weighting a trendline on a scatter graph

[nepentanova]

Hi all,
a quick question.
i have a scatter graph that graphs a correlation of scores over 2
weeks.
x axis is score for week 1, y axis the score for week 2.
if a perfect correlation occurs then the trendline for the graph will
be 45 degrees.
the scores for each week are made up of an average of x measures. is
it possible for the trendline to take into account the # of measures
that makes up the week (so if a score is made up of 10 measures,
weight it 10 times a score made up of 1 measure)
i was going to use a function to amend the source data, so if a score
is made up of 10 measures, then have that number appear 10 times in
the source data, but typically the number of measures is between 300
and 3000, so that is not practical.
Any ideas?

 

[Jerry W. Lewis]

http://groups.google.com/groups?selm=3FFC11D7.3060403@no_e-mail.com

Jerry