Question about unpaired pre/post data

[Alyssa Na'im]

I have a data set of pre- and post-survey data that, because of the
design of the data collection, is not possible to match individually.  I
structured the file so that all pre data are grouped with missing post
data and vice versa.  When testing for significance between the two
groups, I run into warnings stating that the test cannot be completed
due to missing data, etc.  I'm inclined to revise the structure so that
all of the pre- and post-data are grouped by the same variable and and
are identified by a separate variable to indicate the source.  Before
doing so, however, I would like to know if this is the best approach for
this type of analysis.  Thanks in advance.

====================To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Maguin, Eugene]

Alyssa,

I don't know of an alternative. You will have to analyze your data as a
between groups design rather than as the within groups design that it is. In
terms of data management, I think that all you'll need to do is rename the
posttest vars to the same name as the pretest vars or vice-versa and add a
'grouping' variable that designates whether a case is in the pretest group
or the posttest group.

Gene Maguin

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD