Global Journal of Human Social Science, G: Linguistics and Education, Volume 23 Issue 8

The R-square value of this study's model, as shown in table 3, is 0.758. This suggests that teaching quality and its interaction with e-learning has explained 75.8% of the variance in students' satisfaction in Nigeria, while other factors not examined in this study explain the rest. d) Effect size (ƒ²), VIF and Predictive relevance (Q²) The ƒ² value provides an overview of an exogenous construct's effect on the endogenous latent variable. The values are 0.02, 0.15, and 0.35 for small, medium, and large effect sizes, respectively (Selya et al., 2012). The VIF indicates the absence or presence of multicollinearity. Table 4: Effect size (ƒ²), VIF and Predictive relevance (Q²) Constructs f 2 B-Perf Effect Size Servequal 0.664 Large E-learning 0.074 Small VIF Servequal 2.862 E-learning 2.862 Predictive Relevance Indicator SSO SSE Q² (=1-SSE/SSO) E-learning 7812 7812 E-learning*Servequal 279 279 Satisfaction 3627 2230.49 0.385 Servequal 5022 5022 From table 4, SERVQUAL has a large effect, while e-learning has a small effect on students' satisfaction. The VIF for the two constructs indicates no multicollinearity problem, as none has a value greater than 5. The Q² value, which shows the predictive relevance of the model, is greater than zero, as suggested by Duarte & Raposo (2010). e) Importance performance map (IPMA) analysis This study further conducted the importance- performance map analysis (IPMA) of the exogenous variables to the dependent variable, and the result is shown in figure 4: Figure 4: IPMA analysis Volume XXIII Issue VIII Version I 88 ( ) Global Journal of Human Social Science - Year 2023 G © 2023 Global Journals Effect of Teaching Quality on Students' Satisfaction in Nigerian Tertiary Institutions: The Moderating Role of E-Learning Amid COVID-19 Recovery