what those 20 items measure would be interpreted as the motivational SRC in academic essay writing. Table 3 shows factor loadings on the rst unrotated factor of each item in each subscale. 6.2. Motivational SRC among former stu-dents To determine an overall tendency of motivational SRC among the students who completed the EAP program, I computed mans of subscale scores as subscales of ve motivational SRC constructs in ac-Table 2. Internal Consistency Reliability of the Subscales in the Second Study Phase Self-regulatory Capacity Remained ItemsCronbach’s AlphaCommitment  Control10, 16, 21, 24.80Metacognitive  Control1, 11, 18, 34 .84Satiation Control28, 29, 36, 37.88Emotion Control4, 6, 19, 32.79Environment  Control 22, 27, 31, 39.87

元のページ 

10秒後に元のページに移動します

※このページを正しく表示するにはFlashPlayer10.2以上が必要です