Comparison of 2 groups on SPSS

#1
Hi,
I have 2 separate patients data sets, one from patients with a severe disease and another from a group of patients with a milder disease. The same variables were measured for both data set. What SPSS analysis method can I use to analyze the impact of one variable on both datasets. I am trying to find out if this one variable can predict outcomes in both data sets and if so what is the relationships like. Being a Novice as well, I would appreciate help with interpretation of the results when done.
 

Karabiner

TS Contributor
#2
The problem is in no way software-specific, as far as I can see. The appropriate analysis depends on how that variable was measured, how outcome was defined an measured, what exactely you mean by "can predict outcomes in both datasets" (for example, do you want to know whether the prediction is different between sets), and on sample size.

With kind regards

Karabiner
 
#3
The problem is in no way software-specific, as far as I can see. The appropriate analysis depends on how that variable was measured, how outcome was defined an measured, what exactely you mean by "can predict outcomes in both datasets" (for example, do you want to know whether the prediction is different between sets), and on sample size.

With kind regards

Karabiner
Thanks a million for your reply. Lets put it this way I have a certain on day zero and patients response to this treatment was monitored over a period of time (so there are repeated daily measures of certain variables). I have observed a negative dose response relationship between the treatment and the responses (statistically), but it seems like certain doses give a faster response and thereby a better outcome (just by eyeballing the data). I would like to demonstrate or show this using statistics. In addition, I have split the data sets into 2 (still all on the same work sheet) and would like to find out if there is a significant difference between the 2 sets. hope that makes it clearer.
Once again thanks a million for taking out time to respond.
Best regards
 
#4
Thanks a million for your reply. Lets put it this way I have a certain on day zero and patients response to this treatment was monitored over a period of time (so there are repeated daily measures of certain variables). I have observed a negative dose response relationship between the treatment and the responses (statistically), but it seems like certain doses give a faster response and thereby a better outcome (just by eyeballing the data). I would like to demonstrate or show this using statistics. In addition, I have split the data sets into 2 (still all on the same work sheet) and would like to find out if there is a significant difference between the 2 sets. hope that makes it clearer.
Once again thanks a million for taking out time to respond.
Best regards
I use SPSS because I understand it (but not as much as Excel)
 

Karabiner

TS Contributor
#5
I still don't know what kind of variables are concerned. Is outcome measured on a binary scale?
Or categorical, ordinal, or interval scale? The impact variable is measured on an interval scale, I suppose?

By the way, if someone eyeballs data and then tailors a statistical test to the pattern he has seen,
then the resulting p-values are biased. Eyballing is like perfoming a large number of informal "tests"
on the data, and the subsequently performed statistical test does not take that extensive pre-testing
into account. It is some version of the Texas sharpshooter fallacy: the Texan fires some gunshots at the side
of a barn, then paints a target centered on the tightest cluster of hits and claims to be a sharpshooter.
(cf. wikipedia)


With kind regards

Karabiner
 
Last edited: