Leon J. Gleser
Executive Editor, Statistical Science
University of Pittsburgh
Department of Statistics
2732 Cathedral of Learning
Pittsburgh, PA 15260.

Dear Professor Gleser:

Thank you for accepting our paper  "Scientific Method, Statistical Method and the Speed of Light"
(ms 00-01-04). Please find enclosed a hard copy and a disk with the associated files
of a revised version. The files are also available for you to download directly at
http://www.stats.uwaterloo.ca/~rwoldford/papers/statsci  . 
All files in that directory are required or just select "allfiles.zip" and unzip it at your site.

We have changed the style of the references to match that of Statistical Science.   However,
we were unable to find a bibliographic style file to match  that of Statistical Science
(we include our entire reseaqrch database file research.bib in case Statistical Science has
constructed its own style file, e.g. statsci.bst).  As a compromise we have hand edited  the
bibliography file paper.bbl  to produce the desired capitalization (the file PAPER.bbl is a copy
of this file and is provided only as backup).

We have made minor changes to the paper to address some of the referees' concerns.
In other instances we have ignored those concerns for the reasons given below.

1. Find enclosed several plots that demonstrate the clustering identified in the data is not
a weekend effect as suggested by referee #2.  The plots show the clusters based on day and
also based on business day (i.e. with all  intervening days removed).  As can be seen the
clustering is mostly due to the temperature and day variates with some increased dependency based
on speed.  The latter is likely due to the temperature correction Michelson made to the speeds.
If referee #2 had other evidence, or meant something else by the comments, we would be pleased
to have the opportunity to respond before press time.

2. All of the referees took exception to our claim that the PPDAC structure is the statistical
method. We have added a sentence in the conclusion that deals with this concern but we are
deliberately sticking out our necks (what are necks for after all? :-) in the hope of provoking
further discussion of this issue. 

3. In spite of the comments of referee #3, we believe, taken as a whole (with all of the detailed
description and language), that the synthesis of statistical method given by PPDAC is new, We
can find no alternatives of comparable detail in the literature. Our contention is that the
structure is especially useful in the teaching of Statistics.  We have applied it with success 
on problems from a wide range of application areas and use it as the basis 
structure for many applied courses.  Indeed it is this application to real problems that has 
led to its development and refinement  - the single case of Michelson's study was chosen 
for the reasons given in the paper.

4. We have not commented on paragraph 2 of Referee #1's report.  While we agree 
about the  contingent nature of science, the point is not really related to our discussion 
of scientific method except possibly as a variant of falsifiability.  Were we to include it, 
the context of the discussion would have us (as with falsifiability) discuss it as a possible 
demarcation criterion  and then it is debatable whether contingency is sufficiently restrictive 
a criterion to exclude many things we might call non-scientific.



5. We have added a short paragraph in section 6 to clarify the non-cyclical nature of PPDAC to 
deal with the point raised by referee #1 in the final paragraph of the comments.  We hope that 
the text is sufficiently clear to students that executing a single PPDAC is not likely to be 
the beginning and end of  the statistician's role in a scientific enquiry.

We have attempted to write the paper in a lively and provocative style. Some of the sections such 
as 5 and 6 are short and certainly do not do justice to the issues raised. To do so would 
considerably extend the length of the paper.  We hope that these sections will spark the 
interest of the readers to promote further discussion.   Indeed, we hope that some of 
that discussion might take place in the issue in which the paper appears.

Thank you for accepting the paper for Statistical Science and  please express our thanks 
for the contributions of the referees.

Yours sincerely



R.W. Oldford and R.J. MacKay