Date: Fri, 19 Mar 1999 09:50:35 -0300
Reply-To: GLAUBERF@VAREJO.COM.BR
Sender: "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From: Glauber Fonseca <GLAUBERF@VAREJO.COM.BR>
Subject: Test if Samples are Dependent or Independent
Content-type: text/plain; charset=us-ascii
Hi, everbody.
I am facing some problems interpreting SPSS 8.0 for Win results.
This is what I want:
I know the consumption of a certain material for the last 36 months. Based
on this history, I need to foretell the consumption for the next 12
months.Through the Monte Carlo technique I got to the consumption for the
next 12 months, based on the history consumption. So now I have two samples
(1- 36 numbers, 2 - 12 numbers. I have to test if sample 2 (12 elementes)
belong to the same distribution of sample 1 (36 elements).
I have done the T-Test for 2 independent Samples and got the results shown
below:
- Group Statistics:
Grupo 1 (hystorical data):
N: 36
Mean: 4021,305
Std. Deviation: 2913,4024
Std. Error Mean: 485,5671
Grupo 2 (forecast):
N: 12
Mean: 3083,3333
Std. Deviation: 2193,3094
Std. Error Mean: 633,1539
- Independent Samples Test:
Levene's Test for Equality of Variances: F = 3,222
Sig. = 0,079
- Equal variances assumed:
t-test for Equality of Means: t = 1,020
df = 46
Sig. (2-tailed): 0,313
Mean Difference = 937,9722
Std. Error Difference = 919,4538
95% Confidence Interval of the Difference: Lower = -912,7923
95% Confidence Interval of the Difference: Upper = 2788,7367
- Equal variances not assumed:
t-test for Equality of Means: t = 1,176
df = 25,024
Sig. (2-tailed): 0,251
Mean Difference = 937,9722
Std. Error Difference = 797,9093
95% Confidence Interval of the Difference: Lower = -705,2740
95% Confidence Interval of the Difference: Upper = 2581,2185
How can I say sample 2 belongs or does not belong to the same distribution
of sample 1 with a 95% confidence level?
Which results should I compare ? Please describe it in details.
Can you help me ? Is this the best test I can do ? Or should I do another ?
TIA,
Glauber Fonseca
| 
