INTRODUCTION   One of the sources of nasal speech in persons with
cleft lip and palate ("CLP") may be slow reaction time of the
velopharyngeal (VP) flap, which acts as a valve between the nasal cavity
and the pharynx.  A speech pathologist conducted an experiment to
compare VP flap reaction time, digital [finger] reaction time, and labial (lip
closing) reaction in CLP subjects (adults with repaired cleft lip and palate)
vs. comparable non-CLP control subjects.  Digital reaction times were
measured as follows.  A subject sits at the keyboard of a microcomputer. 
The computer sounds a 500 millisecond (ms) low frequency "ready" tone
via the computer speaker.  A short (100 ms) high frequency "signal tone"
is sounded on a random schedule anywhere from 0.5 to 3.5 seconds after
the end of the "ready" tone.  On hearing the signal tone the subject presses
a key on the keyboard as quickly as possible, using the index finger of the
dominant hand.  A computer-driven clock measures the interval from the
signal tone until a key is depressed; the resulting digital reaction time is
reported in milliseconds.  VP reaction time is measured by a tiny
photodetector device inserted into the nasopharynx.  Once the probe is in
position and the equipment calibrated, a subject is asked to phonate the
nasal consonant /m/ (a humming sound) upon hearing the computer's ready
tone, and continue to do so.  Upon hearing the signal tone the subject
phonates a /pi/ sound as quickly as possible.  Changing from the /m/ to the
/pi/ sound requires closing the VP flap.  The computer measures the elapsed
time from the second tone to VP flap closure.  Labial reaction times, also
important to speech in CLP persons, are measured similarly; here the
subject makes an /ah/ ("Say ahhhhh...") sound from the ready tone; when
the signal tone is heard, the subject switches as quickly as possible to an
/m/ sound, which requires closing the lips.  The time (ms) required to close
the lips is reported by the computer as labial reaction time.  For each of the
three reaction time evaluations, each subject performed a series of 10 trials
in succession with 2 seconds between trials.  To account for possible
learning effects, the three tasks were performed in random order within one
session.  Unfortunately, the sequence in which the tasks were performed
was not recorded.

     The objectives of the study were:
     Primary objectives:  
          (1) To compare the means of the three types of reaction times.
          (2) To compare mean reaction times of adult CLP subjects
               (with repaired cleft lip and palate) vs. control subjects.
          (3) To determine if the difference between CLP means and
               control means depends upon the type of reaction time
               (Digital, Labial, VP).
     Secondary objectives:  
          (4) to estimate subject-to-subject variability in overall level of
               response (intercepts) within a session.

     The data are available in CLP01.SSD, a SAS 6.04 dataset, and
CLP01.ASC (ASCII), which contain one observation or record for each
recorded reaction time.  Table 1 contains descriptions of the variables. 
Table 2 contains some lines from a SAS PROC PRINT printout of
CLP01.SSD.  The ASCII version of the data file was obtained by editing
the PROC PRINT printout to delete lines of titles, names of variables, etc.

     For the purpose of this problem you may assume that any missing
data are missing completely at random, for reasons totally unrelated to the
process under study.