I think pwdencrypt uses a hash, and the algorithm is built to make it hard to reverse the hashed string. The standard method for checking a password submitted by a user is to hash it and then compare it to the database's hashed version.
This is how you might check a used table that was entered :
SELECT password_field FROM mytable WHERE password_field=pwdencrypt(userEnteredValue)
(Big surprise) Replace userEnteredValue with the value that the user entered.