Posted by Miguel (4.40.20.153) on September 18, 2000 at 11:38:21:
In Reply to: I have 2 words for you posted by Jack on September 18, 2000 at 11:00:12:
"Regression Analysis
A neat little statistical device that demonstrates the relationship between variables. There is lot of software that will perform
this automatically."
Agreed. However, there is not enough data at this point. I noted
the positive response as 1 and negative responses as 0 in order to
at least quantify v. a total n on a basic percentage scale.
"You need to quantify your results."
I tried above, with very little data I must add.
"Expand your sample size and make it as random as possible. Sampling is a very
powerful tool if done properly. Relatively small sample sizes can accurately predict large populations if the sample is
collected properly. The key to sampling is this: every member of the population has an equal chance of being selected."
Agreed. I can only get so many people to respond
an informal survey. I would love to have enough data
to apply paired-t, regression, correlation coeff, ANOVA, etc.
I guess a threat like; If you do not respond to this survey
I will send my cousin Vito to help your CH with a bat may help, :)
Get actual blood pressure readings from this randomly chosen sample size. Actual blood pressure readings from all
people is available. You would also need to normalize this data and control for age, gender, illness, etc.
Agreed. They weren't provided, nor did I ask for them. Good point.
"Regression analysis will tell you if there is a relationship between BP and clusters. If there is a relationship this many not
be causative but rather associative. A third variable or variables could be causing this BP deal and causing clusters."
It is associative. I doubt that it is causative based on
all literature avilable. Also, note my response to regression above.
Thanx for the input
Miguel
P.S.: Would you like me to e-mail you the data. I
used MS Excell 7.0. I would appreciate your help on this.