Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!uunet!tymix!tardis!oliveb!bunker!wtm
From: rudy@mtqua.att.com (Rudy Vener)
Newsgroups: misc.handicap
Subject: Re: COCHLEAR IMPLANT
Message-ID: <15937@handicap.news>
Date: 2 Jun 91 12:07:54 GMT
References: <15771@handicap.news>
Sender: wtm@bunker.isc-br.com
Reply-To: rudy@mtqua.att.com (Rudy Vener)
Organization: AT&T Bell Laboratories
Lines: 44
Approved: wtm@hnews.fidonet.org
Fidonet: Blink Talk Conference

Index Number: 15937

In article <15771@handicap.news> Stephen.White@p1.f853.n681.z3.fidonet.org
writes:
>Index Number: 15771
>
> JC> I've been attempting to learn lipreading for 44 of my 49 years.  A
> JC> study  sponsored by an oralist school some years ago found that
> JC> lipreading  under the best of conditions is only 26% effective.
> JC> So much for the  glories of lip-reading.
>
>Jay, it isn't much use posting this type of stuff when it is so
>easily proved wrong. It doesn't do much for your credibility, and
>will cause doubt to be cast on other data that you supply.
>
>I, for one, have far more than a meagre 26% effectiveness rate when
>I am lipreading, even under the most adverse conditions. I have
>lipread people successfully under starlight, in disco strobe
>lights, and from just about every angle except from the back of
>their heads. I only have trouble understanding about 1 in 50
>people.

Jay, Steve:

Before we can make any claims about the rightness or wrongness OF
THE STUDY (not the individuals ability) we would need to know how
it was performed.  My experience with clinical tests is that they
try to eliminate extraneous variables and measure only ONE THING at
a time.

When I speechread I do quite a bit more than simply follow the lip
movements and mechanically convert them to words.  I am sure every
speechreader out there makes extensive use of contextual cues,
educated guesswork, knowledge of the speakers speech pattersn, and
a pheneomena which for lack of a better name I call expectancy of
dialogue.  If the subjects were asked to interpret a number of
unconnected words or even a number of unconnected sentences rather
than a self consistent, self reinforcing speech then all the little
additional aids we speechreaders use are lost.

In this case a 26 percent accuracy rate might be pretty good.

--
Rudy Vener  AT&T BTL uucp: att!mtqua!rudy  internet:  rudy@mtqua.att.com