unfortunately, most books are not clear on everything ... hopefully, the 
book will deal with basic issues and concepts reasonably well

ask students the following:

say you were at the race track and ... you have picked your favorite horse 
for race #7 ...

1. how many minutes do you think it will take the horse to cross the finish 
line? ______ (1 time only!)

2. what is range of times (in minutes) within which you are pretty 
comfortable (unless the horse keels over unexpectedly) that the horse will 
make it across the finish line ... fastest to slowest?

1 is a point estimate
2 is an interval estimate

then, you might want to explore HOW you could get either 1 or 2 ...

1 involves having data on the horse ... say from other races of the same 
length .... taking perhaps a random sample of THREE  race times .. (how 
sampled?) ... and using the mean of these 3 times as your best estimate ... 
hold on to your wallet though ... you are going to LOSE this bet!

2 involves again taking data from other races of the same length ... say a 
dozen samples of n=3 ... looking at the means from each of these SAMPLES 
... then lining them up say from slowest to fastest ... and then forming 
some BAND around the MEAN OF THESE MEANS ... (within some distance [enter 
the standard ERROR of the mean]  from the mean of the means) ... where you 
are pretty comfortable that you will have HEDGED YOUR BETS and will have 
included a sufficiently wide range to capture the REAL time that it takes 
the horse to cross the finish line

of course in 2 ... how comfortable you may be will depend of course on how 
many units (of error) you go on either side of the mean of the means ... 
the more units you go ... the more comfortable you will feel (have captured 
the real race result)  BUT ... while you are feeling more comfortable ... 
your PRECISION of narrowing down WHAT time the horse will finish in becomes 
less and less




At 04:47 PM 8/15/00 +0000, Sheila King wrote:
>Several people have pointed out, in response to my original question,
>that



=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
                  http://jse.stat.ncsu.edu/
=================================================================

Reply via email to