As @dnbrv commented, this is not an encoding issue but probably a font problem. Try first changing the font of some problematic text segment to Arial Unicode MS. This probably won’t help (as Word should be able to switch font automatically). In that case, you might check out the identity of some problem character: place the cursor right after a character and enter Alt+X. The character should now turn to its Unicode number, which can be used to analyze the nature of the problem. You could also download and install a very large font like Code2000 to see whether it can show the characters.
On the other hand, if the document is not originally in Word format, there might have been a problem in importing it into Word. The extension .doc does not as such guarantee anything.
Have you tried changing the font all together? Some fonts have more "defined" character sets for their unicode implementations. – Chad Harrison – 2012-01-13T14:10:55.937
Yes, but I am 90% sure this one is supposed to be encoded in Windows-1250 instead of Unicode... – Einsteins Grandson – 2012-01-13T14:18:53.910
MS Word doesn't have encodings but fonts have character sets. The problem is likely that the font doesn't have those characters defined so Word displays empty squares. – dnbrv – 2012-01-13T15:12:09.270