[R] marketing-questionnaire with split-design

[Jim Lemon]

On Tue, 9 Sep 2014 04:42:11 PM Norbert Dörrer wrote:
> Dear Experts
> 
> I'm quite new to the world of R; so maybe my question is kind of
> beginners...
> 
> I am right now trying to analyse questionnaire data (CAWI) for 
producing a
> image/positioning map of the competitors of my company; when 
programming
> the CAWI, a split design was needed; every respondent had to 
answer a
> couple of multiple-choice questions (on reputation of company, 
image,
> sympathy etc.; 1 "perfect" to 5 "absolutely not perfect") for two 
companies
> only - out of a list of several companies; companies were chosen 
randomly.
> 
> So what I have now is two variables that tell me which two 
companies the
> respondent was actually evaluating...
> Company 1:      A, B, C, D, E, F or G
> Company 2:      A, B, C, D, E, F or G
> 
> ... and of course several image-related variables:
> reputation of company 1:  3,4,2,3,5,1,2,3,4,2,3,1,5 and so on
> reputation of company 2:  5,2,3,4,2,5,1,2,3,2,4,21 and so on
> image of company 1:        2,4,3,5,2,3,4,1,2,5,3,4,2 and so on
> image of company 2:        3,4,5,2,5,4,5,3,4,2,5,1,2 and so on
> sympathy of company 1:  1,5,3,4,2,5,3,4,1,2,3,4,5 and so on
> sympathy of company 2:   5,5,4,5,3,3,4,3,3,3,4,2,5 and so on
> etc.
> 
> 
> But what I need in order to do proper calculations is one
> Reputation/Image/Sympathy-Variable for every single company, like 
this:
> reputation of Company A: e.g. 3,4,3,5,NA,3,NA,5,2,1,NA,NA
> reputation of Company B: e.g. NA,4,3,5,NA,3,NA,5,2,1,NA
> reputation of Company C and so on
> image of Company A:...
> image of Company B :...
> image of Company C and so on
> sympathy of Company A:...
> sympathy of Company B :...
> sympathy of Company C and so on
> 
> So I need some syntax to compute a new variable  e.g. 
"reputation_CompanyA"
> that contains the values of "reputation of company 1" whenever 
"Company
> 1"=A and that contains the values of "reputation of company 2" 
whenever
> "Company 2"=A... else is of course missing values!
> 
> The solution has to has something to do with loops and flow control, 
but I
> really can't solve this puzzle.
> 
Hi Norbert,
I'll assume that you have your data (call it nddf) in a form something 
like this:

	rep1  rep2  img1  img2  sym1  sym2  ...
S1         3       4        2        3         1         5
S2         4       2        4        4         5         5
S3         ...

and your allocation of companies to subjects (call it subcomp) is in 
another file like this:

S1         A       C
S2         B       D
S3         ...

If so, you can do something like this. Say you surveyed 10 subjects 
and your overall variable is the sum of the five (?) scores:

# create a list of company by attribute scores
comp_by_attr<-vector("list",7)
names(comp_by_attr)<-LETTERS[1:7]
# step through your rating data
for(subject in 1:10) {
  company1<-
   which(subcomp[subject,1]==names(company_by_attr))
  company2<-
   which(subcomp[subject,2]==names(company_by_attr))
  # add the sum of the subject's company 1 ratings
  comp_by_attr[[company1]]<-c(comp_by_attr[[company1]],
   sum(nddf[subject,seq(1,9,by=2)]))
  # add the sum of the subject's company 2 ratings
  comp_by_attr[[company2]]<-c(comp_by_attr[[company2]],
   sum(nddf[subject,seq(1,9,by=2)]))
}

You should now have a list of seven elements, each labeled with the 
code of the company and containing the overall scores for that 
company. Since I don't have the data, I haven't checked this, but it 
may get you out of trouble.

Jim