Actually I am not sure if you have 1 or 7 variables. You are asking the same question about seven categories of one's core business activity. Anyway if you set up a dummy variable on SPSS, you set one of the activities for example manufacturing as 1 and all the other are RECODED into 0. So you can assess one of your activities "against" all other remaining. It is sometimes handy for regression. You can also run cross-tabs (chi-square) and check if there are significant differences between those seven categories. 

Hope it helps a bit, good luck

Hi,

May be your question needs some elaboration. I will try to answer on different alternatives:

- if the respondents should answer only one of the 7 questions you have with yes {1}, then it is easy to compute a new variable by going to SPSS transform, compute, name the new variable and add the 7 dummy which you already have. the new variable in this case should have only 0 and 1.

- if the new variable have values more than 1, then that person give 2 or more yes answer to the original questions [in this case, it is your role to decide either to delete this person or to decide which answer to keep from the original]

- the third possibility, which I doubt that this what you want, is to keep the 7 dummy 

hope this help

You are dealing with "multiple answers". Apart from using dummy variables, there is a way of analyzing multiple answers in SPSS using the "Analyze - Multiple Response" dialog. This will get you frequencies and crosstabs for a new variable that is a set of your 7 dummies. However, if you want to test for significances you need to stick to the 1-0 dummies.

Hello Edward, It all depends on what you plan to do with the data.  The suggestions around 'recode' are valid but if you then want to use the information in that aggregated variable in correlational and/ or regression  analyses then that is not a valid approach. Why NOt? Because the categories have no inherent value and the resultant Beta -weights will be meaningless.  For instance 1= Manufacturing, 2= Plumber, 3 = laywer, etc... regressed onto 'happiness', will be meaningless as the 1, 2, 3, figurse do NOT mean anything - you could just as easily have coded then the reverse way vis a vis the labels and still have achieved the same statistical result.  So, you need to use 'dummy coding' first variable 0 = you are not manufacturing, 1 = you are manufacturing; second variable 0 = xxx, 1 = xxx; etc. which is what you appear to have already.  Of course you can do frequencies and cross-tables (chi Sq) on these data.

On the other hand, you can use these categories as independent variables in t-tests, Anovas, or Manova's... again coding into the one variable - as per the suggestions made by others - resulting in levels of the IV.  Important, think of how you want to use the data. Cheers, Kate

Edward, based on what you've said, I assume each record in the file will have only one of the 7 indicator variables = 1.  If that is so, something like the following (untested--no SPSS on this machine) syntax should do the trick.  I've assumed variable names core1 to core7 for the 7 indicator variables.  You'll have to replace them with your own variable names.  HTH.

NUMERIC core (F1).

COMPUTE core = 9.

VECTOR c = core1 to core7.

LOOP # = 1 to 7.

- IF c(#) EQ 1 core = #.

END LOOP IF c(#) EQ 1.

VALUE LABELS core

1 "Manufacuring"

2 "R&D"

etc

9 "Missing"

.

MISSING VALUES core (9).

* Check that things worked properly.

CROSSTABS core BY core1 to core7.

I think in your case you are dealing with a multiple response variable. If I think you have to assigned a Yes/No option to each variable then you can go to multiple response set, rename your variable and subsequently perform a frequency or a cross tabulation analysis for your variables.

One way of dealing with multiple checks by respondents is to group them to a different category, as it is easy for analysis at later stage. For example 1 to 7 to actual variables and 9 for multiple checks.

Try excel, easy to use and export.

Put all the variables in order of dummy value that you want to give and assuming that your first value is in cell B2, the formula for new dummy variable can be

=IF(SUM(B2:H2)=1, B2*1+C2*2+D2*3+E2*4+F2*5+G2*6+H2*7,9)

Hello Edward,

As you can deduce from the previous answers all depends om what you want to do  (how you want to analyse/what is the researchquestion) with the new variable(s).

So please clarify your question.

cheers,

Pierre