Global Journal of Human Social Science, G: Linguistics and Education, Volume 25 Issue 3

Table 1: Respondent Profile Survey Respondent (n=286) Measure Item N (%) Measure Item N (%) Gender Female 147 51,4 Level of Studies Undergraduate 124 43,4 Male 139 48,6 Graduate 149 52,1 Phd Candidate 13 4,5 Age 18-24 163 57,0 25-34 115 40,2 35-44 8 2,8 CHATGPT Learning Strongly Agree 130 45,5 ChatGPT Addiction Strongly Agree 120 41,96 Agree 129 45,1 Agree 99 34,62 Neutral 13 4,5 Neutral 32 11,19 Disagree 10 3,5 Disagree 25 8,74 Stronlgy Disagree 4 1,4 Strongly Disagree 10 3,50 ChatGPT Integrations Very Likely 118 41,3 ChatGPT USE Never 3 1,05 Likely 139 48,6 Rarely 32 11,19 Neutral 15 5,2 Occasionally 108 37,76 Unlikely 11 3,8 Frequently 143 50,00 Very Unlikely 3 1,0 ChatGPT Accesibility Absolutely 135 47,2 In some cases 142 49,7 No 9 3,1 V. R esults The analysis of the data in our study was conducted using a statistical method called structural equation modeling (SEM). This approach helps us understand the relationships and causal links within the conceptual model we presented earlier. It involves analyzing numerical data using techniques like multiple regression and factor analysis. Given the complexity of our model, we opted to use the partial least squares method, as recommended by Chetiou et al. (2022), to ensure accurate analysis. To assess the quality and accuracy of our research, we'll be looking at several factors, including indicator reliability, construct reliability, convergent validity, and discriminant validity. These measures help ensure that our data is reliable and valid for drawing conclusions. Once we've established the reliability and validity of our measurements, we'll proceed with the structural modeling test using SMART- PLS 4. a) Assessment of the Measurement Model The results obtained from Table 2, The Cronbach's Alpha values obtained in this research indicate how reliable our measurements are. According to Pallant (2001), if the Cronbach's Alpha value is above 0.6, it means our measurements are quite reliable and acceptable (Nunnally and Bernstein, 1994). Conversely, if the value is below 0.6, it suggests lower reliability. Our study found that all variables had Cronbach's Alpha values higher than 0.6, meaning they are reliable and suitable for analysis. Composite Reliability (CR) measures how consistent and dependable our constructs are. To be considered acceptable for our study, CR values should be higher than 0.7 according to Hair, Hult, Ringle, and Sarstedt (2014). Our analysis revealed that all variables had CR values exceeding 0.7, indicating that they are reliable and suitable for our study. To ensure the reliability of our measurements, we need the Average Variance Extracted (AVE) value for each variable to be greater than 0.5 (Hair et al., 2010). After conducting our analysis, we found that all variables exceeded this value. This indicates confirming the reliability of our measurements. In this studty a method by Fornell and Lacker (2018) is used. They suggest that the average root square of the average variance extracted (√AVE) for each aspect should be higher than its correlation with any other aspect. We checked this using Table 3, and it seems that each aspect is indeed different enough from the others. ChatGPT in Moroccan Education Sector: Examining the Attitude of Student Acceptance and Usage Intent Global Journal of Human-Social Science ( G ) XXV Issue III Version I Year 2025 15 © 2025 Global Journals