Merging Variables

[Marko Obradović]

I am new to SPSS and I am trying to merge two variables. To be precise I have age variable which I recoded into categorical variable, age group. Now I have 5 age groups. Second, I have gender variable. And 3rd I have a question "would you buy a certain brand of car" (I will not say which brand). What I am trying to do is merge age group and gender variable, so that when analyze it along with the question about buying preference it would show how many young males would buy it, or how many middle-aged woman would buy it. I hope you understand my problem, since I am not a native english speaker. If u have any questions, I would try my best to clarify it. Please help!

[Maguin, Eugene]

Crosstabs agegroup by buycar by gender.

Gene Maguin

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Marko Obradovic
Sent: Monday, August 18, 2014 4:40 PM
To: [hidden email]
Subject: Merging Variables

I am new to SPSS and I am trying to merge two variables. To be precise I have age variable which I recoded into categorical variable, age group. Now I have 5 age groups. Second, I have gender variable. And 3rd I have a question "would you buy a certain brand of car" (I will not say which brand). What I am trying to do is merge age group and gender variable, so that when analyze it along with the question about buying preference it would show how many young males would buy it, or how many middle-aged woman would buy it. I hope you understand my problem, since I am not a native english speaker. If u have any questions, I would try my best to clarify it. Please help!



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Merging-Variables-tp5727000.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Bruce Weaver]

Rather than merging two variables into one, you might consider a model that has age and gender (and perhaps their interaction) as predictor variables.  This would allow you to treat age as a continuous variable, assuming you have actual ages.  (Carving continuous variables into categories rarely makes much sense, because it needlessly throws away information.  And very often, the cut-points between categories are completely arbitrary.)  

What is the outcome variable?  If it is Yes-No for each person, I would suggest coding Yes=1, No=0, and estimating a binary logistic regression model with predictor variables as outlined above.  (You can likely find examples of logistic regression with SPSS on the UCLA website--look for their "textbook examples".)

HTH.

Marko Obradović wrote

I am new to SPSS and I am trying to merge two variables. To be precise I have age variable which I recoded into categorical variable, age group. Now I have 5 age groups. Second, I have gender variable. And 3rd I have a question "would you buy a certain brand of car" (I will not say which brand). What I am trying to do is merge age group and gender variable, so that when analyze it along with the question about buying preference it would show how many young males would buy it, or how many middle-aged woman would buy it. I hope you understand my problem, since I am not a native english speaker. If u have any questions, I would try my best to clarify it. Please help!

[Marko Obradović]

Yes, I do use them as predictor variables and I have age for everyone. I just recoded age into age groups, so now I have that variable too. Outcome variable is in fact yes-no, and i coded it like you said. Thats all done, only analysis left. I will look into binary logistic regression. Thank you Bruce.

[Bruce Weaver]

There are many articles that talk about why it is usually a bad idea to carve continuous variables into categories.  Here's one very readable one:

http://isites.harvard.edu/fs/docs/icb.topic477909.files/dichotomizing_continuous.pdf

It talks about dichotomizing, but the points it raises also apply when you carve into more than 2 categories.

Good luck with your analyses.

Marko Obradović wrote

Yes, I do use them as predictor variables and I have age for everyone. I just recoded age into age groups, so now I have that variable too. Outcome variable is in fact yes-no, and i coded it like you said. Thats all done, only analysis left. I will look into binary logistic regression. Thank you Bruce.

[Marko Obradović]

Yeah,  that was it! I already tryied crostabs but I was puting two variables into row, and 3rd into column. I haven't tried puting all 3 into different fields. Thank u very much. This solved my problem.

[Marko Obradović]

Thank you. I will read the article thoug I read about it from my college books and profesors talked about it a lot. My analysis is almost done, I know what to do now, but I will try and get some more participants for the survey. Thaks everyone for help!