Dear Aviva
I always understood that there are in principal two definitions.
1. The analysis of data for which there are several (more than 2)
measurements on each response point.
2. The situation where >2 of these measurements are VARIABLES (i.e. random
and not pre-assigned)
The canonical example is regression analysis.
I am happy to call this bivariate, regardless of what x and y are. The
important question is then to SPECIFY the properties of x and y, and to
choose an appropriate analysis.
But some purists (?) wd have said it is bivariate ONLY if x and y are both
variables. Otherwise it is "only data analysis" But post-Tukey I think this
is a silly line to take.
Please tell me what other people tell you.
Regards
JOHN BIBBY
NB: Due to computer problems, some email was lost over the period 11-14
July. Please re-send any of your email to QED which you think may have been
lost. Thank you!
==========================================
October 23-24th 1998 - the York Maths FunFair.
Please tell your friends, and mark in your diary!
(Sponsors + other support always needed)
Everybody is welcome at The York Maths FunFair on
Saturday 24th October. The events on Friday 23rd are especially for school
years 6 and 7. The social at 8pm on Friday is for everybody.
You can even make a weekend of it in York - a local hotel has offered family
rooms for as little as £35.
Details: Jane Phythian, 01904-421218 or [log in to unmask]
------------------------------------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|