To compute the one-tailed significance probability, first determine whether large values
of t are significant or small values are. Let p denote the significance probability
for the two-tailed test. If large values of t are significant, then the one-tailed probability
is p/2 if t >0, and is 1 − p/2 if t < 0. If small values of t are significant, then
the one-tailed probability is 1 − p/2 if t >0, and is p/2 if t < 0.

[此贴子已经被作者于2008-8-19 15:30:17编辑过]

[此贴子已经被作者于2008-8-19 15:32:31编辑过]

7楼的，能再说清楚点吗？最好用中文表述，不胜感谢！如独立样本t检验H0:µ<=650;H1：µ>650,

用SPSS的双尾检验p值＝0.009，t值＝－2.64，究竟接受还是拒绝原假设？

那如果是配对样本，又该怎么操作呢？谢谢

[此贴子已经被作者于2008-8-20 11:45:43编辑过]

这是一个问题，spss为什么没有单尾呢

如独立样本t检验H0:µ<=650;H1：µ>650,

用SPSS的双尾检验p值＝0.009，t值＝－2.64，究竟接受还是拒绝原假设？

那如果是配对样本，又该怎么操作呢？

当然不能拒绝原假设了……样本均值是小于650；

配对样本应用也是一样的