Combine mutivariables into one

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

vhzv
Posts: 3
Joined: Thu Jul 26, 2007 9:57 am

Combine mutivariables into one

Postby vhzv » Thu Jul 26, 2007 10:07 am

Hello all,

I‘ve got v1, v2 and v3 and all these 3 variables have 5 options. What I want to do is to get just one variable. I have defined multivariable set for these 3 variables, when I run tables and cross with another variable which has e.g. 140 records the result does not add up to 140 nor 100%.

Record v1 v2 v3
1 1 2 2
2 2 3 3
3 3 2 4
4 1 5 1
5 1 2 2


What I did was to export my tables into excel and I have added v1+v2+v3 and for each option (1 to 5) calculated the mean, so I’ve got the result that I needed 140 records and the percentage added up to 100%.

There must be a better way to do it with SPSS.

I hope my question is clear and please HELP.....
andris
Site Admin
Posts: 137
Joined: Fri Jun 22, 2007 7:00 am
Location: Australia

Postby andris » Thu Jul 26, 2007 10:28 am

Hi vhzv,

You can use the Compute function (under Transform -> Compute) to achieve what you want. Make a new variable (fill in the name of the variable in "Target variable"). And use the following compute code:

(V1 + V2 + V3)/3

Where var1, 2 and 3 are the names of your variables. You can use more than three if neccessary. Let us know if that gives you the result you want.
vhzv
Posts: 3
Joined: Thu Jul 26, 2007 9:57 am

Postby vhzv » Thu Jul 26, 2007 12:07 pm

Thanks for you quick reply Andris, I will try that.
yvesvv
Posts: 310
Joined: Wed Jun 27, 2007 8:58 pm
Location: Belgium

Postby yvesvv » Thu Jul 26, 2007 2:38 pm

Hi,

If the three variables measure a same thing, i would advice you to also calculate the cronbach's alpha-coefficient (Analyze - scale - reliability analysis ) as well. When this is higher than 0,6 or 0,7 (depending on the author) you can take the three variables together into one new variable...

Of course, only do this when the three variables measure a same construct...
vhzv
Posts: 3
Joined: Thu Jul 26, 2007 9:57 am

Postby vhzv » Fri Jul 27, 2007 1:03 pm

Hello all,

Andris it didn’t work; I will be more specific with my problem.
I have 3 variables v1,v2 and v3 and each of these variables have 5 options (option 1=coca cola, option2=Pepsi, option 3=Fanta, option 4=sprite, option 5=7 up)

In v1 record 1 the respondent has chosen option 2(coca) and in V2 the same respondent (record 1) has answered option 3 (fanta) and v3 record 1 option 5 (7 up)

Record v1 v2 v3
1 coca Fanta 7 up
2 sprite sprite coca
3
4

Each respondent have chosen different options in v1,v2 ,v3
I have defined multivariable set for these 3 variables, when I run (summary tables for v1,v2,v3) and cross with another variable which has e.g. 140 for Coca, 87 for Pepsi, sprite 60, fanta 50 and 7 up 40 records; the result does adjust to 140 records nor 100% for coca cola and the same thing happens with the other brands.
This is the result that I got:

coca Pepsi sprite fanta 7 up
(v1) what brand do you like 30 19 20 80 70
(v2) what brand you don't like 86 57 20 20 20
(v3) what brand would you buy 61 90 50 40 10
Total 140 87 60 50 40

What I need is that Coca cola must add to 140 and Pepsi also have to add up to 87, etc.
In tables there is not an option where I can get the mean or median for v1,v2 and v3 for each option (coca,fanta,etc.).

Thank you for you helps.
andris
Site Admin
Posts: 137
Joined: Fri Jun 22, 2007 7:00 am
Location: Australia

Postby andris » Fri Jul 27, 2007 1:57 pm

I think I still not get what your end result should be. What mean do you want of what variables? What is your goal? I see what you have measured, but what do you want to get out of your measurements?

Who is online

Users browsing this forum: No registered users and 2 guests

cron