Date: Mon, 23 Aug 1999 17:38:52 -0400
Reply-To: Michele Atkinson <matkinson@DELTATRENDS.COM>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Michele Atkinson <matkinson@DELTATRENDS.COM>
Organization: DeltaTrends
Subject: Interactions
Content-Type: text/plain; charset="us-ascii"
I'm looking at one of our surveys questions that asks about overall
satisfaction. I want to look at whether there are differences among
departments from 98 to 99. It looks like all departments, except for two,
increased from 98 to 99. I ran a univariate analysis to test for
interaction. These are the results:
Tests of Between-Subjects Effects
Dependent Variable: 44
Source Type III Sum of Squares df Mean Square F Sig.
Corrected Model 359.636 13 27.664 36.033 .000
Intercept 89702.230 1 89702.230 116838.762 .000
DEPART 328.651 6 54.775 71.346 .000
YEAR 3.569 1 3.569 4.649 .031
DEPART * YEAR 16.718 6 2.786 3.629 .001
Error 6237.150 8124 .768
Total 140868.000 8138
Corrected Total 6596.786 8137
a R Squared = .055 (Adjusted R Squared = .053)
My questions is in regards to the Sig. If I am interpreting this correctly
I would say there is significant interaction between department and year.
Is this correct? If it is, does it apply to all possibilities? For example,
both Service and Body Shop have a lower mean in 99 compared to 98. Is this
the interaction? Are the differences between means significant based on the
above table or do I have to do further analysis? My goal is to be able to
state: In general satisfaction has increased by department except for
service and body shop. (or make a correct conclusion based on the data,
whatever that happens to be :)
Any insights are greatly appreciated.
Thanks,
Michele
| 
