是不是看的非标准化系数，如果是标准化系数大于1就真就问题了。
参考：https://bbs.pinggu.org/thread-881058-1-1.html
坛子里有一些讨论https://bbs.pinggu.org/thread-544977-2-1.html

我也认为是标准化系数问题

There is nothing wrong with unstandardised solution bigger than 1 only with standardised one. With your next question there is nothing also wrong with negative sign. The only reason where the sign would be a problem is not statistical but if there is a contrast with your theory. That is, if you believe that a variable affects positevely another variable but path coefficient is negative that is there is something strange but not with your statistical analysis. There is a very good site with SEM modelling where you can see.

http://davidakenny.net/cm/sem.htm

Hope it helps

谢谢各位，问题已解决！

谢谢，问题已解决！