Figure 1: Theoretical framework conceptualising ICT implementation in schools
| Australasian Journal of Educational Technology 2011, 27(2), 361-379. |
AJET 27 |
Framing ICT implementation in a context of educational change: A structural equation modelling analysis
Emily M. L. Wong
Vine Education Consultancy, Hong Kong
Sandy C. Li
Hong Kong Baptist University
Despite the common belief that information and communication technology (ICT) has the potential to support certain fundamental changes in learning, few have examined ICT implementation conceptually within a wider context of educational change. Methodologically, we are by and large limited to building simple models that accommodate only a single dependence relationship among variables. Framing ICT implementation as a process of interactions among pedagogical and organisational factors in bringing about changes in student learning, this article used data collected from 1076 teachers in 130 schools to construct a structural equation model (SEM), from which we are able to examine multiple interrelated dependence relationships in a single model. Results indicated that from teacher perspectives, the collegial capacity of ICT implementation strategies played a central and mediating role in effecting changes in student learning, of moving away from a teacher-centred approach to one that is more student-centred. Specifically, ICT brought about these changes in the context of establishing collegiality in fostering pedagogical innovations in schools. Implications for both researchers and practitioners are discussed.
Guided by the theoretical underpinning that implementation is relevant to the interaction of implementation strategies, the changes achieved, and also the factors that influence these changes (Levin, 2001), we have proposed a theoretical model comprising ICT implementation strategies, perceived changes in student learning (outcome variable), and some key contextual factors. With the emerging trend of defining learning outcomes in terms of understanding and performance (Fullan, 1996), constructivism becomes the common lens that people adopt in examining the effectiveness of using ICT in teaching and learning. Based on the understanding that constructivism stresses the importance of each individual's autonomy as a thinker and the importance of the social context of learning, and thereby emphasising the importance of the learner's initiative (Tiene & Ingram, 2001), this paper has conceptualised perceived changes in pedagogy and learning as changes from a teacher-centred approach to a student-centred approach over a period of two years' time, which is parallel to changes in classroom practices from a traditional approach to a constructivist approach.
Figure 1: Theoretical framework conceptualising ICT implementation in schools
| Demographic | Frequency | Percent (%) | |
| Age | 18-25 | 134 | 12.5 |
| 26-30 | 309 | 28.8 | |
| 31-40 | 380 | 35.4 | |
| 41-50 | 166 | 15.5 | |
| Over 50 | 84 | 7.8 | |
| Total | 1073 | 100 | |
| Teaching experience | 0-3 years | 195 | 18.1 |
| 4-10 years | 421 | 39.1 | |
| 11-15 years | 209 | 19.4 | |
| 16-20 years | 97 | 9.0 | |
| Over 20 years | 154 | 14.3 | |
| Total | 1076 | 100 | |
| Gender | Male | 411 | 38.2 |
| Female | 664 | 61.8 | |
| Total | 1075 | 100 | |
The survey instrument for studying ICT implementation in schools consisted of the following six scales: (1) perceived transformational leadership, (2) perceived climate for collaboration and experimentation, (3) collegial capacity of school's ICT implementation strategies, (4) perceived changes in teacher pedagogy, (5) perceived changes in student learning, and (6) perceived effectiveness of government ICT policy. On average, each of the six scales was made up of about 6 items, with a total of 39 items (ultimately reduced to 33 items) on the whole, as shown in Table 3.
All items were measured on a 4-point Likert scale ranging from "strongly disagree" to "strongly agree". This type of ordinal scales measuring the intensity of feeling toward the item generates more information than dichotomous scoring, and it facilitates statistical analysis by more faithfully reflecting the individual differences on the attribute (Nunnally & Bernstein, 1994). On the other hand, statistical analyses of ordinal scales are usually based on a Pearson product-moment correlation (PPM) matrix that assumes the input data as variables of continuous scale. In fact, variables based on ordinal response modes from Likert scale measurements actually depart from the representation of a continuous scale (Joreskog & Sorbom, 2001). To overcome the limitation, the present study used a polychoric correlation matrix as the input matrix for confirmatory factor analysis and structural equation modelling (Flora & Curran, 2004; Joreskog & Sorbom, 2001). Assuming that there is a continuous variable underlying each ordinal variable, polychoric correlations are not correlations computed from actual scores but are rather estimated theoretical correlations of the underlying continuous variables (Joreskog & Sorbom, 2001). Regarding the issue of estimation methods in SEM, diagonally weighted least squares (DWLS) was used as it does not make assumptions about the distribution of the observed variables (Diamantopoulos & Siguaw, 2000; Flora & Curran, 2004), and it was shown to perform better than WLS in small sample sizes in terms of chi-square statistics, parameter estimates, and standard errors.
Despite this, the variance extracted against squared correlation test as suggested by Fornell and Larcker (1981) found that "leadership against climate" as well as "changes in pedagogy against changes in learning" were not clearly separable from one another. The average variance extracted for these two pairs of constructs was smaller than their corresponding correlation (Table 4). On such considerations, modifications were subsequently made to the constructs for leadership, climate, changes in pedagogy, and changes in learning. With a total deletion of 9 items, 3 of which were deleted after conducting EFA, and 6 of which were deleted after conducting discriminant validity checks, the modified measure subsequently contained 33 items (see Table 3).
| Constructs | Indicators | Factor loading | t-value | Composite reliability | Variance extracted |
| Leadership | Ld1 | .82(.82) | 24.02(23.45) | .89(.89) | .53(.54) |
| Ld2 | .81(---) | 25.82(---) | |||
| Ld3 | .74(.74) | 26.98(26.29) | |||
| Ld4 | .56(.58) | 13.13(14.26) | |||
| Ld5 | .61(.62) | 17.84(18.18) | |||
| Ld6 | .80(.80) | 35.48(32.18) | |||
| Ld7 | .73(.75) | 25.91(25.67) | |||
| Cm1/ Ld8 | .80(.79) | 34.58(35.46) | |||
| Climate | Cm2 | .83(---) | 40.02(---) | .85(.85) | .54(.59) |
| Cm3 | .75(.84) | 27.11(30.78) | |||
| Cm4 | .65(.73) | 21.28(22.05) | |||
| Cm5 | .64(.72) | 18.24(18.78) | |||
| Government ICT policy | Gp1 | .78(.78) | 26.65(27.37) | .92(.92) | .58(.58) |
| Gp2 | .83(.83) | 34.44(35.80) | |||
| Gp3 | .68(.69) | 18.60(19.39) | |||
| Gp4 | .66(.68) | 19.03(21.15) | |||
| Gp5 | .85(.85) | 34.77(36.47) | |||
| Gp6 | .68(.70) | 18.07(19.25) | |||
| Gp7 | .81(.80) | 29.36(29.45) | |||
| Gp8 | .78(.77) | 24.52(24.97) | |||
| ICT implement strategies | Is1 | .66(.63) | 17.05(16.23) | .83(.83) | .45(.45) |
| Is3 | .64(.64) | 18.19(18.03) | |||
| Is5 | .71(.68) | 19.45(18.60) | |||
| Is6 | .65(.64) | 18.43(18.50) | |||
| Is7 | .69(.72) | 17.90(20.31) | |||
| Is8 | .69(.72) | 18.85(22.22) | |||
| Changes in pedagogy | Pd1 | .66(---) | 15.62(---) | .79(.78) | .38(.54) |
| Pd2 | .58(---) | 14.02(---) | |||
| Pd3 | .64(---) | 16.53(---) | |||
| Pd4 | .60(.70) | 15.23(15.61) | |||
| Pd5 | .57(.69) | 13.73(14.04) | |||
| Pd6 | .65(.81) | 19.56(21.41) | |||
| Changes in learning | Ln1 | .79(---) | 33.33(---) | .93(.92) | .64(.66) |
| Ln2 | .87(.87) | 37.55(36.50) | |||
| Ln3 | .83(.86) | 38.01(37.42) | |||
| Ln4 | .77(.77) | 29.15(28.69) | |||
| Ln5 | .72(.72) | 27.53(26.55) | |||
| Ln6 | .81(.83) | 34.56(33.77) | |||
| Ln7 | .80(.81) | 28.86(28.62) |
The modified measure for the most part demonstrated unidimensionality, reliability, convergent validity (Table 2), and discriminant validity (Table 4). Results from the structural equation-based approach of CFA and discriminant validity checks indicated that the values all reached an acceptable level indicating construct validity. However, the Cronbach's alpha for the construct of changes in pedagogy (.68) was slightly below the value of .70 as recommended for newly developed measures (Hinkin, 1998). In addition, the average variance extracted for the construct of ICT implementation strategies (.45) was also slightly below the recommended value of .50 (Fornell & Larcker, 1981). These limitations were, however, compensated by the validity of the scales strongly demonstrated in the dimensions of unidimensionality, composite reliability, convergent validity, and discriminant validity.
| Scale | Item | Questions |
| Leadership | Ld1 | My educational beliefs are reflected in the school goals. |
| Ld2 | I feel comfortable to talk to the principal about school matters. | |
| Ld3 | The principal trusts me with school matters. | |
| Ld4 | Teachers have autonomy to make decisions relevant to their teaching. | |
| Ld5 | The principal encourages me to experiment with new ideas in classroom practice. | |
| Ld6 | My school provides adequate resources to support staff professional development. | |
| Ld7 | The principal values staff professional development. | |
| Climate | Cm1 | I have a sense of belonging to my school. |
| Teachers in our school: | ||
| Cm2 | work in a collegial manner. | |
| Cm3 | are willing to experiment with new ideas in classroom practice. | |
| Government ICT policy | The initiative: | |
| Gp1 | provides adequate professional development for teachers. | |
| Gp2 | provides adequate technical support for schools. | |
| Gp3 | enables schools to have autonomy in allocating or recruiting resources relevant to ICT. | |
| Gp4 | provides adequate hardware facilities for schools. | |
| Gp5 | provides adequate support on curriculum resources for teachers. | |
| Gp6 | provides sound network infrastructure for schools. | |
| Gp7 | enables students to learn more effectively. | |
| Gp8 | helps to strengthen teachers' quality of teaching. | |
| ICT implemen- tation strategies | Is1 | My school has a mechanism to disseminate the experiences of using ICT resources for teaching and learning. |
| Is2 | My school has mobilised resources from external parties (e.g. parents/ alumni/ other schools/organisations) to help to implement ICT in teaching and learning. | |
| Is3 | Teachers' opinions can be conveyed to ICT policy-making bodies effectively within school. | |
| Is4 | Colleagues in my school exchange experiences of using ICT to enhance teaching and learning. | |
| I participate in sharing sessions for exchanging experiences (of using ICT to enhance teaching and learning) with: | ||
| Is5 | teachers from other schools. | |
| Is6 | educators from tertiary institutions. | |
| Changes in pedagogy | Compared to the past two academic years, I am: | |
| Pd1 | creating more opportunities for discussions to develop students' expressive and analytical abilities. | |
| Pd2 | encouraging students to explore and to inquire in learning more. | |
| Pd3 | providing more opportunities for students to determine their learning activities. | |
| Changes in learning | Compared to students of the same level in the past two academic years, my students: | |
| Ln1 | are more independent in their learning. | |
| Ln2 | are more active in constructing knowledge. | |
| Ln3 | are able to make better use of collaborative work to facilitate learning. | |
| Ln4 | have more courage to express ideas in class. | |
| Ln5 | are more motivated in their learning. | |
| Ln6 | enjoy learning more. | |
| Construct | Square of correlation (average variance extracted for the pair of constructs) | ||||
| Climate | Govt. ICT policy | ICT strategies | Changes in pedagogy | Changes in learning | |
| Leadership | .79 (.53, .54) .52 (.54, .61)* | .073 (.53, .59) .062 (.54, .58)* | .12 (.53, .44) .12 (.54, .44)* | .18 (.53, .39) .10 (.54, .56)* | .18 (.53, .64) .16 (.54, .66)* |
| Climate | .079 (.55, .59) .072 (.59, .59)* | .15 (.55, .45) .11 (.58, .47)* | .21 (.55, .39) .08 (.59, .56)* | .23 (.55, .65) .16 (.58, .66)* | |
| Govt. ICT policy | .09 (.59, .44) | .084 (.59, .39) .049 (.59, .55)* | .10 (.59, .65) .10 (.59, .66)* | ||
| ICT strategies | .24 (.44, .39) .19 (.47, .54)* | .19 (.44, .64) .18 (.47, .66)* | |||
| Changes in pedagogy | .56 (.39, .64) .39 (.55, .66)* | ||||
| * Values for the modified measure | |||||
In evaluating the alternative models, consideration was given to three aspects: (1) the overall fit measures based on a number of fit indices, namely root mean square error of approximation (RMSEA) (Browne & Cudeck, 1993), non-normed fit index (NNFI), comparative fit index (CFI) (Marsh, Balla, & Hau, 1996); (2) the component fit measures came from parameter estimates (Bollen, 1989), which included the squared multiple correlation (R2) for each pair of relationships and the t-value of the path coefficients; and (3) the model parsimony based on fit indices of the Akaike Information Criterion (AIC), the Consistent Akaike Information Criterion (CAIC), and the Expected Cross-Validation Index (ECVI) (Joreskog, 1993). The essence of the indices was to examine measures of fit that took the number of parameters in the model into account, thereby the fit of model was not necessarily improved as parameters were added to the model (Joreskog, 1993).
Before proceeding to model comparisons, overall fit indices of individual models and component fit measures (parameter estimates and R2) were examined first to check whether any models would be rejected. In other words, overall fit and component fit evaluations were supposed to serve as a preliminary step to screen out unacceptable models. The acceptable models in terms of overall fit and component fit were then compared and evaluated in terms of parsimony indices that take fit as well as parsimony (in the sense of number of parameters) into account.
Figure 2: The basic model examining the pattern of relationship
among the constructs in ICT implementation
Non-mediating model
The non-mediating model was supposed to test against the basic model and other competitive models that were constructed on the basis of theoretical justifications, which supported the mediating effect of organisational variables on changes in student learning. Therefore, in the non-mediating model, all contextual variables were simply proposed as directly related to the outcome variable "changes in learning" (Figure 3).
Climatic model
The climatic model represented the hypothesis that the climate for collaboration and experimentation mediated the effect of organisational interventions on changes in classroom practices (pedagogy and learning). As a competitive model based on alternative hypothesis, it was proposed that a climate for collaboration and experimentation played a central role in the social ecology of ICT implementation by mediating the effect of other organisational interventions on changes in pedagogy and learning. Specifically, "school climate" was proposed as having a direct effect on "changes in learning", and an indirect effect through "changes in pedagogy" on "changes in learning". At the same time, "school climate" was proposed as a variable that mediated the effect of "leadership", "ICT implementation strategies" and "government ICT policy" on "changes in learning" (Figure 4).
Figure 3: The non-mediating model examining the pattern of relationships
among the constructs in ICT implementation.
Figure 4: The climatic model examining the pattern of relationships
among the constructs in ICT implementation
External model
The external model represented the hypothesis that the external input in the form of government ICT policy mediated the effect of organisational interventions on changes in classroom practices (pedagogy and learning). As a competitive model based on alternative hypothesis, it was proposed that the government ICT policy played a central role in ICT implementation by mediating the effect of other organisational interventions on changes in pedagogy and learning. Specifically, "government ICT policy" was proposed as having a direct effect on "changes in learning", and an indirect effect through "changes in pedagogy" on "changes in learning". At the same time, "government ICT policy" was proposed as a variable that mediated the effect of "climate" and "ICT implementation strategies" on "changes in learning". Taking into account the essential role of transformational leadership on shaping school climate, "leadership" was proposed as an exogenous variable that related only to "school climate" (Figure 5).
Figure 5: The external model examining the pattern of relationships
among the constructs in ICT implementation
Pedagogical model
The pedagogical model represented the hypothesis that pedagogical intervention mediated the effect of organisational interventions on changes in learning. It has become clear through research evidence that instructional strategy underlying the use of ICT determines learning effectiveness (Honey et al., 2000; McCombs, 2000; Means, 1994; Mehlinger, 1995). As a competitive model based on alternative hypothesis, it was proposed that in the social ecology of ICT implementation, the effect of organisational interventions on changes in learning was mediated entirely by pedagogical intervention. Specifically, "changes in pedagogy" was proposed as the variable that mediated the effect of "climate", "ICT implementation strategies", and "government ICT policy" on "changes in learning". Taking into account the essential role of transformational leadership on shaping school climate, "leadership" was proposed as an exogenous variable that related only to "school climate" (Figure 6).
Figure 6: The pedagogical model examining the pattern of relationships
among the constructs in ICT implementation
| 1 | 2 | 3 | 4 | 5 | 6 | ||
| 1. | Leadership | 1.00 | |||||
| 2. | Climate | .73 | 1.00 | ||||
| 3. | Govt ICT policy | .25 | .27 | 1.00 | |||
| 4. | ICT implementation strategies | .37 | .37 | .31 | 1.00 | ||
| 5. | Changes in pedagogy | .31 | .30 | .23 | .43 | 1.00 | |
| 6. | Changes in learning | .41 | .44 | .32 | .43 | .62 | 1.00 |
| Goodness of fit index | Recommended good fit values | Basic model | Non-mediating model | Climatic model | External model | Pedagogical model |
| Satorra-Bentler chi-square | --- | 1597.98 | 1636.31 | 1689.54 | 1769.46 | 1714.38 |
| Degree of freedom | --- | 488 | 480 | 486 | 488 | 487 |
| Chi-square per degree of freedom | <3.0 | 3.27 | 3.41 | 3.48 | 3.63 | 3.52 |
| P-value for chi-square test | >.05 | .00 | .00 | .00 | .00 | .00 |
| RMSEA | <.06 | .045 | .047 | .047 | .049 | .048 |
| P-value for test of close fit, RMSEA <.05 | >.05 | 1.00 | .99 | .96 | .79 | .93 |
| NNFI | >.95 | .97 | .98 | .97 | .94 | .97 |
| CFI | >.95 | .97 | .98 | .97 | .95 | .98 |
| ECVI | Smaller value | 1.58 | 1.63 | 1.67 | 1.74 | 1.69 |
| AIC | Smaller value | 1743.98 | 1798.31 | 1839.54 | 1915.46 | 1862.38 |
| CAIC | Smaller value | 2182.54 | 2284.92 | 2290.11 | 2354.02 | 2306.94 |
Component fit evaluation
Results of the parameter estimates were shown in Figures 2 to Figure 6, and Table 7. Upon examination, it can be seen from Table 7 that the non-mediating model had some unreasonable results. The magnitude of the path coefficients from "leadership" to "changes in learning" was very low (.07) with t-value (.89) less than the minimum value of 1.96, indicating that the corresponding parameter was not statistically significantly at the 5% level. In addition, specific R2 for the external model was relatively low, which indicated that the independent variables explained a small portion of variance in the endogenous variables concerned (Bentler, 2007, Bollen, 1989; Diamantopoulos & Siguaw, 2000; Goffin, 2007, Markland, 2009). Specifically, the effect of "government ICT policy" explained only 18% of variance in "changes in pedagogy". Overall, results of parameter estimates supported the rejection of the non-mediating model and the external model.
Model parsimony evaluation
Before proceeding to model parsimony evaluation, the overall fit and component fit evaluations already screened out the external model and the non-mediating model. The other three models (basic model, climatic model, pedagogical model) had a reasonably good fit of model data in terms of RMSEA (<.06), NNFI (>.95), and CFI (>.95) (Bentler, 2007; Hu & Bentler, 1999). Magnitude and t-values of path coefficients (>1.96) as well as R2 (ranging from .23 to .75) for the models were also within the acceptable range (Bollen, 1989; Diamantopoulos & Siguaw, 2000, Markland, 2009). The three models were further compared with reference to selected parsimony indices (ECVI, AIC, and CAIC) that took fit as well as parsimony (in the sense of number of parameters) into account. From Table 6, it can be seen that the values of ECVI (1.58), AIC (1743.98), and CAIC (2182.54) all favoured the basic model, which had the lowest values among the models (Joreskog, 1993). Nevertheless, its magnitude of difference in terms of parsimony indices over the other models was not very significant. As the three competing models were supposed to be approximations to reality, it was worthwhile to investigate similarities among the seemingly different approximations to reality instead of finding the most plausible models.
| Model | Structural equations | R2 | Unreasonable parameter estimates |
| Basic (strategic) | climate = .79 leadership | .64 | --- |
| strategies = .37 climate + .22 policy | .45 | --- | |
| pedagogy = .67 strategies | .29 | --- | |
| learning = .49 pedagogy + .67 strategies | .52 | --- | |
| Non-mediating | learning = .06 leadership + .19 climate + .12 policy + .13 strategies + .60 pedagogy | .48 | Leadership -> learning .06 (t value = .90) |
| Climatic | climate = .50 leadership + .40 strategies + .16 policy | .75 | --- |
| pedagogy = .49 climate | .23 | --- | |
| learning = .54 pedagogy + .48 climate | .50 | --- | |
| External | climate = .77 leadership | .58 | --- |
| policy = .58 climate + .43 strategies | .32 | --- | |
| pedagogy = .43 policy | .18 | "Government ICT policy" explained only 18% of variance in "changes in pedagogy" | |
| learning = .59 pedagogy + .44 policy | .48 | --- | |
| Pedagogical | climate = .81 leadership | .61 | --- |
| pedagogy = .23 climate + .31 strategies + .11 policy | .41 | --- | |
| learn = 1.36 pedagogy | .73 | --- |
The rejection of the external model also gave us insights into the dynamics of organisational interventions. It highlighted the central role of internal organisational interventions, in comparison with external policy, in affecting changes in student learning. Notably, the collegial capacity of school's ICT implementation strategies as perceived by teachers was identified as one of the possible focal points of linking up the organisational interventions to the outcome variable "changes in student learning". Specifically, this study has shown that ICT was able to act as a lever to bring about changes in student learning in the context of establishing collegiality in fostering pedagogical innovations in schools (Figure 2, the basic model).
While other organisational interventions were shown to have no direct impact on the outcome variable, they were capable of exerting indirect influences by shaping the collegial capacity of school's ICT implementation strategies. A transformational leadership had no direct effect on classroom practices (changes in pedagogy and student learning), yet it exerted an indirect effect through shaping the school climate. The study has offered insights into the controversy over the relative importance of changes related to classroom practices and changes related to organisational factors. It has given empirical support to the proposition that attention to organisational changes is essential to changes in classroom practices (Leithwood, 1994).
From the empirical results, schools are advised to place more emphasis on enhancing the collegial exchange and sharing of ICT experiences in order to harness the power of ICT in shifting learning from a teacher-centred to a student-centred approach. The significant path coefficient from changes in pedagogy to changes in learning has clear implication on the crucial role of teachers' pedagogical approach in students' learning outcome. It supported the postulation that ICT implementation strategies will exhibit significant effect on changes in learning outcome if there is a change effected by ICT on teachers' pedagogy in classroom practices. The results have been consistent with previous empirical work, which have identified the impact of ICT as highly related to how teachers exploit it efficiently for pedagogical purposes (Balanskat et al., 2006). These empirical findings actually echo the contemporary concepts of educational change, of which ICT is seen as an enabler to reshape the delivery of instruction. As such, greater impact of ICT implementation is achieved if teachers strengthen ICT use with pedagogical strategies.
In addition, empirical evidence from the present study supported that a transformational leadership was highly influential in establishing a cohesive and proactive school climate that enhanced the collegial capacity of ICT implementation strategies. It alerts school administrators and policymakers to the positive effect of cultivating collegiality in schools upon enhancing changes in student learning, which at the same time also offers insights into the appropriate direction of education reform. It is worth pausing to consider whether to adopt a control strategy, or whether to adopt a commitment strategy that seeks to develop innovative working arrangements supporting teachers' decision making, and increasing teachers' engagement in the tasks of teaching (Rowen, 1990). Apart from that, an effective government ICT policy in enhancing professional development for teachers and in providing curriculum and resource support was also very important in facilitating effective ICT implementation in schools, though the impact was not as significant as that exerted by the school climate. In this connection, policymakers are urged to review the impact of adopting private sector practices to tighten control of educational process and the work of professionals, and to consider whether the dominant ideas of economic rationalism and managerialism will really increase or worsen educational outcomes (Mok & Welch, 2002).
Despite the large sample size of over 1076 teachers, the study was possibly limited by the common method variance since the measurement scales were based on a single method of questionnaire survey (Bagozzi, Yi & Phillips, 1991). To avoid the common method variance, future studies may also consider adopting multiple measures obtained with multiple methods. Hence, construct validation can be done with multitrait-multimethod (MTMM) matrix (Bagozzi et al., 1991). The present study aims at examining the pattern of relationships among the constructs rather than about the predictive ability of the constructs, and the analysis is at the teacher level. As pointed out by Kreft and De Leeuw (1998) and Raudenbush and Bryk (2002), if the school is regarded as a unit of study, the school-level variables may also play a role in contributing to the variation of the outcome variable. It is thus worthy to further explore the school effect on ICT implementation through the building of multi-level structural equation models. It would be interesting to find out whether the multi-level results obtained from each individual school are consistent with those in the SEM analysis and how much variation found in the outcome variable can be attributed to different school-level variables.
Bagozzi, R. P., Yi, Y. & Phillips, L. W. (1991). Assessing construct validity in organisational research. Administrative Science Quarterly, 36(3), 421-458.
Balanskat, A., Blamire, R. & Kefala, S. (2006). The ICT Impact Report: A review of studies of ICT impact on schools in Europe. [viewed 11 Apr 2011] http://ec.europa.eu/education/doc/reports/doc/ictimpact.pdf
Bentler, P. M. (2007). On tests and indices for evaluating structural models. Personality and Individual Differences, 42(5), 825-829 (Special issue on Structural Equation Modeling). http://dx.doi.org/10.1016/j.paid.2006.09.024
Bollen, K. A. (1989). Structural equations with latent variables. New York, Chichester: Wiley.
Brown, J. S., Collins, A. & Duguid, P. (1989). Situated cognition and the culture of learning. Educational Researcher, 18(1), 32-42.
Fornell, G. & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39-50.
Fullan, M. (1992). Successful school improvement: The implementation perspective and beyond. Buckingham, Philadelphia: Open University Press.
Fullan, M. (1996). Leadership for change. In K. Leithwood, J. Chapman, D. Corson, P. Hallinger & A. Hart (Eds.), International handbook of educational leadership and administration. Dordrecht, Boston, London: Kluwer Academic Publishers.
Goffin, R. D. (2007). Assessing the adequacy of structural equation models: Golden rules and editorial policies. Personality and Individual Differences, 42(5), 831-839 (Special issue on Structural Equation Modeling). http://dx.doi.org/10.1016/j.paid.2006.09.019
Hair, J. F., Anderson, R. E., Tatham, R. L. & Black, W. C. (1998). Multivariate data analysis. New Jersey: Prentice Hall, Upper Saddle River.
Hinkin, T. R. (1998). A brief tutorial on the development of measures for use in survey questionnaires. Organizational Research Methods, 1(1), 104-121.
Hu, L. T. & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1-55.
Joreskog, K. G. (1993). Testing structural equation models. In J. S. Long (Ed.), Testing structural equation models. Newbury Park, London, New Delhi: Sage.
Joreskog, K. G. & Sorbom, D. (2001). LISREL 8: User's reference guide. Lincolnwood: Scientific Software International.
Kreft, I. & De Leeuw, J. (1998). Introducing multilevel modeling. London, Thousand Oaks: Sage.
Lave, J. & Wenger, E. (1991). Situated learning: Legitimate peripheral participation. Cambridge, New York: Cambridge University Press.
Leithwood, K. (1994). Leadership for school restructuring. Educational Administration Quarterly, 30(4), 498-518.
Levin, B. (2001). Reforming education: From origins to outcomes. New York: RoutledgeFalmer.
MacCallum, R. C. (1995). Model specification: Procedures, strategies, and related issues. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications. Thousand Oaks CA and London: Sage.
Markland, D. (2007). The golden rule is that there are no golden rules: A commentary on Paul Barrett's recommendations for reporting model fit in structural equation modeling. Personality and Individual Differences, 42(5), 851-858 (Special issue on Structural Equation Modeling). doi:10.1016/j.paid.2006.09.023.
Marsh, H. W., Balla, J. R. & Hau, K. T. (1996). An evaluation of incremental fit indices: A clarification of mathematical and empirical properties. In R. E. Schumacker (Ed.), Advanced structural equation modeling. Mahwah, New Jersey: Lawrence Erlbaum Associates.
McCombs, B. L. (2000, 11-12 September). Assessing the role of educational technology in the teaching and learning process: A learner-centred perspective. Paper presented at the Secretary's Conference on Educational Technology 2000: Measuring Impacts and Shaping the Future, Alexandria, VA.
Means, B. (1994). Introduction: Using technology to advance educational goals. In B. Means (Ed.), Technology and education reform. San Francisco: Jossey-Bass Publishers.
Mehlinger, H. D. (1995). School reform in the information age. Bloomington: Center for Excellence in Education, Indiana University.
Mok, J. K. H. & Welch, A. R. (2002). Economic rationalism, managerialism and structural reform in education. In J. K. H. Mok & D. K. K. Chan (Eds.), Globalization and education: The quest for quality education in Hong Kong. Hong Kong: Hong Kong University Press.
Nunnally, J. C. & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw-Hill.
Raudenbush, S. W. & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods. Thousand Oaks, London, New Delhi: Sage.
Rowen, B. (1990). Commitment and control: Alternative strategies for the organizational design of schools. Review of Research in Education, 16, 353-389
Senge, P. M. (1990). The fifth discipline: The art and practice of the learning organization. New York: Doubleday/Currency.
Smylie, M. A., Miretzky, D. & Konkol, P. (2004). Rethinking teacher workforce development: A strategic human resource management perspective. In D. Miretzky (Ed.), Developing the teacher workforce. Chicago: NSSE.
Spillane, J. P., Reiser, B. J. & Todd, R. (2002). Policy implementation and cognition: Reframing and refocusing implementation research. Review of Educational Research, 72(3), 387-431.
Tiene, T. & Ingram, A. (2001). Exploring current issues in educational technology. Boston: McGraw-Hill.
Valli, L., Cooper, D. & Frankes, L. (1997). Professional development schools and equity: A critical analysis of rhetoric and research. Review of Research in Education, 22, 251-304.
Windschitl, M. & Sahl, K. (2002). Tracing teachers' use of technology in a laptop computer school: The interplay of teacher beliefs, social dynamics, and institutional culture. American Educational Research Journal, 39(1), 165-205.
Wong, E. M. L. & Li, S. C. (2008). Framing ICT Implementation in a context of educational change: A multilevel analysis. School Effectiveness and School Improvement, 19(1), 99-120.
Zhao, Y. & Frank, K. A. (2003). Factors affecting technology use in schools: An ecological perspective. American Educational Research Journal, 40(4), 807-840.
| Authors: Emily Mei Ling Wong, Director, Vine Education Consultancy 13/F, Po On Commercial Building, 198 Nathan Road, Hong Kong Email: info@vineedu.com.hk Sandy C. Li (correspondence author), Associate Head and Associate Professor Department of Education Studies, Hong Kong Baptist University Kowloon Tong, Hong Kong. Email: sandyli@hkbu.edu.hk Please cite as: Wong, E. M. L. & Li, S. C. (2011). Framing ICT implementation in a context of educational change: A structural equation modelling analysis. Australasian Journal of Educational Technology, 27(2), 361-379. http://www.ascilite.org.au/ajet/ajet27/wong.html |