英语论文网

tive advantage, we calculated the Pearson correlations coefficient in order to measure the linear association between variables (see Table 3). Based on the correlation analysis, there is no multicollinearity problem among independent variables.

According to Burns and Bush (2003), it proposed the rules of thumb about correlation coefficient size. That is, correlation coefficient that fall between +1.00 and +0.81 or between -1.00 and -0.81 are generally considered to be strong association. Those correlation that fall between +0.80 and +0.61 or -0.80 and -0.61 generally indicate a moderate association. Those fall between +0.60 and +0.41 or -0.60 and -0.41 are typically considered as weak association. Subsequently, those fall between the range of +0.21 and +0.40 or -0.21 and -0.40 indicating a very weak association between the variables. Finally, any correlation that is equal or less that ± 0.20 indicates rarely identifies a meaningful association between variables.

Among the factors contribute to internet adoption; result reveals that the relationship between relative advantage and internet adoption is significant positive (significant at 0.000, p < 0.01), the relationship between compatibility and internet adoption is significant positive (significant at 0.000, p < 0.01), the relationship between trialability and the internet adoption is significant positive (significant at 0.000, p < 0.01), the relationship between observability and internet adoption is significant positive (significant at 0.000, p < 0.01), and the relationship between complexity and internet adoption is significant positive (significant at 0.000, p < 0.05). The results indicated that the factors contribute to internet adoption for small businesses in Malaysia are high and relationships among the variables are strong.

Multiple Regressions

According to Churchill (1995), a regression model is considered to be valid when the predicted variables are not correlated among themselves (i.e. multicollinearlity) and the linearity of the relationship between dependent and independent variables. The main purpose of hypotheses testing is to test the four hypotheses that were developed earlier in the research methodology section.

The results of the multiple regression analysis shown in Table 5 indicate that the coefficient of determination (R square) of this model is 0.786, indicating a 78.6 percent of the factors contribute to internet adoption is explained by relative advantage, compatibility, trialability, observability, and complexity. In other words, these five independents variables predict a strong change in perceived value. Further, the F value of this model is 68.917.

From the results, complexity is the most influential determinants in the factors contribute to Internet adoption for small businesses in Malaysia. It has high positive impacts on Internet adoption (Beta=0.454) at 0.000 significant level. As such, hypothesis 5 is accepted. However, relative advantage and compatibility do not have significant influence on internet adoption (significant level=0.370 and 0.158). It is concluded that relative advantage and compatibility are not determinants of Internet adoption and本论文由英语论文网提供整理，提供论文代写，英语论文代写，代写论文，代写英语论文，代写留学生论文，代写英文论文，留学生论文代写相关核心关键词搜索。