Again, the results of Kaiser-Meyer-Olkin MSA (.937) and Bartlett’s Test of Sphericity (x2 = 2384,, p =.000) indicated that the data set satis-
 |
 |
fied the assumptions for factorability. Principle Components Analysis was chosen as the method of extraction in order to account for maximum variance in the data using a minimum number of factors. A three-factor solution was extracted with eigenvalues 9.335, 1.286 and 1.146 and was supported by an inspection of the Scree plot. These
3 factors accounted for 69.220% of the variance as shown in Table 3.
The 3 components were rotated using the Varimax procedure and a simple structure was achieved as shown in the Rotated Component Matrix in Table 4.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
f Improved У Communications
The results of the factor analysis on the drivers produced 3 underlying factors — improvement to medical and business efficiency and effectiveness, pressure and improvement to communications (see Table 2). Similarly the results of the factor analysis on the benefits produced 3 underlying factors — improved medical and business efficiency, improved communications and improved medical and business effectiveness (see Table 4). A simple model was developed using the combined male and female data (see Figure 3). Figure 3 provided all possible associations between driving force factors and benefit factors for ICT adoption. It was now appropriate to determine whether those associations differed between male and female GPs. The data was subdivided into male and female GP responses and the two sets of data formed the basis for testing the model. The model was tested using partial least squares (PLS) with PLSGraph. PLS is a combination of principal components analysis, path analysis and regression. PLS offers a number of advantages. It is suitable for exploratory studies (Chin 1998, Gefen et al 2000), it has minimal requirements on sample size and residual distribution (Gefen et al 2000) and it is an appropriate procedure for small response levels, other meth
ods requiring greater than 200 responses (Lai 2004). The results can be seen in Figures 4 and 5 and Tables 5 & 6.
An examination of Figure 4 and Table 5 shows that those respondents that placed a higher priority on improvement to medical and business efficiency and effectiveness noted a higher level of benefit in terms of improvement to efficiency, improvement to effectiveness and improvement to communications. The data in Figure 4 shows that placing a higher priority on improvement to communications or pressure to adopt ICTs did not significantly alter the perception of any of the three groups of benefits.
In the PLS analysis, the square roots of the Average Variance Analysis (AVE) values for all constructs are higher than the correlations between constructs and the composite reliability values are above 0.70 (Gefen et al 2000). These results indicate good convergent and discriminant validity and reliability.
An examination of Figure 5 and Table 6 shows that those female respondents that placed a higher priority on improvement to medical and business efficiency and effectiveness or improvement to communications saw no significant differences in the perception of any of the three groups of
Figure 4. Partial Least Squares Model of Drivers and Benefits (Males)
|
1.296* (2.31) |
|
0.973* (1.79) |
|
.218* (2.50) |
|
|
||
|
|||
|
|||
|
|||
|
|||
|
|||
|
|||
|
|||
|
|||
|
|||
|
benefits. However, those respondents that had adopted ICT primarily through pressure did show a significantly positive difference in the perception of all three benefit groups.
Again, in the PLS analysis, the square roots of the Average Variance Analysis (AVE) values for all constructs are higher than the correlations between constructs and the composite reliability values are above 0.70 (Gefen et al 2000). These results indicate good convergent and discriminant validity and reliability.