In case of model fit the value of chi-square(CMIN/DF) is less than 3 but whether it is necessary that P-Value must be non-significant(>.05).If my sample size is very large it is not mandatory that I have found in one book.As a non-significant chi-square shows that my model is not different than the default one and it fits well.I have found some article where they have taken a highly significant chi-square value.The only focus is that the value must be less than 3.I want to know which approach is correct one-A significant chi-square(p-value) less than 3 or a Non-significant Chi-square with value less than 3. Whether a large sample size play any role here.