Structural Equation Modeling
As the current study involves partial least square–structural equation modeling, further analysis was conducted based on two sections, including inner-model and outer-model analyses [
45]. First, the validity and reliability of the measurement tool was tested using Confirmatory Factor Analysis (CFA).
Table 3 represents the results of CFA. As shown, loading values of Public Relations Campaigns (PRC1, PRC2, PRC3, and PRC4) were relatively high, ranging from 0.725 to 0.798, indicating a robust consistency between the constructs. The AVE value of 0.577 implies that the Public Relations Campaigns explain 57.7% of the variance in the observed variables, indicating sufficient convergent validity. The high CR 0.758 and CA 0.845 values further establish the construct reliability.
For Environmental Education and Awareness, items EDU1, EDU2, EDU3, and EDU4 show strong loadings, ranging from 0.714 to 0.787, demonstrating a strong internal consistency. The AVE value of 0.556 further implies that the construct explains 55.6% of the variance in the observed variables, showing satisfactory convergent validity. In addition, the high CR 0.744 and CA 0.834 values indicate adequate construct reliability.
Regarding the Attitudes Towards Environmental Sustainability construct, items ATT1, ATT2, ATT3, and ATT4 show loadings ranging from 0.213 to 0.878. While the loading of ATT4 is somewhat low, the other items indicate strong loadings, suggesting a generally strong internal consistency. The AVE value of 0.527 indicates that the construct explains 52.7% of the variance in the observed variables. Meanwhile, the 0.722 CR and 0.765 CA values confirm the construct reliability.
Finally, the items regarding Behavioral Intention (INT1, INT2, INT3, and INT4) show loadings ranging from 0.026 to 0.937. While the loading of INT1 is low, the other items show strong loadings. The AVE value of 0.522 implies that the construct explains 52.2% of the variance in the observed variables. The high values for CR, 0.726, and CA, 0.742, establish the construct reliability.
The acquired values for the SEM fit indices exhibit the adequacy of the model fit in this study [
46]. The Standardized Root Mean Square Residual (SRMR) value of 0.074 falls below the threshold of 0.08, suggesting an acceptable fit between the proposed model and the observed data. The Tucker–Lewis Index (TLI) value of 2.150 surpasses the threshold of 0.90, indicating a good fit of the model to the data. The Chi-square value of 1.041 is below the threshold of 3.0, indicating a non-significant difference between the model-implied covariance matrix and the observed covariance matrix. In addition, the Normed Fit Index (NFI) value of 0.926 falls within the range of 0 to 1, explaining the variance of the observed data. Altogether, the acquired values indicate a generally acceptable fit of the SEM model to the data.
Table 3 represents the findings of goodness of fit.
A number of corrective actions were taken to improve the quality of the model. These included deleting items with low factor loadings, re-examining the relationships between variables based on modification indices, and simplifying the model structure to align with the theoretical framework. Data quality and consistency with the underlying statistical hypotheses were also confirmed. After applying these modifications, the model was re-estimated, and improved fit indices emerged, enhancing the model’s reliability and the validity of the results obtained.
The discriminant validity is assessed using a two-step approach, including the Fornell–Larker criterion and the Heterotrait–Monotrait ratio scale [
47]. As shown in
Table 4, the diagonal values describe each construct’s square roots of the AVEs. It is found that all diagonal values are more significant than the off-diagonal values, suggesting discriminant validity among the constructs. Further, the HTMT values should be less than 1 to confirm discriminant validity. Here, the HTMT values for all constructs are less than 1, indicating acceptable discriminant validity among the constructs in the model.
Table 5 provides the summary of Heterotrait–Monotrait ratio.
The (HTMT) criterion was used to assess the discriminant validity of the latent variables in the conceptual model. The HTMT value is an important indicator to ensure that each variable measures a distinct dimension, independent of the other variables. Based on accepted values in the literature (less than 0.85 or 0.90), all bivariate correlations between the variables fell within the acceptable range, indicating strong discriminant validity.
The HTMT value between “Behavioral Intention” and “Attitudes” was 0.510, a moderate value indicating a moderate conceptual relationship, but still within the permissible limit. The correlation between “Education and Awareness” and “Behavioral Intention” was 0.078. Similarly, the correlation between “Media Campaigns” and “Attitudes” (0.362) and its relationship with “Behavioral Intention” (0.512) also strengthens the discriminant validity.
These results indicate that the variables in the model do not suffer from conceptual overlap, and that each measures a unique theoretical dimension, which enhances the accuracy of the results obtained in the structural analysis of the model. This is also confirmed by
Table 6.
Before conducting the path analysis, the predictive power of the independent construct was examined [
48]. As shown in
Table 7, Attitudes show that the Public Relations Campaigns explain 29.6% of the variance. For Behavioral Intention, Education, and Awareness, Public Relations Campaigns show 18.8% and 19.6% of the variances, respectively. Overall, moderate variance is explained by the Public Relations Campaigns. The obtained R-square values imply that the independent variable collectively explains moderate variability in the dependent variables.
Finally, the path analysis as shown in
Figure 2 is conducted to test the proposed effects of Public Relations Campaigns on Environmental Education and Awareness, Attitudes Towards Environmental Sustainability, and Behavioral Intention. As indicated in
Table 8, Public Relations Campaigns significantly affect Environmental Education and Awareness (β = 0.443,
p < 0.001), implying that Public Relations Campaigns affect environmental education and awareness. This means that the first hypothesis was confirmed. Concerning the effect of Public Relations Campaigns on Attitudes Towards Environmental Sustainability, the obtained values β = 0.544,
p < 0.001 suggest that Public Relations Campaigns significantly affect attitudes towards environmental sustainability. Finally, it is found that Public Relations Campaigns significantly affect Behavioral Intention (β = 0.433,
p < 0.001), indicating a robust influence of Public Relations Campaigns on behavioral intention toward environmental sustainability. Altogether, these results supported assuming the positive effects of Public Relations Campaigns in the current research study.
The second hypothesis, “Public Relations Campaigns positively affect attitudes towards environmental sustainability among the university students in the UAE and Egypt”, also remained supported. Study respondents agreed that Public Relations Campaigns are critical in shaping positive attitudes toward environmental sustainability and preservation efforts. They noted that the effect of Public Relations Campaigns was observed in the public’s adoption of environmentally sustainable behaviors. In addition, respondents expressed that Public Relations Campaigns encourage a culture of environmental accountability and awareness, leading to augmented support for sustainability initiatives. The information shared through Public Relations Campaigns perceived positively affected environmental sustainability attitudes among the respondents. These findings are also consistent with the argumentation in the current literature [
49,
50]. For example, Barter argues that Public Relations Campaigns aim to lead thought processes and actions toward a distinct goal. How individuals internally process the communicated content affects changes in their attitudes, beliefs, and intentions. This inner processing of communication has been a preceding area of interest for social cognition and communication research scholars [
51].
Finally, regarding the third hypothesis, “Public Relations Campaigns positively affect behavioral intentions among the university students in the UAE and Egypt”, the respondents agreed that Public Relations Campaigns significantly affect their intentions to engage in environmentally friendly behaviors, i.e., recycling and reducing energy consumption. They also revealed that the information provided through Public Relations Campaigns positively influences their choices for adopting sustainable lifestyle choices. In addition, respondents expressed that Public Relations Campaigns promote translating their positive perspectives toward environmental sustainability into concrete actions. They reported that orientation to Public Relations Campaigns enhances their participation in pro-environmental activities, reflecting intentions to contribute to environmental conservation efforts. Thus, these results are consistent with the proposition to run Public Relations Campaigns for environmental sustainability and preservation. As Megantari noted, environmental communication is a rudimentary tool for promoting practical efforts to improve human understanding of the environment and human interactions with nature [
52]. It serves as a medium through which environmental issues are created and different societal responses are negotiated [
7].