Rank Question

[Omar Farook]

Dear Experts,
  We asked our staff to select the right approach for communication between the staff, the staff should select three options out of 10, and they have to rank these options from one to three based one his or her preference.
  For example, e-mail=1, newsletters=2, and one to one meeting=3.
  My question is, by using SPSS how can we enter and analysis these data?
  Thanks.
  Omar.


---------------------------------
No need to miss a message. Get email on-the-go
with Yahoo! Mail for Mobile. Get started.

[Hans de Kleine]

Hallo Omar,

I am also dealing with this problem. I think the answer can be found in the
multiple response analyse. Because multiple answers can be given to a single
question.

Create separate variables for each answer option
Q(uestion)a(nswer) Variable label  and Value labels

q1a1  email                      1 'third preferred' 2 'second preferred' 3
'first preferred'
q1a2 newsletter                1 'third preferred' 2 'second preferred' 3
'first preferred'
q1a3 one to one meeting    1 'third preferred' 2 'second preferred' 3 'first
preferred'
q1a4 etc                          1 'third preferred' 2 'second preferred' 3
'first preferred'
q1a5 etc                          1 'third preferred' 2 'second preferred' 3
'first preferred'
q1a6 etc                          1 'third preferred' 2 'second preferred' 3
'first preferred'
q1a7 etc                          1 'third preferred' 2 'second preferred' 3
'first preferred'
q1a8 etc                          1 'third preferred' 2 'second preferred' 3
'first preferred'
q1a9 etc                          1 'third preferred' 2 'second preferred' 3
'first preferred'
q1a10 etc                        1 'third preferred' 2 'second preferred' 3
'first preferred'

Now you have got to make a grouping variables with multiple response sets

In your case you have got three counted values. So you have got to make
three separate grouping variables.

Select in SPSS multiple response sets… (analyse, tables)

Select variables q1a1 to q1a10  and put them into 'Variables in set'

You have got to specify the counted value. This is the value that you want
to count as 'yes'  For the first new variable choose 1

Call this new variable q1ans1 (SPSS assigns automatically a $ to the
variable, this shows you that it is a variable containing several
variables.)

Now you have got to do this also for the second and third answer, this time
the counted value is 2 and 3 Call the variable q1ans2 and q1ans3

You can analyse it with a multiple response table. (frequency or cross)

At the moment I'm try to figure out if it is possible to get one multiple
response variable instead of three separate ones.  But I'm not sure if it
can be done.

Greetings

Hans

[Omar Farook]

Dear Hans,
  First thanks for the information.
  Actually the staff first should select 3 out of 10, and after that his or she should rank the selected approach from 1 to 3.
  So in case if I will create 10 variable (q1a1 to q1a10) for each question, the staff will fill 3 variables and leave the rest.
  q1a1=3, q1a7=1, and q1a9=2.
  I think it is not a yes/no variable.
  Any feedback please?
  Omar.


Hans de Kleine <[hidden email]> wrote:    Hallo Omar,

I am also dealing with this problem. I think the answer can be found in the multiple response analyse. Because multiple answers can be given to a single question.

Create separate variables for each answer option
Q(uestion)a(nswer) Variable label  and Value labels
  q1a1  email                      1 'third preferred' 2 'second preferred' 3 'first preferred'
q1a2 newsletter                1 'third preferred' 2 'second preferred' 3 'first preferred'
q1a3 one to one meeting    1 'third preferred' 2 'second preferred' 3 'first preferred'
q1a4 etc                          1 'third preferred' 2 'second preferred' 3 'first preferred'
q1a5 etc                          1 'third preferred' 2 'second preferred' 3 'first preferred'
q1a6 etc                          1 'third preferred' 2 'second preferred' 3 'first preferred'
q1a7 etc                          1 'third preferred' 2 'second preferred' 3 'first preferred'
q1a8 etc                          1 'third preferred' 2 'second preferred' 3 'first preferred'
q1a9 etc                          1 'third preferred' 2 'second preferred' 3 'first preferred'
q1a10 etc                        1 'third preferred' 2 'second preferred' 3 'first preferred'
  Now you have got to make a grouping variables with multiple response sets
  In your case you have got three counted values. So you have got to make three separate grouping variables.
  Select in SPSS multiple response sets… (analyse, tables)
  Select variables q1a1 to q1a10  and put them into 'Variables in set'
  You have got to specify the counted value. This is the value that you want to count as 'yes'  For the first new variable choose 1
  Call this new variable q1ans1 (SPSS assigns automatically a $ to the variable, this shows you that it is a variable containing several variables.)
  Now you have got to do this also for the second and third answer, this time the counted value is 2 and 3 Call the variable q1ans2 and q1ans3
  You can analyse it with a multiple response table. (frequency or cross)

At the moment I'm try to figure out if it is possible to get one multiple response variable instead of three separate ones.  But I'm not sure if it can be done.
  Greetings
  Hans



---------------------------------
Food fight? Enjoy some healthy debate
in the Yahoo! Answers Food & Drink Q&A.

[meljr]

This is what I would try. Set up 1 variable for each of the 10 questions. For each respondent give them a score of 10 for the first choice, 9 for the second choice, and 8 for the third choice. For the other questions either leave as missing values or give a value of zero. When you run frequencies on the 10 variables you should see clearly who has the highest mean score. Even if you run correlations the values should hold.
My first post here, so it's worth a try.
meljr