I have a survey that was conducted 2 years ago and a survey conducted this
year, each time we took a random sample of our population, and aside from a
very small number of individuals who randomly happened to take the survey
both times they are two distinct groups. We used the same screening criteria
and sampling methodology both times, but got a relatively low incidence rate
both times, (good response rate but not everyone qualified for the survey)
and both times we got a small sample sizes n~250.
Between the two surveys there are 11 similar questions that I would like to
compare, but I would like to somehow isolate whether the changes are
actually changes in opinion and behavior within the population or changes
due to the cross section of the population being different. I have looked at
the 95% confidence intervals around certain questions and it appears that
there are a few shifts, that need further examination.
Right now I have:
2 select all that apply questions which are essentially a string of about 10
dichotomous yes/no questions.
5 Likert scale questions on a 5 point likert with don't know and n/a set as
missing
4 Nominal level select the best option questions
Can anyone let me know what tests are best to run in this situation to test
for differences between the two years?
=====================
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
Locked
