Have data from survey that is 2 different setups.
One is from dBASE and is in the form:
Respondent1 Var1(Hi/Lo) Var2(Hi/Lo) Var3(Hi/Lo) Val1(1 to 7 Likert scale)
Respondent1 Var1 Var2 Var3 Val2(different scenario)
Respondent1 Var1 Var2 Var3 Val3
so one has for each row the same respondent but the different Val(ues)
for the scenarios, 15 scenarios total for this, for each respondent.
I read here about some code that could transpose this into one long line,
harder to read or scan but probably easier for analysis.
Then there is a separate 16th scenario for the same respondent:
Respondent1 Var6 Var7 Var8 Val1A(16th 1 to 7 Likert scale)
plus demographic variables of education, gender, views, and usual such
for about 10 Demographic Variables.
What would be the best approach. To fill out the first with 15 rows and
repeating the results of Var6.. with Val1A seems unwise and inelegant
(this worked with a merge function with SPSS 9.0.1). I thought of having 2
copies of SPSS running, one for the 15 rows and one for just a regular one
row for each case but SPSS does not permit this, having two copies
simultaneously - for the data entry and then figure out how to combine
the files for the later analysis.
Should I use SAS which is more powerful and then port it to SPSS for the
quick, preliminary analyses?
Probably the best way is just one longish row so each respondent has just
one line? I have to keep track of respondents for the demographic
variables as controls but since the survey is randomized, the results
could be looked at, for certain analyses, as random variables whose means
have meaning, given a randomized factorial survey (the sampling is
another problem and not what I am referring to in regards to randomness.)
Sorry to be slightly incoherent, but just working this out as I muddle
through before I actually start the coding.
Adam Sundor
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
