Page 707 - ISC PROCEEDINGS 21.4

These indicators confirm that the dataset is suitable for factor analysis. The cumulative
explained variance reaches 62.963%, indicating that the five extracted factors account for
a substantial proportion of the variance in the data. In addition, all factor loadings remain
within acceptable ranges: PEOU from 0.741 to 0.796, PU from 0.730 to 0.784, SI from
0.776 to 0.796, TR from 0.735 to 0.840, and AISE from 0.727 to 0.840.
These EFA results suggest that, after removing SI3, AISE4, AISE5, TR3, and TR4, the
measurement structure becomes more concise and more clearly reflects the five
intended theoretical constructs.
Table 3. Summary of EFA results for the independent variables after refinement

Range of
Factor Observed variables Note
loadings

PEOU PEOU1–PEOU5 0.741–0.796 Retained

PU PU1–PU5 0.730–0.784 Retained

SI SI1, SI2, SI4, SI5 0.776–0.796 SI3 removed

TR TR1, TR2, TR5 0.735–0.840 TR3, TR4 removed

AISE4, AISE5
AISE AISE1–AISE3 0.727–0.840
removed
Source: Compiled from SPSS outputs
4.4. Pearson correlation analysis
Pearson correlation analysis indicates that all explanatory variables are positively
associated with intention to use AI tools for learning, and all relationships are statistically
significant. Specifically, PEOU correlates with IU at 0.488, PU at 0.589, SI at 0.440, AISE at
0.625, and TR at 0.575, with all significance values equal to 0.000. At the bivariate level,
the results therefore suggest that students are more likely to intend to use AI tools for
learning when they perceive the tools as easy to use and useful, receive stronger social
support, feel more confident in using AI, and trust the tool more strongly.
Table 4. Pearson correlations with intention to use AI tools for learning
Variable Correlation with IU Sig.
PEOU 0.488 0.000
PU 0.589 0.000
SI 0.440 0.000
AISE 0.625 0.000
TR 0.575 0.000
Source: Compiled from SPSS outputs.
4.5. Multiple regression analysis and hypothesis testing
Multiple linear regression was performed to estimate the simultaneous effects of
the five explanatory variables on intention to use AI tools for learning. The model is highly
significant, with F = 173.046 and Sig. = 0.000. The adjusted R² equals 0.805, meaning that
the five independent variables explain 80.5% of the variation in IU. The Durbin-Watson
statistic is 1.865, suggesting no serious autocorrelation in the residuals. In addition, VIF
values range from 1.076 to 1.237, far below common warning thresholds, indicating that
multicollinearity is not a concern in the model.

706