Date: Mon, 1 Sep 2003 15:47:50 +0200
Reply-To: Rita Clivio <Clivio@databank.it>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Rita Clivio <Clivio@databank.it>
Subject: R: Segmentation
Content-Type: text/plain; charset="iso-8859-1"
are two variables enough to make a cluster analysis?
What about giving a code to every combination (15 : 5x3), may be using the formula
group =100 x buying situations code + ways code ?
In this way you can see which group you can consider together because of the small (and inconsistent) number .
Da: David Lindsay [mailto:email@example.com]
Inviato: luned́ 1 settembre 2003 15.16
I would appreciate some help with this.
I am about to set up 3 different buying situations (everyday,
once a year, and once in 5 years) and I will ask respondents to
tell me which one of 5 ways they would respond (impulse, some
thought, planned using the internet etc.). Clearly, I will get
a variety of responses and some people will respond to two or
more buying situations in exactly the same way.
I want to segment respondents, but unsure if cluster analysis is
the appropriate route, or what should I use.