sas >> Statistical question -- compare two columns for statistical difference

by ni14 » Tue, 25 Jan 2005 04:33:00 GMT

Hello,
I have two columns of continuous data, col1 & col2. I would like to test them to see if they are statistically different or the same. (I son't if I should test to see if they are the same or different). Can anyone please tell me how you do this in SAS? Is it the t-test? If yes, what do you look for in SAS output? I think you use t-test only if col1 & col2 are normally distributed (col1 & col2 by represent probabilities, output form logistic--also I have two more vars col3 &col4 and these are just raw scores out of a neural network package and I need to do the same for these as well). Thanks much.
NICK
___________________________________________________________
Sign-up for Ads Free at Mail.com
http://promo.mail.com/adsfreejump.htm

sas >> Statistical question -- compare two columns for statistical difference

by art297 » Tue, 25 Jan 2005 07:05:44 GMT


Nick,

Bob already gave you good advice but, since you are comparing two columns
of data, I presume that the independence issue can't be ignored. If your
data are suitable for a t-test, I would guess that you want to look
up 'paired t-test' and meet those assumptions.

And, I'd strongly recommend doing at least some brief reading of a
statistic's text, as the first question should be what kind of statistic
can be used to analyze the type of data you have (e.g., choosing between a
parametric vs a non-parametric test) .. along with whether you meet the
test's assumptions.

Art
--------


them to see if they are statistically different or the same. (I son't if I
should test to see if they are the same or different). Can anyone please
tell me how you do this in SAS? Is it the t-test? If yes, what do you look
for in SAS output? I think you use t-test only if col1 & col2 are normally
distributed (col1 & col2 by represent probabilities, output form logistic--
also I have two more vars col3 &col4 and these are just raw scores out of a
neural network package and I need to do the same for these as well). Thanks
much.

Similar Threads

1. Statistical test for difference between two numbers

2. Any statistical test for comparing phase shift?

Hello,

I would like to statistically compare phase shifts obtained from cross
spectral analysis .  Does anyone know any references about this topic and
how to do it in SAS?   Or how to construct CI on a phase shift?

Thank you for your help!

Katrina

3. Two Senior level SAS Developer/Statistical Analyst opportunities in Columbus, Ohio

4. Question about the relationship between statistical

Typically, though not always, we do not observe the population of
interest. We usually observe a sample. We have to observe outcomes and
predictors in training sets, test sets, and validation sets, so we
consider them randomly-selected and perhaps stratified partitions of a
sample.

A survey sample means a subset of a population that hopefully has enough
power to yield sufficient statistical estimates of population
parameters. In observational research we also speak of observations of a
population as a sample as well.

We can only observe a sample of a very large population, or one that
extends into the future. We may not be able to sample at random from
some populations. In these situations the training set ---> test set
----> validation set progression helps prevent overfitting a model to
the specific characteristics of a larger sample. The terms "population"
and "sample" have special meanings in statistical theory. The
training-test-validation terms refer to elements of good practice in
statistical modelling.
S



-----Original Message-----
From:  XXXX@XXXXX.COM  [mailto: XXXX@XXXXX.COM ]
On Behalf Of Minze Su
Sent: Monday, January 14, 2008 6:36 PM
To:  XXXX@XXXXX.COM 
Cc: Minze Su
Subject: Question about the relationship between statistical Population,
Sample, training set, validation set and test set.


The training set + validation set + test set together  = sample or The
training set + validation set + test set together  = population? Can
anyone help?

5. Statistical Challenge Question

6. Question about the relationship between statistical Population,

The training set + validation set + test set together  = sample or
The training set + validation set + test set together  = population?
Can anyone help?

7. Somewhat OT statistical programming question

8. Statistical question pertaining to missing values