GLOBETROTTERDK  Trapeze unfortunately appears to be the best software on offer for Mac users wishing to convert .pdf files to .doc or .rtf format.

It really is a shame as the developer is very responsive to users needs and problems (also developer of Amnesty). Only about 1/3 of the .pdf files originally formatted in .doc or .rtf are reconverted in passable shape. Most either end up with mangled formatting or with unreadable text (maybe in Symbol font).

There are Windows programs that are able to convert from .pdf to .doc or .rtf around 98% - 100% of the time with perfect format recreation. Be careful if you go that route as there is a fair amount of vaporware in this category.  
(Version 1.3.1)

ANONYMOUS  I used the trial version (converts up to 3 pages only) to convert a pdf to rich text. I used Simple Text to make a few quick changes, saved as pdf again and had no problems at all. Worked flawlessly.  
(Version 1.2.3)

ANONYMOUS  In a random test of a PDF, when the converted RTF file was opened in Word, a substantial amount of the text appeared in the Symbol font which made it look like Greek. Not worth spending money on an application that cannot get its fonts right.  
(Version 1.2.3)

ANONYMOUS  It seems to do the job nicely on english PDF files. However, it will produce an error ("The file 'your file name.rtf' because a file system error occurred while trying to write the converted document.") if a converted file of the same name already exists. So, you must either trash the converted file of the same name or rename it before attempting to convert its PDF again. It would be more convenient if Trapeze simply overwrote the file or prompted the user for a new file name.

The error message also causes Trapeze to become stuck in the background if another app is brought to the front. One must click on Trapeze in the dock to bring it to the front in such a case as clicking on it with the mouse won't work. Then to dismiss the error message you must hit return on the keyboard.

All in all, it does a nice job of extracting text, but minor bugs could be frustrating over time. It's no doubt a small price to pay for those who must do this sort of text extraction on a regular basis, and for casual use it's small text tag at the bottom of the converted document is unobtrusive.  
(Version 1.2.2)

JEAN-PIERRE OLIVIER  I just tried light version of Trapeze 1.2.1, on a french document, the obtained word document is really OK  
(Version 1.2.1)

ANONYMOUS  Easy to use, but iIt only produces continues text without spacing between words. Not good.   
(Version 1.1.1)

ANONYMOUS  This app seems to work as advertised when used on documents comprising english text. However, if you have special characters like german umlaute or french accents etc., the app gets into deep trouble, ie. the results are unusable.

This should be mentioned somewhere, I guess.

Maybe it is due to the program that generated the PDF to be converted? I only tried PDF to RTF or HTML or plain text; my own PDF's are generated by OS X own PDF-engine steered by some TeX-application like teTeX, TeXShop. Though this produces PDF's which are perfectly readable by Preview or Acrobat, er, Adobe Reader, etc.  
(Version 1.1)

MESAD  PDFs created by scanned images do not contain any extractable text information. As a result, Trapeze cannot "extract" and "convert" anything from your scanned pages. FYI, this is answered in the FAQ section of Trapeze Help. It sounds like you are interested in obtaining an OCR tool, which is an imperfect solution, but may help you nonetheless.  
(Version 1.0.3)

SKYHORSE  I tried this out with regular black and white text PDF's, grayscale text and image PDF's, and text on multicolor background PDF's, all created by a Microtek 4850 scanner and readily accessible in Preview or Appleworks. Trapeze in all cases failed to open the files, reporting that they weren't PDF's.  
(Version 1.0.3)