روابط ساختاری میان عوامل موثر بر انگیزش زبان آموزان الکترونیکی برای انتقال مهارت
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|30033||2014||8 صفحه PDF||سفارش دهید||5730 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Computers in Human Behavior, Volume 32, March 2014, Pages 335–342
This study investigates the structural relationships among the following factors: e-learners’ internal value, learning usefulness, learning environment, satisfaction, learner achievement, and motivation for skill transfer. To answer the research questions, the researchers administered online surveys to 584 students enrolled in two courses, Conflict Management and Negotiation and Communication Skills, at S Cyber University. According to the results of structural equation modeling, the structural relationships among e-learners’ internal value, learning usefulness, learning environment, learner satisfaction, learner achievement, and motivation for skill transfer were significant and showed positive influence. However, the relationships among learning usefulness, learning environment, learner satisfaction, and learner achievement and those of learning environment, learner satisfaction, and motivation for skill transfer wer
Although cyber universities are recognized as educational organizations of the future, their educational outcomes have not yet been fully studied. Particularly, despite various discussions on the design aspects of educational systems and programs for producing positive educational outcomes, research is scarce on the transfer of cyber university educational outcomes to career fields (Lim, 2009). Recently, transfer of training was used for educational evaluation in the enterprise educational environment. In cyber universities, however, it is difficult for most cyber learners to directly implement their knowledge and skills to their job situation. It is difficult to evaluate learning outcomes because the purpose of cyber universities is to provide higher education to adults who have not previously had the opportunity to attend tertiary institutions, due to either personal or economic reasons (Lim, 2009). Motivation to transfer learned skills has proven to be an important factor predicting learners’ actual behavioral change (transfer) in numerous research studies (e.g., Baldwin and Ford, 1988 and Burke and Hutchins, 2007). It was confirmed that the motivation to transfer occurred prior to the transfer of training (Axtell et al., 1997 and Chiaburu and Lindsay, 2008). Additionally, Gegenfurtner, Veermans, Festner, and Gruber (2009) emphasized the importance of research on the motivation of transfer by mentioning that major interests of human resource development (HRD) theory and practices include training failure, namely, low return on investment due to learners’ low motivation to transfer. As it is confirmed that motivation to transfer is the main variable determining educational effects, in conjunction with learning motivation (Shin & Oh, 2004), measuring educational outcomes of cyber universities through motivation to transfer should provide meaningful insights. The factors affecting transfer or motivation to learning transfer can be classified into three main types: learner characteristics, training design, and external environment. Baldwin and Ford (1988) proposed the transfer process model, in which they presumed that personal factors, training-related factors, and organizational factors affect transfer of learning both directly and indirectly. Neo (1986) argued that positive perception of the organizational environment affects transfer motivation by demonstrating the effects of learning motivation on educational training outcomes. Additionally, Holton (1996) reported that learning, expected usefulness of the training, job attitude, learner satisfaction, and the transfer environment directly affect motivation to transfer by presenting the HRD evaluation research and measurement model. Moreover, Gegenfurtner et al. (2009) divided the factors affecting motivation to transfer as personal, training-related, or organizational. They also extended Baldwin and Ford’s (1988) transfer process model by categorizing the factors as occurring before training, in the middle of training, or after training. However, Neo (1986) pointed out most studies that analyzed learning outcomes according to the learner’s personal characteristics were mainly focused on the learner’s intellectual ability, and research on learner motivation and environment factors remains insufficient. Campbell (1988) and Tannenbaum and Yukl (1992) proposed that the concept of training effects should be extended to the personal variables of trainees and the research should include trainees’ self-efficacy and motivation. Meanwhile, Warr and Bunce (1995) mentioned that learners’ responses to the usefulness of their learning can be effective on three learning design principles (i.e., same element, stimulus variation, general principles). They also pointed out that measuring only learner enjoyment is problematic; the instructor and job-related usefulness of training contents should also be studied as important response measurement estimates (Alliger and Janak, 1989 and Warr and Bunce, 1995). Until now, learner satisfaction has been frequently used to evaluate training results due to measuring convenience. Moreover, as the effects of external environment on motivation to transfer were studied with a focus on the workplace environment (e.g., seniors, colleague support, organizational environment) in previous works (Facteau et al., 1995, Huczynski and Lewis, 1980, Kirwan and Birchall, 2006 and Seyler et al., 1998). It is thought that additional studies focused on the learning environment (e.g., instructor, colleague support, learning atmosphere) should be conducted to provide a more complete picture. Therefore, the current study aims to investigate the effects of internal value as a personal characteristic of learners. It includes internal value as a motivational variable, learning usefulness as learning content variable, and learning environment as an external environmental variable possibly affecting learner achievement, learner satisfaction, and motivation to transfer. We adopt an integrative model and confirm the structural relationships among the variables. Moreover, we identify the effects employment status on learning by investigating the differences in structural relationships among the variables according to learners’ employment status. Cyber universities typically have high proportions of learners who are employed than do traditional universities. Although many previous studies have investigated the effects of having a job on university students’ academic achievement, the results are contradictory. That is to say, some researchers reported that simultaneously holding a job and studying at a university is potentially be harmful to one’s learning (Astin, 1993 and Lammers et al., 2001). Other researchers expressed positive opinions (Dallam and Hoyt, 1981 and Lucas and Lammont, 1998). Furthermore, some proposed that it is not employment status but the difference in distribution of the learner’s time (Dundes & Marx, 2006/2007; Gleason, 1993 and Orszag et al., 2001). The purpose of this study is to examine the effects of internal value, learning usefulness, and learning environment on learner satisfaction, learner achievement, and motivation to transfer of learning to their workplace. Additionally, we will investigate the structural relationships among e-learners’ internal value, learning usefulness, learning environment, achievement and motivation for transfer depending on their employment status. The independent variables are internal value, learning usefulness, and learning environment. The dependent variable is the motivation to transfer of learning. The moderating variables are achievement and satisfaction. The specific research questions are as follows: (1) Do e-learners’ internal value, learning usefulness, and learning environment affect learner satisfaction? (2) Do e-learners’ internal value, learning usefulness, learning environment, and learner satisfaction affect learner achievement? (3) Do e-learners’ internal value, learning usefulness, learning environment, learner satisfaction, and learner achievement affect the motivation for skill transfer?
نتیجه گیری انگلیسی
4.1. Descriptive statistics To confirm the normalization of multivariate distribution of data, we analyzed average, standard deviation, skewness, and Kurtosis. The variables’ means ranged from 3.67 to 53.43, standard deviation ranged from .64 to 3.97, skewness ranged from −.71 to .42, and Kurtosis ranged from −.73 to 6.24. This satisfied the basic assumptions of structural equation modeling examination, as the skewnesses of the measurement variables were less than 2 and their kurtoses were less than 7 (Curran, West, & Finch, 1996). We confirmed that all variables have a significant correlation at the α = .05 level. Since the standard Kurtosis is smaller than 3, and Kurtosis is smaller than 10, all the normalization standards are satisfied ( Kline, 2005). Thus, the current data satisfy the assumption of multivariate normal distribution. As seen in Table 1, the correlation analysis results among variables are as follows. Table 1. Correlation analysis results among variables (n = 584). Variables 1 2 3 4 5 6 7 8 9 10 11 1. Internal Value 1 1 2. Internal Value 2 .78a 1 3. Learning Usefulness 1 .39a .44a 1 4. Learning Usefulness2 .39a .43a .83a 1 5. Learning Environment 1 .51a .55a .62a .63a 1 6. Learning Environment 2 .52a .58a .62a .61a .90a 1 7. Satisfaction1 .50a .56a .52a .53a .63a .67a 1 8. Satisfaction2 .50a .55a .54a .57a .67a .69a .91a 1 9. Achievement .20a .21a .16a .14a .15a .15a .14a .14a 1 10. Transfer Motivation1 .51a .57a .50a .53a .67a .71a .77a .78a .12a 1 11. Transfer Motivation 2 .50a .55a .48a .48a .63a .68a .71a .71a .15a .84a 1 Mean 3.93 4.07 3.86 3.73 3.67 3.81 4.23 4.05 53.43 4.05 3.91 SD .64 .64 .77 .72 .67 .66 .68 .72 3.97 .71 .67 Skewness −.25 −.39 −.22 −.11 .12 .03 −.71 −.48 −1.48 −.46 −.24 Curtoses −.43 −.39 −.73 −.42 −.49 −.54 .04 −.14 6.24 .02 −.31 a p<.05. Table options 4.2. Measurement model Before examining the structural model, we evaluated the fitness of the measurement model by Maximum Likelihood, according to the confirmation procedure of the second level model estimate (Kline, 2005). As seen in Table 2, all fitness indexes of the measurement model seemed desirable. Table 2. Fitness examination results of the measurement model (n = 584). χ2 df TLI CFI RMSEA (90% Confidence Interval) Measurement Model 46.174 25 .993 .996 .038 (.020–.055) If RMSEA is less than .05, the model is close; if less than .08, it is reasonable; and if less than .10, it is poor (Browne & Cudeck, 1993). Table options The standard factor loading index among the paths of measurement variables ranged from .838 to .965, and every path was statistically significant at the α = .05 level. Under the condition that the standard factor loading index should be greater than .30 ( Hair, Anderson, Tatham, & Black, 1992), every measurement variable appeared to properly measure the corresponding latent variable. 4.3. Structural model As the fitness index of the measurement model satisfied the fitness index criteria and the estimate possibility of the structural model was theoretically confirmed, we estimated the fitness of the initial research model. As a result of confirming the fitness index of the initial research model, we were able to confirm the generally good level seen in Table 3. Table 3. Examination results of fitness of structural model (n = 584). χ2 df TLI CFI RMSEA (90% Confidence Level) Initial Research Model 49.615 30 .994 .996 .033 (.015–.050) If RMSEA is less than .05, the model is close; lf less than .08, it is reasonable, less than .10, it is poor (Browne & Cudeck, 1993). Table options Given the hierarchical model relationship between the initial and revised models, we examined the difference in χ2 to confirm the statistical difference. The χ2 test showed the difference in χ2 value between the two models was not statistically significant (Δχ2 (5, n = 584) = 1.155, p = .949), as shown in Table 4. Accordingly, since the revised model is more succinct and generally better fit than the initial model, we selected the revised model as the final research model and re-estimated the fitness and path coefficient of model. Table 4. Fitness examination results of the initial and revised models (n = 584). χ2 df TLI CFI RMSEA (90% Confidence Level) Revised Model 50.770 35 .996 .997 .028 (.006–.044) Initial Research Model 49.615 30 .994 .996 .033 (.015–.050) If RMSEA is less than .05, the model is close; if less than .08, it is reasonable; if less than .10, it is poor (Browne & Cudeck, 1993). Table options We were able to confirm that the fitness index of the revised model is generally good, as shown in Table 5. The results of examining the statistical significance of path coefficient are as shown in Table 4. The relationships among variables according to the structural estimate of the revised model are as follows. Table 5. Structural coefficient estimate of the revised model (n = 584). Paths between variables Un-standardized estimate Standardized estimate Standard deviation t p Satisfaction ← Internal Value .276 .229 .051 5.453⁎ .000 ← Learning Usefulness .452 .439 .055 8.152⁎ .000 ← Learning Environment .208 .206 .049 4.219⁎ .000 Achievement ← Internal Value 1.715 .230 .320 5.365⁎ .000 Motivation to Skill transfer ← Internal Value .146 .115 .045 3.217⁎ .001 ← Learning Usefulness .297 .275 .044 6.686⁎ .000 ← Satisfaction .605 .576 .043 14.213⁎ .000 ⁎ p<.05. Table options First, internal value (t = 5.453, p < .05), learning usefulness (t = 8.152, p < .05), and learning environment (t = 4.219, p < .05) had significant positive effects on learner satisfaction in the following order: learning usefulness (β = .439), internal value (β = .229), and learning environment (β = .206). Second, internal value (t = 3.217, p < .05), learning usefulness (t = 6.686, p < .05), and learner satisfaction (t = 14.213, p < .05) had significant positive effects on motivation to transfer in the following order: satisfaction (β = .576), learning usefulness (β = .275), and internal value (β = .115). Fig. 2 shows the path coefficient estimate of revised model as follows. Full-size image (69 K) Fig. 2. Path coefficient estimate of revised model.