What do scatter plots of pollutants vs attributes look like?
-------------------------------------------
Emil M Friedman, PhD
2304 Richmond Road
Beachwood, OH 44122
[log in to unmask]
216-591-1750 (voice)
775-249-6744 (fax)
----- Original Message -----
From: "Kieran Towey" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Thursday, May 22, 2003 10:56 AM
Subject: Trend Analysis
Allstaters
An analyst undertook a survey in Spring, Summer, Autumn and Winter to
determine the levels of pollutants in water. I have been given the
aggregated data, i.e. the mean values of each attribute at each of the time
points and the pollutant score for each time point. Hence I have only 4 data
points and have no access to the underlying survey data (analyst has left
and no-one can find the underlying data :) ). I have been asked to comment
on the relationship between the attributes and the polutant score, i.e study
the trend and say which variables are most relatated to changes in the
pollutant score.
As there are so few data points there is not much I can do except show the
relationship between the attribute means and the pollutant score over time.
The pollutant score "looks" semi circular...ie low in Spring, rises in
Summer and Autumn and falls in Winter, whereas most of the attribute means
are the inverse of this, high in Spring, low in Summer, Autumn and high
again in Winter. Is there any summary correlation type statistic that may
indicate this relationship. I understand that the data set size is extremely
limiting but if there was some statistic that might indicate the functional
relationship it would help. There are 50 attribute means and therfore 50
plots so some indicator of the "strong" relationships would be of
assistance.
Regards
Kieran
|