The study provides valuable insights into the factors influencing solar power technology adoption, particularly emphasizing urban-rural disparities and ecological considerations. However, certain aspects of the methodology and findings remain unclear:
The use of PLS-STEM is well-justified for small samples, but additional details on statistical validation, such as bootstrapping parameters, confidence intervals, or collinearity diagnostics, would clarify the robustness of the results. Additionally, while the findings focus on ecological and social drivers of adoption, practical barriers like system reliability, maintenance challenges, and financing options may significantly impact rural adoption decisions. Could the authors elaborate on whether these aspects were explored or might be included in future studies?
Finally, the focus on Sindh province provides context-specific insights, but its unique conditions may limit generalizability. A comparison with other regions or a discussion of potential confounders, such as grid connectivity or local incentives, would enhance the broader applicability of the findings. These clarifications would strengthen the study’s contributions to the policy and academic discourse on renewable energy transitions.
1. Table 2 shows the square root of AVE (SRAVE) values are quite close to the inter-construct correlations, and in some cases (e.g., ECAS vs. SODS in the total sample: SRAVE for ECAS is 0.88, correlation is 0.528; ECAS vs. ECCO: 0.88 vs. 0.500). More critically, the VIF values in Table 4 are very high for several constructs (e.g., SODS = 6.473, ADBE = 4.583, ECAS = 4.865), approaching the common threshold of 5. Given the high VIFs and only marginal discriminant validity, is it possible that the structural model suffers from multicollinearity, which could inflate the path coefficients and undermine the reliability of the findings for the driving factors?
2. The study uses a PLS-SEM with a complex model featuring six independent constructs and 27 indicator items. The final sample size is 311, which is adequate but not large. For the sub-group analysis (rural and urban), the sample is split into roughly 155 and 156 respondents per group. The key driving factors for the urban sample (ADBE, OPNI, ECAS) are found to have neutral effects in the rural sample. Can the authors confirm if a power analysis was conducted for these sub-group analyses? Given the much smaller sample size per group and the number of parameters to estimate (e.g., Table 3 shows many external loadings near 0.7), is it possible that the neutral findings in the rural sample are simply a result of low statistical power or Type II error, rather than a true theoretical difference?
3. The discussion section in 4.3 correctly states that ADBE, OPNI, and ECAS are driving factors in urban districts butneutral in rural districts. However, in the conclusion (Section 5), it is stated: the MDFs driving the households’ intention… included ADBE, OPNI, ECCO, SODS, and ECAS… Within rural districts, ADBE, OPNI, and ECAS were shown to have a neutral influence. This part is consistent. However, later in the same paragraph, the conclusion states: the findings indicated that while urban residents were open to innovation and were aware of the advantages… rural dwellers did not prioritize those advantages… and showed a lack of openness to innovation. The latter statement is an over-interpretation. A lack of a statistically significant effect (neutral influence) does not equate to a lack of openness or a failure to prioritize ; it simply means the data did not detect a significant relationship. Can the authors clarify why they interpret a non-significant path coefficient as a definitive lack of a characteristic, rather than simply an inconclusive finding?