It is a common mistake in the literature that people concern small values of R^2. However,I have no problem to observe that value of R bar squared. If the variation of your regressand is very low, R^2 can also be low.

Therefore, 0.62 is fine. If you have further question about my answer, contact me or send me your independent and indepent variables.

[此贴子已经被作者于2006-6-9 19:03:32编辑过]

在用OLS估计时,R平方只有0.64.该怎么使R值升高了: this is not a very meaningful question, it depends on your model, cross-sectional or time series one. and for models like logit and probit, R-square has little information.