Global Journal of Computer Science and Technology, G: Interdisciplinary, Volume 23 Issue 1

According to the results derived in Table 5 based on the Pearson correlational analysis, it was found that, the User Training and Education, Communication, and Implementation Strategy factors seem to have a statistically significant correlation with the remote ERP implementation success (Sig. (2-tailed) < 0.05). Among these three factors, User Training and Education has the strongest positive correlation (r=0.878) with remote ERP implementation success, while a moderate correlation is witnessed between Communication (r=0.542) and remote ERP implementation success, and a weak positive correlation is witnessed between Implementation Strategy (r = 0.249) and remote ERP implementation success. Although Top Management Commitment, Change Management, and Project Management seem to have positive correlations with remote ERP implementation success, they are not statistically significant, since the Sig. (2-tailed) has been greater than 0.05. c) Multiple Linear Regression Analysis According to the previous section, it was identified that a positive relationship exists between all the six independent variables and remote ERP implementation success. This section proceeds with a deeper investigation based on assessing the impact of the identified factors that ensure remote ERP implementation success. For this, the study accommodated the multiple linear regression model. First, the data were analyzed to check on the convenience for regression analysis. Accordingly, the assumptions of normality, linearity, and absence of collinearity were tested as prerequisites for a multiple regression analysis. With reference to (Pallant, 2001), there are a few main identified assumptions to be tested for a multiple regression which include, 1. The required sample size for a regression test 2. No multicollinearity between independent variables 3. Normality distribution of data set and test for outliers 4. Linearity between independent and dependent variables 5. Homoscedasticity of independent variables Assumption 01: The Required Sample Size for a Regression Test As cited in (Pallant, 2001), it is recommended for social research to have at least 15 subjects per predictor for a valid regression test. The ideal sample size for a regression can also be calculated using the following formula: n> 50 + 8m (m = number of independent variables) (Pallant, 2001). Accordingly, the study consists of 6 main independent variables with primary data collected from 269 ERP system users. When outliers are comprised in a data set common, those data points should be removed (Osborne and Overbay, 2004). However, no outliers nor missing values were spotted in the data set, and therefore, 269 responses were used for a regression analysis. Table 6: Multicollinearity Test Based on Tolerance and VIF Value (Constant) Tolerance VIF Top Management Commitment 0.978 1.023 Change Management 0.859 1.164 Communication 0.672 1.488 User Training and Education 0.700 1.429 Implementation Strategy 0.700 1.429 Project Management 0.978 1.023 Source: Authors’ Representation Based on SPSS Results © 2023 Global Journals Global Journal of Computer Science and Technology Volume XXIII Issue I Version I 39 ( )G Year 2023 Assumption 02: Multicollinearity Test of Predictor Variables The collinearity diagnostics confirm whether there is a serious problem with multicollinearity. Condition Index Values greater than 15 indicate a possible problem with collinearity; greater than 30, a serious problem. A tolerance value < 0.10 suggests a concern with (multi) collinearity. VIF is simply the tolerance's reciprocal value. As a result, VIF values > 10 suggest concerns with collinearity. The results of these analyzes are presented in Table 6. It could therefore be seen that the tolerance levels are > .10 and VIF values are < 10 for all independent variables. The study also revealed a safer facet in its Condition Index (CI) values as shown in Table 7. All independent variables had a CI < 15, which thereby indicates that the predictor variables are free from multicollinearity. Critical Success Factors of Remote ERP Implementation: From System Users’ Perspective