Unable to match octal character

[Boylan, Ross]

I have a file with some characters that display in the emacs buffer as \203.  
Yet when I search for that, using
C-s C-q 203 <ret>
I can't match it.  Likewise if I use search and replace.

I have verified that \203 is the display of a single character both by using 
the arrow key (it's a single step to move over it) and by searching on the 
literal string \203 or 203.

What's going on?

When I put the cursor on one of these characters and do describe-char I get
            position: 474 of 48736 (1%), column: 25
            character: \203 (displayed as \203) (codepoint 4194179, #o17777603, 
#x3fff83)
    preferred charset: tis620-2533 (TIS620.2533)
code point in charset: 0x83
               syntax: w        which means: word
             category: L:Left-to-right (strong)
             to input: type "C-x 8 RET HEX-CODEPOINT" or "C-x 8 RET NAME"
          buffer code: #x83
            file code: #x83 (encoded by coding system raw-text-unix)
              display: not encodable for terminal

Character code properties: customize what to show
  general-category: Cn (Other, Not Assigned)
  decomposition: (4194179) ('?')

Is the fact that the codepoint is not 203 significant?

History of the file:
SAS running on MS Windows produced rtf output.  SAS has its own font it likes 
to use,
and the horizontal bar used in tables does not travel well (even on Windows).  
That's the character that's causing trouble.
Opened in Wordpad on Windows and exported text.
Move the file to Debian GNU/Linux (UTF-8 environment) and opened the text file 
in emacs 24.4.1.

Using a slightly different procedure I was able to replace octal characters:
Started with same rtf output.
On linux ran unrtf -- text to convert to text.
Open text file in emacs.  In this case the character was \220.
I tried wordpad because unrtf did not preserve the column alignment

Thanks.
Ross Boylan.