variable set up...ordinal
hdhdhd123
My query is:
In Spss, is it possible to put the following response under a single
variable.
Presently, I know that I need to break this question into three
variables like-
col_red,col_gre and col_blu
But obviously if I could able to accomodate these three variables in a
single one;that would save time regarding data entry and making
variables! Again, i suppose variable measure is ordinal for this
question.
Any response to the above two queries will be highly appreicated and of
great help.
Q. Please rank the colour you like from the following?(e.g Red=2;
Green=1;Blue=3 or anothe respondent may put it this way Red=3;
Green=2;Blue=1 etc )
A. Red
B.Green
C.Blue
Thanks,
hd
Richard Ulrich
wrote:
> Hi,
>
> My query is:
> In Spss, is it possible to put the following response under a single
> variable.
> Presently, I know that I need to break this question into three
> variables like-
> col_red,col_gre and col_blu
> But obviously if I could able to accomodate these three variables in a
> single one;that would save time regarding data entry and making
> variables! Again, i suppose variable measure is ordinal for this
> question.
If you have *3* variables with ranks, then each is ordinal.
If you combine choices, then it will be categorical --
"For the colors, red, blue and green, which do you like best
and next best?
1. Red and blue
2. Red and green
3. Blue and red
4. Blue and green
5. Green and red
6. Green and blue."
> Any response to the above two queries will be highly appreicated and of
> great help.
>
> Q. Please rank the colour you like from the following?(e.g Red=2;
> Green=1;Blue=3 or anothe respondent may put it this way Red=3;
> Green=2;Blue=1 etc )
> A. Red
> B.Green
> C.Blue
--
Rich Ulrich, wpi...@pitt.edu
http://www.pitt.edu/~wpilib/index.html
hdhdhd123
Thank you so much Rich for your prompt response. Your reply provides a
new direction for me to think about my query.
A further query on this issue is-
Suppose a respondent just ranks Green= 2. He does not touch Red and
Green. So in SPSS, I need to put on Green 2nd as one of the value of
'Color' variable as well as the others you mentioned(e.g.
blue-red-green;green-blue-red etc). Thus it will need me to put all the
combinations as values of 'Color' variable and obviously it will be a
long list. Just wondering, if there is any way that could deal this
issue in a shorter manner.
Besides, is there any option in SPSS so that values could be copied and
pasted to another variable value? If possible, this will obviously save
lots of time on keying in the values.
Any info on these will be highly appreciated.
Regards,
hd
Richard Ulrich
wrote:
> Hi,
>
> Thank you so much Rich for your prompt response. Your reply provides a
> new direction for me to think about my query.
>
> A further query on this issue is-
> Suppose a respondent just ranks Green= 2. He does not touch Red and
> Green. So in SPSS, I need to put on Green 2nd as one of the value of
I suppose that this is one more reason why "ranking"
is a crappy way to collect data. Avoid it, when there
is any alternative.
Also, it is one more reason why compressing data to
the minimum number of variables is a fairly bad idea.
It sounds like you are taking a set of data that are
collected in a (more) reasonable way, and trying to force
them into a compact format -- That hasn't been a target
of commonplace database design since we moved away
from using computer cards.
> 'Color' variable as well as the others you mentioned(e.g.
> blue-red-green;green-blue-red etc). Thus it will need me to put all the
> combinations as values of 'Color' variable and obviously it will be a
> long list. Just wondering, if there is any way that could deal this
> issue in a shorter manner.
>
> Besides, is there any option in SPSS so that values could be copied and
> pasted to another variable value? If possible, this will obviously save
> lots of time on keying in the values.
You can define a set of value labels for several variables
at once, if that is what you are asking.
Value list Pick1 to Pick3 1 'red' 2'green' 3'blue' .
JKPeck
which treats a set of variables as a single multiple response set.
This is available in basic ways through the Multiple Response procedure
in the Base, and there are much more extensive facilities in the Custom
Tables option. For these you will needs separate groups of variables,
but the multiple dichotomy and multiple category set types handle these
situations. Which is best depends on the dimensionality of your answer
space and some other characteristics.
mca...@hotmail.com
and choice 3. The possible answers for each are red, green blue.
This would take care of the missing values dilemna. When someone
choses only green for choice 2, then their three variables would be
Choice one - user missing
Choice two - green
Choice three - user mising
I have no idea if this is adequate for your purposes but I agree that
trying to condensce these things is usually not a good idea.
Marc
monia9PL
your respondent can invent ;) You can make it as mcam54 proposed, or
make 3 variables:
GreenRate
RedRate
BlueRate
with 1,2,3 or missing values, and these would be ordinal.
The choice depends on what analysis you are going to perform.
This way would well handle respondents who can't decide and absolutely
want to give 1 to red and blue.
In this case you don't need labeling in fact, but you can do:
Val lab GreenRate RedRate BlueRate
1'first choice'
2'second choice'
3'third choice'.
exe.
Analogous way for mcam54 proposition. (If does not work, write it for
each variable separately)
Monika
mca...@hotmail.com napisal(a):